阅读说明：本技术 减轻神经肌肉失调的影响 (Mitigate the influence of neuromuscular disorders ) 是由 尤里·克里蒙 谢里沙·米德拉 卡塔林·巴特菲-沃尔科特 英格里德·墨菲 瓦穆斯·瓦德翰·奇乌 于 2019-05-16 设计创作，主要内容包括：本公开涉及减轻神经肌肉失调的影响。在一些实施例中,公开的主题是具有可穿戴辅助设备的辅助系统,可穿戴辅助设备减轻神经肌肉失调(例如非预期运动或损失力量)的影响。当处于主动/预测模式时,辅助系统使用基于情境、操作和历史上下文的预测分析。当处于反应模式时,辅助设备在不改变用户的力量的情况下减轻非预期运动。辅助系统可以具有锻炼模式,用于评估用户的各种肌肉和关节的力量和灵活性,并且促进锻炼以避免进一步损失,或维持当前力量和灵活性。辅助系统利用来自耦合到辅助设备的传感器(以及可选地来自耦合到移动设备和环境中的传感器)的传感器数据。辅助设备上的致动器基于推断的预期动作或对非预期移动的反应来控制设备的移动。(This disclosure relates to mitigate the influence of neuromuscular disorders.In some embodiments, disclosed theme is the auxiliary system with wearable ancillary equipment, and wearable ancillary equipment mitigates the influence of neuromuscular disorders (such as unintended movements or loss strength).When have the initiative/prediction mode when, auxiliary system use the forecast analysis based on situation, operation and historical context.When being in reaction pattern, ancillary equipment mitigates unintended movements in the case where not changing the strength of user.Auxiliary system can have exercise mode, for assessing the various muscle of user and the strength and flexibility in joint, and promote to take exercise to avoid further loss, or maintain current strength and flexibility.Auxiliary system utilizes the sensing data from the sensor (and optionally from the sensor being coupled in mobile device and environment) for being coupled to ancillary equipment.Actuator in ancillary equipment is based on the expected movement acted or control the reaction of unexpected movement equipment of deduction.)

1. a kind of for mitigating the system of neuromuscular disorders, comprising:

Ancillary equipment, comprising:

Ancillary device sensor, for measuring the movement for wearing the muscle of user of the ancillary equipment, pressure or contraction and putting At least one of in pine；And

Actuator, for enhancing the muscle activity of the user；And

Processing circuit, for performing the following operations:

Sensing data is handled to infer the desired movement of the user, the sensing data is sensed from the ancillary equipment Device and environmental sensor are received；And

The actuator is controlled to realize the desired movement by the muscle for enhancing the user.

2. system according to claim 1, wherein the sensing data includes at least one of the following Measurement result: movement, object, gesture, voice, audible sound than speech, position or the degree of approach.

3. system according to claim 1, wherein in order to control the actuator by the muscle for enhancing the user Realize the desired movement, the processing circuit modifies the control based on operation mode, wherein the operation mode is One in the following terms: passive-reaction pattern, active-reaction pattern, active-prediction mode, override mode temper mould Formula.

4. system according to claim 3, wherein the passive-reaction pattern mitigates unintended movements, the active- The user of reaction pattern auxiliary loss strength, the active-prediction mode predict the desired movement, and the exercise Mode is used to that strength and flexibility to be promoted to retain and monitors the current ability of the user.

5. system according to claim 1, wherein in order to handle the sensing data to infer the expection of the user The sensing data is converted to contextual information by movement, the processing circuit, and the contextual information includes the following terms At least one of in: high probability situation context, high probability operation context or high probability move context.

6. system according to claim 5, wherein in response to by the user response in the control of the actuator And the audible order made, the override mode of the system are realized by the processing circuit, the override mode makes the place It manages circuit and executes following operation:

The control to the actuator is modified to defer to the audible order；And

Using from the sensing data current context and it is described it is audible order come re -training machine learning model, with Improve following deduction.

7. system according to claim 6, comprising: communication component, for matching to the physical object for including customization mode It sets the corresponding object of file and sends operation requests, the operation requests are in response to corresponding with the operation of the object described pre- Phase moves and is sent.

8. system according to claim 1, wherein in order to control the actuator, the processing circuit is based on the use The current ability at family come modify it is described control to adjust strength subsidiary level.

9. system according to claim 1, wherein in order to handle the sensing data to infer the expection of the user Movement, the processing circuit are realized:

Desired movement estimator, the desired movement estimator using derived from the sensing data context to generate Desired movement is stated, the desired movement includes one or more movements.

10. system according to claim 9, wherein the desired movement estimator includes the operation for the system Multiple level of accuracy of mode, wherein level of accuracy depends on the available sensor data and history in current context Data available in context, and wherein, institute is distributed on to the analysis of the available sensor data and historical context It states between the first processing circuit for including in ancillary equipment and the second processing circuit of the separate ancillary equipment, wherein described First processing circuit Internet access includes the memory of object profile known to the user, and the second processing is electric Road Internet access includes the memory of the object profile and historical context data for the unfamiliar object of the user, Wherein, first processing circuit, which is arranged in, disconnects Shi Yiyu with the second processing circuit and is communicatedly being connect Relatively low level of accuracy infers the desired movement when to the second processing circuit.

11. a kind of for mitigating the method for neuromuscular disorders, which comprises

Movement, pressure or the receipts of the muscle of the user of the ancillary equipment are worn using the device sensor measurement of ancillary equipment At least one of in contracting and loosening；

Processing sensing data to infer the desired movement of the user, the sensing data be from ancillary device sensor and Environmental sensor is received；And

The actuator of the ancillary equipment is controlled to realize the desired movement by the muscle for enhancing the user.

12. according to the method for claim 11, wherein the sensing data includes at least one of the following Measurement result: movement, object, gesture, voice, audible sound than speech, position or the degree of approach.

13. according to the method for claim 11, wherein control the actuator with by the muscle for enhancing the user come Realize that the desired movement includes: that the control is modified based on operation mode, wherein the operation mode is in the following terms One: passive-reaction pattern, active-reaction pattern, active-prediction mode, override mode or exercise mode.

14. according to the method for claim 13, wherein the passive-reaction pattern mitigates unintended movements, the master The user of dynamic-reaction pattern auxiliary loss strength, the active-prediction mode predict the desired movement, and described Exercise mode is used to that strength and flexibility to be promoted to retain and monitors the current ability of the user.

15. according to the method for claim 11, wherein the environmental sensor includes at least one of the following: Microphone, accelerometer, gyroscope, global positioning system (GPS) sensor, proximity sensor, position sensor, guide Needle, camera or biosensor.

16. according to the method for claim 15, wherein contextual information be provided to housebroken machine learning model with Infer the desired movement of the user.

17. according to the method for claim 11, wherein handle the sensing data to infer the expected fortune of the user Dynamic includes: that the sensing data is converted to contextual information, and the contextual information includes at least one in the following terms : high probability situation context, high probability operation context or high probability move context.

18. according to the method for claim 17, comprising: in response to by the user response in the control of the actuator and The audible order made realizes that override mode, the override mode include:

The control to the actuator is modified to defer to the audible order；And

Using from the sensing data current context and it is described it is audible order come re -training described in machine learning mould Type, to improve following deduction.

19. according to the method for claim 17, comprising:

Physical object configuration file is stored on a memory, and physical object configuration file includes normal modal configuration file, auxiliary At least one of mode configuration file or customization mode configuration file；And

The high probability, which is created, according to the physical object configuration file operates context.

20. according to the method for claim 18, comprising: to include customization the physical object configuration file of mode it is corresponding Object send operation requests, which is sent out in response to the desired movement corresponding with the operation of the object It send.

21. according to the method for claim 11, wherein controlling the actuator includes: the current energy based on the user Power come modify it is described control to adjust strength subsidiary level.

22. according to the method for claim 11, wherein handle the sensing data to infer the expected fortune of the user It is dynamic include: using derived from the sensing data context generate the desired movement, the desired movement includes one A or multiple movements.

23. according to the method for claim 22, wherein using derived from the sensing data context to generate Stating desired movement is performed using the technology with multiple level of accuracy for different operation modes, wherein accurate The horizontal data available depended in available sensor data and historical context in current context of degree, and wherein, it is right The analysis of the available sensor data and historical context is distributed on the first processing circuit in the ancillary equipment included Between the second processing circuit far from the ancillary equipment, wherein the first processing circuit Internet access includes the use The memory of object profile known to family, and the second processing circuit Internet access includes not yet done for the user The object profile for the object known and the memory of historical context data, wherein first processing circuit is arranged to Shi Yiyu is being disconnected to be communicatively connected to the second processing circuit phase relatively low with the second processing circuit Level of accuracy infer the desired movement.

24. at least one machine readable media, including instruction, described instruction by processing circuit when being executed, so that the processing Circuit perform claim requires method described in any one of 11-23.

25. a kind of system, the device including requiring method described in any one of 11-23 for perform claim.

Technical field

The embodiment of this theme relates in general to ancillary equipment, and more specifically but is not limited to, and is related to by inferring or pre- Desired movement is surveyed to mitigate the equipment of the influence of neuromuscular disorders (for example, tremble or lose strength).

Background technique

Millions of people suffers from degeneration (degenerative) motion control (for example, Parkinson's disease, multiple sclerosis etc.), Which suppress the abilities that people executes basic task (for example, feed, wear the clothes, write).Many people suffer from upper limb tremor.On Limb and lower limb also occur losing strength, flexibility and control force.Rock, tremble and muscle strength loss may be various nerves The result of obstacle.The basic reason of essential tremor (essential tremor) is unclear, also without effectively treatment side Method.With aging of population, it is contemplated that these statistical data will deteriorate.

Previous research and solution may include: various drugs；Deep brain stimulation (needs to perform the operation)；Wearable vibration Equipment；Spoon with stabilization；Weighing gloves；Or EMG signal filtering.Each technology in these technologies has one Or multiple defects.Drug can only may partly solve the problems, such as, in many cases, lack validity.The body of one people can Resistance can be generated to drug, and dosage must be constantly increasing.Some drugs may bring significant antagonism pair to make With for example, (for example, aggressive) etc. on (for example, kidney) physiologically, behavior.For being implanted into the operation of deep brain stimulation Cost is high, and carries out the invasive surgical with high fatal risk and serious side effects.Wearable vibratory equipment (for example, Emma project of Microsoft) not yet establishes curative effect.In addition, Emma does not have auxiliary mode (for example, giving user's strength To enhance the muscle of atrophy), and there is no exercise mode.Special spoon or fork with self-stable function, which can mitigate, only fits Solution for tableware.Weighing gloves are passively, and only to aggravate in gloves.The solution cannot be delayed completely Solution is trembled, and effect is poor for severe tremor and muscle loss.Gloves do not provide analysis, patient-monitoring or exercise Ability.Although carrying out the early stage research to filtering EMG signal to find the method for inhibiting to tremble, its effect is not yet true It is fixed.EMG filtering is also possible to cause problem because interfering desired movement when attempting filtering and trembling.EMG filtering will not monitor user Situation will not supplement the strength of partial loss.

Summary of the invention

One embodiment of the disclosure provides a kind of for mitigating the system of neuromuscular disorders, comprising: ancillary equipment, It include: ancillary device sensor, for measuring the movement for wearing the muscle of user of ancillary equipment, pressure or contraction and loosening At least one of in；And actuator, for enhancing the muscle activity of user；And processing circuit, for performing the following operations: Sensing data is handled to infer the desired movement of user, sensing data is connect from ancillary device sensor and environmental sensor It receives；And control actuator is to realize desired movement by the muscle for enhancing user.

Another embodiment of the present disclosure provides a kind of for mitigating the method for neuromuscular disorders, and method includes: to use During the device sensor measurement of ancillary equipment is worn movement, pressure or the contraction of the muscle of the user of ancillary equipment and is loosened At least one of；Sensing data is handled to infer the desired movement of user, sensing data is from ancillary device sensor and ring Border sensor is received；And the actuator of control ancillary equipment is to realize desired movement by the muscle for enhancing user.

The another embodiment of the disclosure provides at least one non-transitory machine-readable media, including for mitigating Neuromuscular The instruction of meat imbalance, instructs when being executed by processing circuit, so that processing circuit executes the operation including the following terms: using auxiliary In helping movement, pressure or the contraction of the muscle of the user of the device sensor measurement wearing ancillary equipment of equipment and loosening extremely One item missing；Sensing data is handled to infer the desired movement of user, sensing data is from ancillary device sensor and environment Sensor is received；And the actuator of control ancillary equipment is to realize desired movement by the muscle for enhancing user.

Detailed description of the invention

In the attached drawing being not drawn necessarily to scale, same reference numerals can describe the similar component in different views. Same numbers with different letter suffix can indicate the different instances of similar component.In the diagram of the following drawings, pass through Example rather than limitation mode show some embodiments, in which:

Fig. 1 is the diagram of the problem of being caused by neuromuscular disorders according to the embodiment and advanced solution and application；

Fig. 2 shows the various scenes according to the embodiment that ancillary equipment can be used；

Fig. 3 is the diagram for showing the sophisticated method according to the embodiment for the movement modulating device in auxiliary system；

Fig. 4 shows the dynamic non-individual body (dynamic continuum) of auxiliary equipment system according to the embodiment；

Fig. 5 is the block diagram for showing the various assemblies of auxiliary equipment system according to the embodiment；

Fig. 6 is the block diagram for showing the substitution inference engines of desired movement estimator (IMI) according to the embodiment；

Fig. 7 A-7C includes the various assemblies according to the embodiment for further showing auxiliary system as shown in figures 5-6 Block diagram；

Fig. 8 A-8D is according to the embodiment for assisting user to mitigate neuromuscular mistake in various operation modes including showing The flow chart of the method for the influence of tune；And

Fig. 9 is the exemplary block diagram for showing the machine that can realize one or more embodiments on it.

Specific embodiment

In the following description, for illustrative purposes, elaborate various details in order to provide to some example embodiments It understands thoroughly.It is apparent, however, to one skilled in the art, that in these no details or can have slightly The case where change, which gets off, practices this theme.

The embodiment of this theme is to be related to improving with those of the neurodegenerative movement control function of people and extend it solely The system and method for vertical viability.In addition, in embodiment, this theme includes that will be helpful to reduce the physics of muscle deterioration to control Treatment method.

Embodiment described herein several stages being in progress in degenerative disease to provide many benefits for user, for example, pushing away The desired movement of disconnected user is more accurately to execute the movement, mitigate and tremble, supplement the muscle strength of loss or suitably to use Family provides treatment and takes exercise.Furthermore, it is possible to it is customed to carry out to monitor the progress of user, or to medical professional, pharmacy mechanism, Or research institution etc. is reported.It should be noted that embodiment be not limited to the mankind application, can also be implemented to such as dog, horse, The animals such as cat.

The specific spy for combining the embodiment to describe is meaned to the reference of " one embodiment " or " embodiment " in specification Sign, structure or characteristic are included at least one embodiment of this theme.Therefore, (each place throughout the specification Occurring) appearance of the phrase " in one embodiment " or " in embodiment " is not necessarily all referring to identical embodiment, or not Necessarily all referring to different or mutually exclusive embodiments.The feature of various embodiments can be combined in other embodiments.

For illustrative purposes, elaborate concrete configuration and details in order to provide thorough understanding of the subject matter.However, right In it will be evident to one of ordinary skill in the art that the embodiment of described theme can be not presented herein specific It is carried out, or can be carried out with various combinations as described herein in the case where details.Furthermore, it is possible to be omitted or simplified known Feature, in order to avoid fuzzy described embodiment.Various examples can be provided throughout the specification.These examples are only to tool The description of body embodiment.The scope of the claims or meaning are not limited to the example provided.

Embodiment described herein different from the technology of such as virtual arm, low-pass filter or simple ectoskeleton etc.It is real Applying example may include following element, for example, context-aware, intent inference, the physical state for perceiving user, selective amplification (for example, needing the amount of mankind's strength reduction of selected muscle being related to corresponding with Given task) or continue to monitor use The strength or flexibility at family.

Ectoskeleton solution itself (assuming that there are the solutions) may promote muscular atrophy.Embodiment is by giving Auxiliary is limited in the case where settled preceding muscular states (for example, degree that muscle is weakened) to complete the degree of required by task It is interior, and the exercise by promoting the muscle for reduction, to help to postpone expected muscle deterioration (deterioration). Supplemental force be targetedly (for example, its correspond to reduction muscle) and be adjusted (for example, the amount of power is only replaced The muscle strength of loss, or be not enough to replace the muscle strength lost).It is directed to although following description is usually focused on The solution of hand, it should be appreciated that, technique described herein and method can be used for trembling inhibition or strength enhancing to it It is beneficial any limbs.For example, embodiment can be integrated with leg support frame, to increase stability, enhancing " upwardly and forwardly " Mobile or improvement posture.In another example, embodiment can provide balance exercise, and assess the dominant Parkinson's disease of posture State.In another example, embodiment may include neck or neck equipment, to keep a people when musculi colli is weaker Head it is upright etc..

Fig. 1 is 110, Yi Jizhen of the problem of neuromuscular disorders in the hand or arm according to the embodiment by patient cause To the diagram of the force level solution 120 (for example, strength of modification user) for being suitable for several applications 130 of these problems 100.When the hand of a people trembles, everyday tasks become more difficult.For example, when a people suffers from severe tremor, Holding coffee cup 101 may be nearly impossible without splashing out boiling hot coffee.Similarly, it buckles shirt 103, shake hands 105 or on keyboard typewriting 107 may become heavy task.

It in embodiment, include gloves 121 for the solution of hand tremor 120.Gloves 121 may include connector 123A-E is (for example, cable, ribbon cable, hydraulic line, pneumatic wire, shape change material (for example, marmem, intelligence are poly- Close object etc.), piezoelectric type electric wire etc.), for controlling each finger-joint according to actuator 127 (for example, motor, pump etc.). For example, gloves thumb may include pairs of connector 123A and 123B and for the corresponding of two thumb joint 125A-B Actuator 127.127 control connector 123A-E of actuator is to adjust (moderate) is directed to given joint or which direction vector How much pressure or power will be applied.In this example, actuator 127 can initiate or stimulate contraction of muscle or loosen.Gloves 121 can be with It is a part of wearable arm reinforcement, which uses equipped with sensor, actuator 127 and intelligence In determining desired movement, and arm, hand or other limbs are adjusted, correctly to realize the desired movement of user.For example, this can To include applying different grip pressures (for example, big grasping the grip pressure applied when bowling according to the object being crawled The grip pressure applied when grasping orange).

The application 130 for enhancing equipment may include various tasks.For example, user can pick up glass 131, without making glass Glass is fallen, or will not apply the pressure for being enough to make glass to rupture.It can improve automaticity (dexterity), such as allow user Buckle shirt 133.User can pick up the slender body 135A of such as smart phone etc, and then easily using intelligence Mobile phone 135B.The various shape factor of gloves 121 may be implemented, for example, the not gloves 137 of finger, or extended to from finger tip The shape adaptation gloves 139 of ancon.Depending on the obstacle (for example, tremble and lose strength etc.) of user, different shape adaptation It can be useful.

As described above, gloves 121 or other limb support parts (for example, skeletal muscle enhancing equipment) can be used to assist in use Desired movement is realized at family, for example, picking up cup, inhibiting to tremble.In addition, calculate node can be used to explain sensor information To infer the desired movement (for example, hold one glass of water, buckle shirt, shake hands etc.) of user.

Calculate node can maintain to perceive the history of user and current muscle strength and flexibility.It can be by being set to control The corresponding movement in the position that standby actuator sends signal selectively to expand with user's strength has failed.It is saved by calculating The system of point control can also provide exercise mode for user: a) execute effective exercise routine or b) strength of record user, Flexibility or the variation of motion range.When inferring user's desired movement, calculate node can maintain situation and historical context Perception.

Other than auxiliary user carries out everyday tasks and promotes that exercise is concentrated to fail to avoid further muscle, by The historical information that equipment (for example, in calculate node) is collected may be valuable in lasting patient care.For example, anti- Feedback data may have very big value for following personnel or industry: medical expert for progression of disease, drug (for example, titrate Deng)；Nursing staff's (for example, for creating or monitoring therapeutic scheme)；Or pharmacy industry (for example, for testing, studying).Feedback Data can processing locality (for example, passing through equipment) or by remotely processing (for example, the initial data from equipment is passed to Cloud component) to generate such as feature extraction, historical data or vertical analysis, can by various sources (for example, medical expert, Researcher etc.) consumption.

Fig. 2 shows the various scenes according to the embodiment that ancillary equipment can be used.It should be noted that in order to illustrate mesh , describe the auxiliary for hand or hand and handle composite.However, ancillary equipment can be used for various limbs, core trunk, Or neck support.Auxiliary can be provided in family, office or public environment based on available training and sensor information.It retouches The environment 200 (for example, home environment) for the various objects that can be interacted with user is stated.The center of diagram is that user can With the object seen in environment 200.In this case, coffee cup 203 is shelved on desk 209.It can be seen that in coffee There are tea bags 205 and spoon 207 in cup 203.Auxiliary system can be based on as described more fully below because of usually deduction or pre- User is surveyed to the intention of the object in environment 200.

In embodiment, environment 200 may include camera sensor 211, which can be with auxiliary needle to field Context (context) exploitation 213 of scape.Motion analysis and setting/scene determine that 210 can depend on context.For example, such as Visual aspects for the environment 200 that camera sensor 211 is captured and 213 can be auxiliary based on object or the exploitation of the context of situation Help the analysis to scene.In this example, cup 203 (object) can indicate the setting for having a drink, and possible movement It is to pick up cup 203.

In embodiment, environment 200 may include audio sensor 221.Audio sensor 221 can detecte the sense of hearing and mention Show, for example, the prompt of ambient sound (for example, breaking glass), speech or override (override) 223 are (for example, " stop！", " " etc. Admiration) or non-karst areas expression (for example, breathe, moan, roar), as described below.Desired movement infers that 220 can be used sound Frequency prompt carries out self calibration (for example, tuning) 220.Just attempt to help user's for example, user can perceive ancillary equipment Hand 201 is mobile to spoon 207, but cup 203 is caught in being actually intended that for user.In this case, for example, user can be with By saying that " being cup, be not spoon " or " I wants cup " carry out override devices.Self calibration and tuning block (example can be passed through Such as, pass through the training in terms of the context based on scene) identify override.

Position 231 and environmental context 233 can be used for auxiliary movement weighting 230.For example, user more likely plans to drink Coffee (for example, catch cup 203 and take cup 203 to lip), rather than execute movement related with door is opened.When with Family at kitchen or dining room, it is related with coffee cup 203 it is some movement more likely (with when user is in retail shop or bathroom Compared to).For example, empty coffee cup 203 is more likely placed in dish-washing machine (at the restaurant with user by user when user is at kitchen When compare).In this example, the list of possible movement can be defined, for example, hold cup/spoon/tea bag, turn the handle, Or buckle clothes.Weight can be distributed for each movement based on the situation and operation context discovered.It can initially manually Weight is distributed, or weight can be distributed based on training standard and machine learning.For some contexts, the power of movement is distributed to It can be 0 or 1 (for example, the range for being directed to percentage 0 to 1) again.Therefore, object can be grabbed using different force levels； For example, using stronger grasping on ceramic coffee cup, lighter grasping is used on foam cup.

As described above, having the device memory 241 of the information about prior actions can be with historical context 243 together It is used by next movement inference engines 240, to predict next movement according to similar context-sensitive prior actions.Example Such as, if prior actions are " with spoon mixing cup ", it may infer that next movement is " from cup water ".It should manage Solution, when auxiliary system is in passive and reaction pattern, predictability, which is inferred, to be skipped, to support alleviation of simply trembling.

Environment 200 (for example, using auxiliary system) may include eyes and body tracker 251.System may include view Line tracking equipment 253.It can be known by the user focus determined by eyes/body tracker 251 or view line tracking device 253 Other 250 interested object.For example, the sight as user is located on cup/spoon/tea bag (203/207/205) object composition When, inference engines 240, which can give lower weight distribution, is related to desk 209, family mark (home sign) 208 or clock 206 movement.

Environment 200 can also include proximity sensor 261, motion detector 263 or pressure sensor 265.Movement side The information from (one or more) sensor 261,263 or 265 can be used to analysis engine 260 to identify relative to each other User movement or object movement.It can be used together with space interpretation 273 relative to X, Y, Z coordinate position or mobile 271, with logical The degree of approach is crossed to identify object 270.For example, can detecte when user moves down their hand 201 to cup 203 It moves downward.When the relationship between the movement in processing space or the object in processing three-dimensional space, absolute position can be used Set (for example, X, Y, Z coordinate) or relationship positioning (for example, the distance between momentum, object or direction (for example, pitching, yaw, Roll) etc.) consider usual available six-freedom degree in three-dimensional setting.

It should be appreciated that and not all embodiments all will include such as all the sensors described in environment 200.For example, working as User outdoors when, fail to be communicably coupled to ancillary equipment (for example, Intelligent glove 121) or portable device (OTG) (example Such as, smart phone, tablet computer, wearable device, HMD etc.) sensor can not to auxiliary system provide input.This In the case of, prediction or real-time inference analysis may be limited to available sensor information.In environment 200, for example, based in kitchen It is directed to the prior actions of these objects in environment, the movement for picking up cup 203 can be predicted towards user's movement of cup 203. User may be intended to alternatively catch spoon 207.For example, for actively and predict helps, if strength auxiliary help too by force and Can not be without mobile direction be auxiliarily changed, then user may need to issue override.It helps to tremble to mitigate in passive and reaction In the case where, user can easily take cup 203 in the case where no clear override.In the ring with gaze tracker In the example in border, the sight of user towards tea bag 205 can be weighted with override historical context, and be correctly predicted user and wanted Catch tea bag 205 rather than the handle of coffee cup 203.In this example, being equipped with when user is in (can be integrated into auxiliary system In system) in the family of sensor or known environment when, various cameras and microphone can be fixed in the environment, and by coupling Close the mobile device (for example, wearable device, HMD, smart phone, ancillary equipment etc.) worn by user or held.When with When family is in the environment with limited sensor or without integrated sensor, the operation of ancillary equipment may be restricted, Some of them analysis and inference engines are omitted or reduce.It will be understood by those skilled in the art that upon reading this disclosure, depending on In available sensor and historical data, specific function can be optional or omission.

Fig. 3 is the diagram for showing the sophisticated method according to the embodiment for the movement modulating device 300 in auxiliary system. Moving modulating device 300 includes the Intelligent glove 310 (for example, ancillary equipment) with several actuators 301.Actuator 301 by Small circle instruction on Intelligent glove 310.It should be noted that and not all actuator 301 is all identified, so as not to keep attached drawing complicated Change.Actuator 301 can correspond to pressure or neural point in the limbs (for example, appendage, hand etc.) of user.When signal is sent When to actuator 301, the limbs of user can be applied a current to promote or inhibit the movement of muscle or joint.Intelligent glove 310 can also include firm and flexible ribbon, to help user to obtain additional strength.It is acted based on expected, ribbon can connect Required shape or juxtaposition (juxtaposition) are tightened, relaxed or is hardened to by signal.

In embodiment, comprehensive sensor network 330 can provide to desired movement estimator and movement modulation engine 320 The data of various sensors from ancillary equipment and in environment.(being carried out using the sensor in ancillary equipment) hand exercise Tracking 331 and (being carried out using the data from 337 sensor of position 333, voice 335 and vision) environmental modeling can be by For providing the environmental information from sensor to desired movement estimator and movement modulation engine 320.

In embodiment, it is contemplated that inferred motion device and movement modulation engine 320 can detecte movement, and identify object 321 and user for object expected movement 323.As described below, various housebroken machine learning models can be used to base History, situation in Analysis of Policy Making and report engine 340 and operation context infer the intention of user.Analysis of Policy Making and report Accusing engine 340 can be assisted by sending actuation commands (actuation command) 341 to movement modulation engine 320 Real-time control 307.In embodiment, parsing action modeling 343 can be used to provide actuation commands in Analysis of Policy Making and report engine 340 341.Analysis of Policy Making and report engine 340 can execute report generation 345 for drug titration, fine tuning, early warning etc.. The output of report generation 345 can be by various people or mechanism use preferably to look after user.For example, medical expert can monitor The situation of user, carer or relative can receive event alarms (for example, falling cup), and plant maintenance personnel can receive Equipment fault alarm etc..

Fig. 4 shows the dynamic non-individual body 400 of auxiliary equipment system according to the embodiment.Gloves can be passively 410 Or 420, or both combination of active.In completely passive mode 410, equipment be may be considered that in disengaged position (disengaged)411.When being in disengaged position 411, equipment will not both assist that user will not be hindered mobile.The mode can With used when not needing the help to particular task in user or the mode can during emergency or the system failure quilt It uses, for example, being initiated by user display order or according to the self diagnosis of instruction failure.

Equipment may be at reaction pattern 413, and in this mode, equipment for example in real time carries out user's desired movement anti- It answers.The natural movement that auxiliary equipment system can correspond to user operates, while inhibiting unconscious and trembling.The mode can be with It is used when user has elements but is trembled and is influenced.Another reaction pattern can provide and the flesh with strength reduction The corresponding selective amplification of meat.Level of amplification corresponds to the level of muscle loss, and can be stored in can be visited by calculate node In the data storage asked.The mode is applicable when user has part strength residue.

Equipment may be at prediction mode 415.In prediction mode 415 (for example, autonomous mode), ancillary equipment can be with base In situation context or when user start under certain situation movement when predictably (for example, user by they muscle send out Out before motor message) initiate movement.The mode is applicable when user largely loses strength.For example, if user is difficult to Setting in motion (for example, hand or arm are lifted away from body), then ancillary equipment can be by identifying nearby object, identification time of day Possible situation etc. carrys out prediction action.For example, user may usually have breakfast in 8 a.m..In morning 8:15, user faces coffee Cup.Based on context, ancillary equipment can trigger muscle response to reach for and catch coffee cup.Equipment can be trained to To identify various common scenes or be learnt from iterative task.

Fig. 5 is the block diagram 500 for the various assemblies for showing auxiliary equipment system according to the embodiment.For illustrative purpose, It can be shown in component used in existing system with dotted line frame.It is exclusively used in the embodiment of ancillary equipment as described herein Component is shown with dotted border.These components be integrated into the user with neuromuscular problem provide it is improved auxiliary set It is standby.Diagram is divided into five funtion parts, layer or mode: sensed layer 510；What layer 520 has occurred in determination；Judgement (or user's meaning Figure) layer 530；Controlled motion layer 540；And update step 550.

In embodiment, auxiliary equipment system includes the force applicator that can be realized with various technologies (for example, ectoskeleton Shape gauntlet), for example, controlling hydraulic or pneumatic wire, cable or ribbon etc. or intelligence using electrodynamic pump, motor or valve Polymer, biologic artifact etc. increase the electromechanical assembly of the muscle strength of user to applied force.Force applicator may include Multiple sensors (being not shown, for example, movement, pressure etc.) and actuator (for example, actuator 127 shown in Fig. 1).At one In embodiment, 127 active suppression of actuator is confirmed as the unexpected movement trembled.In another embodiment, gloves 121 are kept Rigidity (is trembled) for example, substitution is actively resisted, so that user feels that their hand is fixed in gloves 121.It is such Equipment may be comfortable for the people of hand tremor.The mechanical organ of the mobile gloves of actuator is (for example, 123 He of ribbon Joint 125), it is corresponding with the expection muscle activity of user.Actuator movement corresponds to the muscle strength of user's application (that is, not having Have amplification) or " amplification " movement corresponding with the amount of muscle loss in specific muscle of the actuator property of can choose ground.

Referring again to FIGS. 5, in embodiment, can cumulatively aggregating sensor data to determine user in space and environment In real time kinematics (for example, motion detector), be used for context-aware.Sensing data can be from all related and can It is polymerize with sensor device.Sensor may include: sensor array 511 in Intelligent glove equipment (for example, acceleration Meter, gyroscope, heat sensor, pressure sensor, other physiology and movable sensor)；(or other can wear the OTG equipment of user Wear equipment, smart phone or mobile device) in sensor 513；Sensor 515 in external equipment is (for example, camera, close Spend sensor etc.).Sensor 513 in OTG equipment may include physics and biosensor (for example, global positioning system (GPS), accelerometer, gyroscope, proximity sensor, microphone, headset equipment (HMD) camera, heart rate monitor, other Biosensor etc.).Sensor 515 (for example, camera, proximity detector, microphone etc.) in external equipment can be used for Situation context is provided.

In embodiment, the sensor 511 in equipment, user OTG equipment in sensor 513 and outside set The sensor information of sensor 515 in standby is provided to layer 520, with determine about user have occurred what or in the environment What (for example, situation context of user) has occurred.Motion detector 521 can identify the movement of user and close to user Object movement.Object and gesture identification component 523 identify user gesture and the object close to user.Object identifying can be auxiliary Situation context is helped, for example, auxiliary is picked up and holds the coffee cup with liquid.Object identifying can identify that user is connecing Nearly stair and may need auxiliary to catch railing (or in the case where the ancillary equipment of leg, auxiliary to step on upward or downward It steps on).Natural language processor (NLP) 525 can identify the voice from audio sensor (for example, microphone).In embodiment In, user can provide audible (for example, speech or non-karst areas) order or feedback.Embodiment can identify than speech can Listening.For example, can identify doorbell and infer that user will stand up to open the door.Or in another example, the teapot blown a whistle Can produce user will enter kitchen to close the deduction of stove.There are location components 527 where to identify user.Example Such as, when the user is at home, compared with when user is in work or shopping, different situation contexts may be relevant.Another In example, proximity sensor can sense user and approach the door with particular kind of locking mechanism.In the example In, close to the aspect for locking mechanism being situation context.

Once various movements, identification, language and location context are polymerize and are identified in layer 520, so that it may by information It is supplied to and judges layer 530.Judge layer 530 based in movement, environment object, gesture identification, auditory cues (for example, speech or Non-karst areas expression, ambient sound etc.) or position come identify user be intended to.In embodiment, it is contemplated that inferred motion device (IMI) 531 The information from layer 520 can be used to determine movement expected from user, then can for example be used by controlled motion layer 540 The user is expected to be moved to provide control or movement modulation intelligence, to complete desired movement.In embodiment, 531 IMI Determine which element of the movement of user is it is contemplated that rather than by unexpected caused movement of trembling in real time.IMI 531 passes through By movement with situation context relation (for example, passing through analysis motion profile) and optionally from user's acquisition prompt (for example, language Sound and eye sight line) and prompt (for example, sound) is obtained from environment to make a determination.In embodiment, lose in muscle strength Under the extreme case of mistake, IMI 531 can infer that desired movement is (such as following in the case where the muscular movement independent of user It is discussed with reference to Fig. 6).

In embodiment, IMI 531 includes situation context-memory (SCM) component 532, situation context identifier (SCI) 533, functional object configuration file (FOP) 534 and study/inference engines 535.IMI 531 can receive as explicit The input of user's interaction override component 536, for example, in the case where user intends to carry out uncertain movement, or in user Intend in the case where IMI 531 when accurately inferring the expected movement of user without correcting.IMI 531 can also be received Self calibration and tuning (SCT) information 538, to improve the deduction to desired movement.

In embodiment, SCI 533 is based on real time sensor data and comes from (for example, from the acquisition of SCM 532) first The prompt of preceding situation context is determined for what user and environment have occurred.SCI 533 provide to historical pattern, position It sets, the identification of time and situation context and the relevant situation of real time sensor data.Pass through correlation, deduction and deviation The standard deviation of norm calculates high probability result.SCI 533 can receive input element, comprising: in environmental context, situation Hereafter and real time sensor data.Environmental context provides information related with previous position and time index, appoints to identify Daily mode and the associated location of anticipating and time-based event.Situation context is associated with preceding events, for example, ring Relevant information between border, position and the correlate occurred in the position.Real time sensor data is provided for context Additional accuracy and details.SCI 533 determines high probability context using these inputs, and ensures to more and more accurately Model improves.

In embodiment, SCM 532 provides historical context storage device.Historical context is sent out in environment and situation The history of raw event, and be used to improve a possibility that event occurring under given current context.Therefore, Yong Huke It is had breakfast with 8 points of every morning in history.IMI 531 can be by the historical context from SCM 532 (for example, having sent out Raw event) it is integrated with the current context from SCI 533, more accurately to infer that user is intended to.For example, using Historical data corresponding with breakfast activity in SCM 532, and movement of the user from SCI 533 in corridor, IMI 531 may infer that user is moving towards kitchen just to start from coffee.In another example, IMI 531 can be in user's morning 7: Infer that user will carry out dress when 30 opening wardrobes or dressing table drawer.But if it is at 3 points in afternoon, user is opened Dressing table drawer may indicate that user will pack up clean clothing.It can be various based on time of day, position, movement etc. Possible situation allocation probability.SCI 533 and SCM 532 can provide high probability situation context, act as with exercise data one It is inputted for the core of IMI 531.

In embodiment, SCM 532 assists SCI 533 by defining the relationship between the movement having occurred and that, thus Continuous inferred motion is provided.In this example, the hand of user stretches to the direction of cup, and puts and begin to write beside cup.Known use A piece of paper is put on desk shortly before family, this helps to be inferred to user may to go to take pen.Known this is breakfast time And user just holds fork, this helps to be inferred to user may to go to take cup.SCM 532 can pass through the phase of movement Close property (for example, direction, vector, closing rate etc.) and object refine to assist providing the identification to situation and event, thus Predefine the possibility object that will be manipulated in context.532 data of SCM can be used in various ways.For example, fortune Dynamic inference engines (for example, being located in IMI 531) can assess object and serial correlation, so as to based on object properties and for grasping Vertical sequence of events responds to distinguish.In another example, approximating sequence fallout predictor can provide specific (specificity) And process, the selection to which object will be manipulated is reduced to involve the object of upcoming movement based on most probable.

FOP 534 may include the list of special object, the special object have on how to control (for example, operation) this The specific respective attributes of a little objects.When applicable, information of the auxiliary system based on such as user location etc executes in lists It searches.When applicable, profile information auxiliary IMI 531 determines the expected movement of user.For example, auxiliary system can be examined User location is surveyed in the aisle ends of second floor.Lookup in the profile list provides following information, this is to lead to bathroom Door, opener be in knob shape, need to rotate clockwise a quarter circle to open door.The object-based function of FOP 534 The identification of operation context to the object in list is provided with operation mode.This may include standard object (for example, general It is publicly available), modification object (for example, auxiliary object) (for example, wherein power, grasping, surface be modified to enhance or assist it is fixed To) and custom-built system (it is modified to the limitation for specially enhancing single user).FOP 534 may include for master die State, auxiliary mode, automatic modality or the Object Operations element for customizing mode.Here, automatic modality is grasped automatically including wherein object Make to mitigate user's applied force with those of operation object mode.Example may include can be by building (for example, family Front yard) automated network control automatic door open device.Therefore, the user's degree of approach or speech order detected can open or close Door.

Normal modal element or component can based in the case where no any enhancing everyday objects in the situation Mode makes object Attended Operation and prediction/positive action.Assist Modal elements can be based on some auxiliary to everyday objects Enhance (it allows better simply manipulation, for example, enhancing relevant to Americans with disabilities act (ADA) auxiliary) participate in object Operation and prediction/positive action.Customizing Modal elements can be based on the distinctive customization enhancing of user (for example, family improves and vapour Vehicle/driving enhancing) make object Attended Operation and prediction/positive action.In this example, communication component is (for example, transmitter, receipts Send out device etc.) signal or order can be sent to object to request the object to assist user in a manner of predefined customization.For example, In this example, request can be on or off automatically-controlled door.FOP 534 can provide high probability operation context, be IMI 531 third core input.

Study or inference engines 535 may include feeding back with the decision-making capability of sustained improvement auxiliary system.For example, working as user When being intended to (for example, being intended to override prediction intention with practical) by speech interference prediction, for example, when IMI 531 judges user's meaning by accident When figure, self study adjustment can be triggered.Inference engines (as discussed below with reference to Fig. 6) in 535 may include passing through It practises, analysis and feedback cycle provide the high probability context component of improved accuracy.

In embodiment, many movements shown in Fig. 5 can be handled in ancillary equipment, so that user does not need Rely on access cloud or internet connection.Some processing can be performed on this destination edge cloud, for example, in user family or handling official business On local server in room, or even on smart phone or other mobile devices.In order to provide a user more independence, The common object of user, situation and movement can have the local profile in the memory being stored in ancillary equipment.It is auxiliary The training of auxiliary system or user's monitoring and report may need to access internet or cloud, however, training, monitoring or report can also It can be in local progress.In this example, training may need have foot using complicated and intensive training pattern is calculated Preferably off-line execution in the calculate node of enough power.Cloud or Internet Server may be needed to access receive doctor input, The configuration file etc. of new object.Self calibration input (for example, override) can be applied immediately, or be saved to and connected in ancillary equipment It is connected to the time later for calculating powerful equipment with re -training model when.It should be appreciated that with Local or Remote Cloud Server phase Than the various combination of processing or component may reside in ancillary equipment, this is necessary for providing a user auxiliary in real time.

Fig. 6 is the block diagram for showing the substitution inference engines of IMI according to the embodiment.In embodiment, inference engines 630 with The form of probability receives core output from SCM 532, SCI 533 and FOP 534, respectively include: high probability moves/close to upper Hereafter 632, high probability situation context 633 and high probability operate context 634.Using probability be because user movement not It is deterministic.User movement is simulation, rather than number, and almost there is unlimited possibility.Accordingly, it is possible to feelings Border context, operation context and movement/close to context can be assigned probability before association.Study analysis is fed back to Road inference engines 635 receive continuous contextual feedback using these high probability context-predictions, and by that will operate Hereafter, situation context and movement/be associated close to context and carry out training system.It should be appreciated that these can be used in association Any of context type is as the principal element (for example, as seed, beginning or starting point etc.) in training.For example, Training can focus first on situation (for example, situation context is seed), and by situation and historical operation (for example, in operation Hereafter) it is associated with movement (for example, movement or close to context).Alternatively, training can be paid close attention to associated with situation and movement Historical movement.In embodiment, various contexts can be trained to reciprocity variable.It should be appreciated that various machines can be used The context relation of three types is variable input by learning model.

Once the machine learning model in study analysis and feedback cycle inference engines 635 can be to IMI by training 631 provide one or more high probabilities deduction for situation, movement/movement and operation.Although should be appreciated that inference engines 630 are illustrated in figure 6 as individual block or component, but inference engines 630 may be implemented as the sub-component of IMI 631, or Inference engines 630 and the various assemblies of IMI 631 may be implemented as individually processing or distributed treatment or combination.IMI 631 (for example, for situation, movement/movement and operation) high probability is used to infer, and to kinetic control system 640 (for example, figure Layer 540 or MME 545 shown in 5) motion information is provided.

Referring again to FIGS. 5, IMI 531 can to movement modulation engine (MME) 545 provide about user intention (for example, Desired movement) it is what high probability information, and MME 545 uses the information to execute movement (for example, enhancing user oneself Muscle is inputted to realize desired movement).In embodiment, MME 545 activates ancillary equipment (for example, electromechanical assembly, Intelligent glove Deng) force applicator to realize desired movement.MME 545 for example receives the information about user movement from motion detector 521, And desired movement is received from IMI 531.Based on operation mode or other parameters (for example, user's strength and limitation), or prediction With react user's auxiliary mode.545 determination of MME will by actuator (for example, actuator 127) or pass through other signaling (examples Such as, for operating the radio signal of automatically-controlled door or lamp switch) which is activated move.MME 545 can determine which movement will It needs strength to expand or assist using the congenital or current strength of user and which movement, and which actuator is needed Complete expected movement.Each actuator can provide muscle auxiliary or inhibition of trembling to single muscle or one group of muscle.Therefore, such as The degree that fruit user only loses strength or muscle loss in some muscle is different, then can be swashed with different strength levels Actuator living.It trembles if the specific muscle of user has, actuator corresponding with the muscle that trembles is susceptible to suffer from can be activated To inhibit to tremble.In another embodiment, no actuator is activated for the desired movement of user, and user can be with congenital Strength is completed to move.However, in this case, inhibition can be used for unintended movements, for example, tremble, twitch or other Dyskinesia.

In embodiment, MME 545 helps to realize three previously described operation modes: system initialization, standard are used Family operation mode and exercise mode.MME 545 can based on receive system operation modes label or identifier 541 and differently Operation.System initialization mode can be executed by kinsfolk, nurse, therapist or other people, with imitate user operation/ Wear daily routine when ancillary equipment.The imitation to daily routine can be executed to train auxiliary system and provide for inciting somebody to action Situation, movement and the baseline for operating context relation.The training can also fill the object profile in FOP 534.Standard User operation mode is the mode for user's auxiliary.MME 545 may include machine learning to learn and finely tune its operation. When ancillary equipment is by user's operation, initial baseline is can be enhanced in study when operation.

In embodiment, auxiliary system includes exercise mode.The mode for user provide workout scheme (for example, stretching, extension and Strengthen).In this mode, system can prompt user before user participates in resistance exercise course first.MME 545 can root Locally adjusted according to remaining strength in (one or more) specific muscle using how many resistance or stretching, and record user Progress is to generate report to user and health worker.Workout scheme and recommendation may include doctor's input 542.Record data may be used also To be used to the operation of the calibrator (-ter) unit under Standard User operation mode.

MME 545 can also utilize adeditive attribute, for example, user's physiological parameter (UPP) 543 and doctor's input 542.UPP 543 may include the database of user's current health state, for example, corresponding to the strength and flexibility of specific muscle and joint. UPP 543 continuously update based on exercise mode by system, or from external source (for example, doctor, physiotherapy Teacher) it receives.Doctor's input 542 can provide and specifically temper which (which) muscle and recommend which (which) exercise related Information, including frequency and duration or number of repetition.In embodiment, UPP 543 can also include specific instruction with more Energetically inhibit to tremble (for example, because this be comfortable for users), or do not inhibit to tremble on the contrary (for example, because of In some cases, this is uncomfortable for users).

Controlled motion layer 540 can also include actuator control manager (ACM) 546, home automation controllers 548, And move or move starter 547.ACM 546 can drive the actuator network in ancillary equipment (for example, Intelligent glove) To execute special exercise based on guide/instruction from MME 545.Movement or mobile starter 547 can be to each actuatings Device sends signal or instruction to realize movement.Home automation controllers 548 can with the network of home automation type and set It is standby to be communicated.For example, IMI 531 can determine that user intends to open the door for leading to bathroom.As requiring user to open manually Door alternative solution, IMI 531 can to home automation controllers 548 send order, digital door opener can with when send out Order is played to open automatically-controlled door via home automation network.

In embodiment, auxiliary system may include emergency override 537, is used to indicate auxiliary system and stops and move immediately Except any inhibition/amplification/interference to user movement.The override 537 can (keyword sighs with feeling, moans based on special sound order Chant, breathe, shout), gesture or the activation of specific " stopping " button/switch etc..For example, when user feels that equipment is not assisted When making or providing incorrect auxiliary/inhibition, user can deactivate ancillary equipment (for example, controlling manager 546 to actuator Real-time signal is sent to deactivate actuator).It can be deactivated to trigger by physics or virtual switch or button or voice command. In embodiment, deactivating (for example, emergency override) can be by other people triggering in addition to a user, for example, kinsfolk, nursing Personnel, emergency response personnel, medical expert etc..

In embodiment, auxiliary system may include user's interaction override component 536, for perceiving equipment just in user When making incorrect prediction but not needing emergency cut-off, correction is provided to IMI 531.For example, IMI 531 may infer that use Family travels to take cup.User can roar " I just attempts to pick up spoon ".The order or mobile suggestion can use NLP 525 identifications, and corrective action is used to by IMI 531.It can be entered for the situation of the correction, operation and movement context To study/inference engines 535 for additional training.It should be noted that other prompts, which can be used, carrys out override IMI 531.For example, Specific action (for example, hand one draws or pull back and forth hand or other gestures suddenly) can be pre-defined to indicate that IMI 531 is pushed away Disconnected decision inaccuracy.

In embodiment, auxiliary system may include self calibration tuner (SCT) 538.SCT 538 may include being used for The continuous self study platform of IMI 531 and MME 545.In a simple form, for IMI 531, as user clearly override IMI When 531 deduction, the machine learning weight of the adjustable model of SCT 538.When user indicates that special exercise is insufficient (for example, hand Finger does not hold door handle fully with rotating handles) or moving keeps user's generation uncomfortable (for example, due to from actuator Excessive power) when, the self calibration of MME 545 can be executed.

In embodiment, MME 545 can provide historical information to update step 550.Update step 550 may include user's prison Survey and report (UMR) 551 component.UMR 551 can (for example, via cloud 553) to external staff or entity (for example, therapist, Research institution etc.) report or analysis are provided.UMR 551 can also be related with the force level of user more to the offer of UPP 543 New information and other current operating informations, for the use of MME 545.(for example, operating mould in user's specification during auxiliary mode During formula and especially during exercise mode), auxiliary system can with the current muscle strength of continuous assessment user, flexibly Property and engine control.UMR 551 can update 543 database of UPP.UMR 551 can also monitor a series of physical attribute (for example, strength, intensity of trembling, bioassay and specific activities).

It can be by according to predefined scale (for example, unified Parkinson's disease measuring scale (UPDRS)) addition objective evaluation To enhance UMR 551.In this example, fixed time test can be executed to check that patient can be in appointed interval by their index finger The number or user that touch thumb can rotate upwards/be rotated down the number of their hand, the measurement as flexibility. The test subset that progress can be easily automated by ancillary equipment can be with periodic operation (for example, once a week or every two weeks one It is secondary).(this is the current typical case of patient for these strength and flexibility test and the test only executed in year/semi-annual physical examination Situation) it compares and can more frequently be performed.These test results can report cloud 553, for doctor, therapist and its His interested party uses.The data may have very big value for doctor, therapist, pharmacy, research institution etc., for providing Better drug titration, physiotherapy (for example, exercise), or it is used to help the individual of other similar situation.In embodiment, believe Cease it is processed, and (one or more) specific feature set be utilized suitable degree of secret protection or encryption provide.

Fig. 7 A-7C includes various groups for further showing substitution auxiliary system as shown in figures 5-6 according to the embodiment The block diagram of part.Fig. 7 A shows 533 component of (for example, shown in Fig. 5) SCI according to the embodiment.Situation context identifier 533 can be used high probability context (situation) input 741 to identify high probability situation context 633, in the high probability situation It 633 can hereafter be used by IMI (631, Fig. 7 C).Several 730 Hes of input data can be used in situation context identifier 743 741 and derived input (for example, context identifier 703,713 and 723) identify situation context.High probability context (situation) 741 can be exported from environmental context 703, previous scenario context 713 and historical context 723.Environmental context 703 can be used previous position and time index 701 come environment-identification (for example, family, office, grocery store, dining room, corridor, Kitchen, bathroom etc.).Preceding events 711 can be used to identify previous scenario context 713 (for example, can mean in kitchen Need to have a meal or have a drink using tableware).Historical context 723 can be used foregoing history may information 721 (for example, Mean that cup may be picked up in there are the kitchen of cup object).Situation context identifier 743 can be used on high probability Hereafter (situation) 741 and real time sensor data 730 generate high probability situation context 633.In embodiment, housebroken Machine learning model can be used to determine at one or more points/calculate various context identifiers 703,713,723 Or 741, wherein model specific to for exporting context identifier 703,713,723,743 or 633/741 input 701, 711,721,730 or 741 data.

In this example, user enters the known local cafe using dixie cup.Situation context can be indicated a level pressure Power is placed on cup safely to hold it.Another cafe may use the ceramic cup for needing different pressures to hold.If User often patronizes two caves, then can save as two positions and default cup configuration file applied to special object Previous scenario context.Therefore, when user enters the shop using ceramic cup, location probability will be with object or evolutionary operator probability Association is to apply pressure appropriate.Possible that day shop will will use different types of cup or cafe is unknown.At this In a little situations, user may say " paper coffee cup is used in I " to remind ancillary equipment to select correct object configuration text Part, as discussed in more detail below.

In embodiment, user can indicate that user intends the coffee that access uses ceramic cup by the sense of hearing or other modes Shop.Historical context may infer that a series of actions for going to that cafe, point are single and drink coffee.When user is away from home, auxiliary The subset that historical action and functional object configuration file only can be used is helped in the local storage of equipment.For example, user can It can know that the network connection of cafe is unreliable.The situation context can be stored as to remote cloud server or this destination edge cloud Position or movable configuration file are directed in server.It in embodiment, can be based on expected situation and operation context (example Such as, cafe is gone to) extract various historical actions and functional object configuration file or a series of expected movements in advance from cloud.Cause This, when extracting data in advance, be necessarily dependent upon usually in the subset for the information being locally stored compared with, ancillary equipment can be with Higher level of accuracy is operable to infer desired movement and movement.

Referring now to Fig. 7 B, the block diagram diagram of the associated situation by movement and event according to the embodiment is shown. The mark of situation and event can be exported, from the association of movement (direction, vector, closing rate) and the refinement of object so as to pre- First determine the possibility object to be manipulated within a context.As described above, SCM 532 (Fig. 5) may include inferred motion engine 752 With approximating sequence fallout predictor 754.Inferred motion engine 752 assesses object and serial correlation 751, to be based on object properties and use It is responded in the sequence of events of manipulation to distinguish.Approximating sequence fallout predictor 754 provides specificity and process, to be involved based on most probable The object of upcoming movement reduces the selection to the object that will be manipulated.Object meaning refinement 753 is a kind of technology, is led to Object Selection (for example, which object will be used) can be refined by crossing the technology.For example, working as close to several object (examples Such as, the pen on desk, notepad, newspaper and cup) when, it may be difficult to determine that user wants and which of several objects pair As interacting.However, to the user behavior pattern for putting down pen and catching cup later is scheduled on, for Object Selection can be refined Such as there is when just putting down more maximum probability selection cup (compared with selecting newspaper).SCM 532 can be to above and below high probability Literary (object) 755 provides input.High probability object context 755 can be used to determine with inferred motion and about object The situation context-memory 756 of approximating sequence information (for example, operation context).Situation context-memory 756 can be to IMI 631 (Fig. 7 C) provides high probability movement/close to context 632.

Referring now to Fig. 7 C, the frame of the identification of the operation context according to the embodiment for showing object 634 (Fig. 6) is shown Figure.It can identify that high probability operates context 634 based on object functionality and operation mode.Object may include standard object (for example, general public is available, such as papery or ceramic coffee cup), modification object (for example, auxiliary object) (for example, wherein power, Grasp, surface be modified to enhancing or auxiliary directional) and customization mode object (it is modified to specially enhance single user Limitation).High probability identification (operation) 769 can be grasped from Object Operations (standard) 762, Object Operations (auxiliary) 764 or object Make (customization) 766 to export.Normal modal configuration file 763 can be used to identify user for mark in Object Operations (standard) 762 The possible operation of quasi- object.Auxiliary mode configuration file 765 can be used to identify user for auxiliary in Object Operations (auxiliary) 764 Help the possible operation of object.Object Operations (customization) 766 can be used customization mode configuration file 767 and carry out identity user for fixed The possible operation of object processed.For example, object can be identified as the coffee cup of not handle, this can be to dixie cup or foaming polyphenyl Vinyl foam cup distributes the probability of higher (compared with ceramic cup).

The operation of (for example, shown in Fig. 5) FOP 534 based on for the determining function of high probability identification (operation) 769 with Mode is operated to provide the identification of the operation context to object.Functional object configuration file 534 can provide defeated to IMI 631 Enter, to determine that high probability operates context 634 (Fig. 6).Functional object configuration file 534 is commonly stored with object and object such as What is operated related information.For example, the functional object configuration file for kitchen microwave ovens may include: kitchen microwave ovens position In where (for example, be located at what room, from the ground it is how high, be located at sales counter above etc.)；Handle is located at where (for example, being located at Left side is located at top etc.)；Start button is located at where, which type of signal specific which can issue (for example, sound Sound, lamp etc.) etc..It in another example, may include the type for the opening mechanism that door has for the configuration file of door, including Following details, for example, whether position, rotation mode (for example, clockwise or counterclockwise) or door that handle is placed there is electromechanics to open Door ability.As described above, IMI 631 provides desired movement information, motion control layer 540 and then control to motion control layer 540 Force applicator is to realize desired movement.

Fig. 8 A-C is according to the embodiment for assisting user to mitigate neuromuscular mistake in various operation modes including showing The flow chart of the method for the influence of tune.In block 801, auxiliary system identifies operation mode.Ancillary equipment can be initial with system Change or Setting pattern 810, user's auxiliary mode 820 or exercise mode 830 operate.

Referring now to Fig. 8 B, the flow chart of the method for system initialization 810 according to the embodiment is shown.In frame 811, It can identify and register the sensor that can be used for inputting.Method for finding automatically can be used to identify the sensing in range Device.Alternatively, user can manually position and register sensor.Sensor may include ancillary equipment, other OTG or can wear Wear the sensor on equipment and environmental unit.In frame 812, the data correlation for the sensor being registered for can establish Mapping.Sensing data association can establish the relationship between sensor (for example, sensor position relative to each other, sensing Device relative to the position (for example, position, height etc.) of environment, the data of sensor it is how interrelated (for example, in time, In the angle of override etc.)), and intent inference is carried out based on this.In frame 813, environment profile data can be created Library 840.In frame 814, situation configuration files database 842 can be created, and be used to capture situation context, and For example, notice environment configuration files database 840.Environment profile and situation configuration files database 840 and 842 can be from Environment collects some sensing datas, and fills configuration file using training data.In this example, nurse or kinsfolk can To walk about in the home environment.Camera is (for example, be coupled to head-mounted display (HMD) (for example, glasses, goggles, the helmet Deng) or be installed on wall) can be with capturing ambient information, for example, being located in corridor, bathroom, kitchen, family room etc.. Common activities of the user in these environment can be used to input or train situation context.

In frame 815, the initial physiological parameter of user can be pre-loaded to user's physiological parameter (UPP) database 844 In.User's strength and restricted information can be assumed, receive from health care professionals or export from test, and defeated Enter as baseline.It will be understood that can before environment profile database 840 and situation configuration files database 842 or it UPP database 844 is filled afterwards.

Referring now to Fig. 8 C, the flow chart of the method for user's auxiliary mode 820 according to the embodiment is shown.It should manage Solution, user's auxiliary mode 820 can assist in dynamic non-individual body with the passive of various levels, reaction or active, such as Fig. 4 institute Show.When ancillary equipment is in user's auxiliary mode 820, in frame 821, auxiliary system is received for various contexts or movement Sensing data, including but not limited to, movement, Object identifying, position, environment and user feedback.Sensing data can be gathered It closes to be associated with the information from various sensors.Historical context can also be associated with to export in real time up and down with sensing data Text.As set forth above, it is possible to carry out the association of auxiliary data using various machine learning models.When being in prediction mode, in frame 822 In, it may infer that desired movement.In reaction pattern, desired movement and unintended movements also may infer that, to inhibit non-pre- Phase movement.In this example, user unable may provide from mobile predicted motion.In such a case, it is possible to base Carry out predicted motion in context.In frame 823, movement, which can be modulated to mitigate, trembles, mitigates unintended movements or to be expected Movement provides auxiliary strength.Movement modulation can be (rather than prediction) of reaction, for modulating and execute movement.Movement is adjusted System can strength and limitation parameter 844 based on user.Movement modulation may include send signal 825 to ancillary equipment (for example, Intelligent glove, flexible leg bracket, neck support etc.) in actuator.It should be appreciated that movement modulation can be the (example of active Such as, strength assists) or (for example, tremble subtract) of reaction, or both, as by operation mode and the identification of user's physiological parameter (for example, using UPP database 844).As described above, user's monitoring and report can be executed in frame 824.In frame 826, The calibration and tuning of auxiliary can be executed, as described above.

Referring now to Fig. 8 D, the process of the method for the exercise mode 830 according to the embodiment for showing ancillary equipment is shown Figure.In frame 831, user or system can choose exercise to practice.When Systematic selection exercise mode 830, system can To prompt user to start or participation activity.In embodiment, forging can be obtained from the list that defined physiotherapy tempers 846 Refining.It can periodically suggest tempering (or needing to manually select) to user.User, which can pre-select, various is automatically reminded to mould Formula.Once being selected or being proposed, in frame 832, user can be prompted to start to take exercise.In frame 833, user movement is identified Sensing data 848 can be collected and be used to infer desired movement.In frame 834, it can according to need to modulate fortune It is dynamic, and in frame 835, actuator is sent signal to assist desired movement or inhibit unintended movements.In frame 836, Sensor in ancillary equipment monitors user.Strength water can be inferred based on movement, pressure and the other sensors in equipment It is flat.In frame 837, the variation in user's strength or limitation or other physiological parameters can be updated in UPP database 844.In In frame 838, the report of the relating to parameters with the state of workout scheme and update can be generated.Report be can store for looking into later It sees, or is transmitted to user or medical care and health personnel as needed.

Fig. 9 shows the example that can execute any one or more technologies (for example, method) being discussed herein on it The block diagram of machine 900.In embodiment, ancillary equipment (for example, Intelligent glove or leg support frame etc.) and machine 900 are (for example, clothes Business device machine) communicated, machine 900 can be used for executing housebroken model and based on context data based on deduction Desired movement provides motion control.Machine 900 can be Local or Remote computer, or such as smart phone, tablet computer Or the processing node in the OTG equipment of wearable device etc.Machine 900 can be used as autonomous device and be operated, or can be with It is connected (e.g., networked) to other machines.In embodiment, machine can be with ancillary equipment direct-coupling or integrated.It should manage Solution, when processor 902 is directly coupled to ancillary equipment, some components of machine 900 can be omitted to provide light weight and spirit Equipment (for example, display equipment, UI navigation equipment etc.) living.In networked deployment, machine 900 can be operated as server- Server machine, client machine in client network environment, or both.In this example, machine 900 may be used as equity (P2P) peer machines in (or other are distributed) network environment.Machine 900 can be personal computer (PC), tablet PC, machine Top box (STB), personal digital assistant (PDA), mobile phone, Web appliance, network router, interchanger or bridge can (sequentially or otherwise) executes instruction any machine of (the specified movement to be taken by machine of these instructions).In addition, Although only showing individual machine, term " machine " should also be as being understood to include that executing one group (or multiple groups) alone or in combination refers to It enables to execute the arbitrary collection of the machine for any one or more methods being discussed herein, for example, cloud computing, software service (SaaS) or other computer set group configurations.

As described herein, example may include logic or multiple components or mechanism, or can by logic or multiple components or Mechanism operates.Circuit is the set for the circuit realized in tangible entity comprising hardware is (for example, ball bearing made, door, patrol Volume etc.).Circuit member can be flexibly with the variation of bottom hardware over time.Circuit includes can be in operation The member of specified operation is executed either individually or in combination.In this example, the hardware of circuit can be designed immutablely to execute spy Fixed operation (for example, hardwired).In this example, the hardware of circuit may include the physical assemblies of variable connection (for example, executing list Member, transistor, ball bearing made etc.), the computer-readable medium (example of the instruction of specific operation is encoded including physically being modified Such as, magnetic, electrical, removable constant aggregate particles placed etc.).When connecting physical assemblies, the basis of hardware component Electrical characteristics are changed, for example, becoming conductor from insulator, vice versa.The instruction so that embedded hardware (for example, executing list Member or loading mechanism) member of circuit can be created within hardware by variable connection, to execute specific operation in operation Part.Therefore, when the device operates, computer-readable medium is communicably coupled to the other assemblies of circuit.In this example, arbitrarily Physical assemblies can be used in the more than one member of more than one circuit.For example, in operation, execution unit can be One time point is used in the first circuit of the first circuit unit, and in different time by the first circuit unit Two circuits reuse, or are reused in different time by the tertiary circuit in second circuit unit.

Machine (for example, computer system) 900 may include hardware processor 902 (for example, central processing unit (CPU), Graphics processing unit (GPU), hardware processor core or their any combination), main memory 904 and static memory 906, some or all of which can communicate with one another via interconnecting link (for example, bus) 908.Machine 900 can also wrap Include display unit 910, Alphanumeric Entry Device 912 (for example, keyboard) and user interface (UI) navigation equipment 914 (for example, Mouse).In this example, display unit 910, input equipment 912 and UI navigation equipment 914 can be touch-screen display.Machine 900, which can also comprise storage equipment (for example, driving unit) 916, signal generating device 918 (for example, loudspeaker), network, connects Jaws equipment 920 and one or more sensors 921 (for example, global positioning system (GPS) sensor, compass, accelerometer, Or other sensors).In this example, sensor 921 may include wearable sensors, the sensor based on ancillary equipment and Environmental sensor, as described above.Machine 900 may include o controller 928, for example, serial (for example, universal serial bus (USB)), parallel or other wired or wireless (for example, infrared (IR), near-field communication (NFC) etc.) connection is to communicate or control one A or multiple peripheral equipments (for example, printer, card reader etc.).

Store equipment 916 may include machine readable media 922, be stored on the machine readable media 922 one group or Multi-group data structure or instruction 924 (for example, softwares), one or more groups of data structures or instruction 924 embody any one or Multiple technique described hereins or function, or utilized by any one or more technique described hereins or function.Instruction 924 exists By machine 900 execute during can also completely or at least partially reside in main memory 904, in static memory 906 or In hardware processor 902.In this example, hardware processor 902, main memory 904, static memory 906 or storage equipment One in 916 or any combination may be constructed machine readable media.

Although machine readable media 922 is shown as single medium, term " machine readable media " may include by Be configured to store one or more instructions 924 single media or multiple media (for example, centralized or distributed database or Associated cache and server).

Term " machine readable media " may include that can store, encode or carry to be executed by machine 900 and make machine Device 900 execute the disclosure any one or more technologies instruction or can store, encode or carry by it is this kind of instruct make With or to this kind of arbitrary medium for instructing relevant data structure.Non-limiting machine readable media example may include that solid-state is deposited Reservoir and optics and magnetic medium.In this example, large capacity machine readable media include possess it is multiple have it is constant (for example, It is static) machine readable media of the particle of quality.Therefore, large capacity machine readable media is not transient state transmitting signal.Large capacity The specific example of machine readable media may include: nonvolatile memory, for example, semiconductor memory devices are (for example, electricity can Program read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM)) and flush memory device；Disk, for example, Internal hard drive and moveable magnetic disc；Magneto-optic disk；With CD-ROM and DVD-ROM disk.

Instruction 924 can also utilize a variety of biographies via network interface device 920 using transmission medium by communication network 926 Defeated agreement is (for example, frame relay, Internet protocol (IP), transmission control protocol (TCP), User Datagram Protocol (UDP), hypertext Transport protocol (HTTP) etc.) in any one be transmitted or received.Example communication network may include local area network (LAN), wide Domain net (WAN), packet data network (for example, internet), mobile telephone network (for example, cellular network), ordinary old style telephone (POTS) network and radio data network are (for example, be known as Institute of Electrical and Electric Engineers (IEEE) 802.11 Family of standards is known as802.16 family of standards of IEEE), IEEE 802.15.4 family of standards, equity (P2P) network etc.. In this example, network interface device 920 may include one or more physical receptacles (for example, Ethernet, coaxial or phone are inserted Hole) or one or more antenna to be connected to communication network 926.In this example, network interface device 920 may include multiple days Line, to use in single input and multi-output (SIMO), multiple-input and multiple-output (MIMO) or multiple input single output (MISO) technology extremely Lack one to carry out wireless communication.Term " transmission medium " be understood to include can store, encode or carry for by Any intangible medium for the instruction that machine 900 executes, and including number or analog communication signal or for promoting this kind of software Communication other intangible mediums.

Additional information and example

Example 1 is a kind of for mitigating the system of neuromuscular disorders, comprising: ancillary equipment, comprising: ancillary equipment sensing Device, for measure wear movement, pressure or the contraction of the muscle of user of ancillary equipment and in loosening at least one of；And it causes Dynamic device, for enhancing the muscle activity of user；And processing circuit, for performing the following operations: processing sensing data is to push away The desired movement of disconnected user, sensing data is received from ancillary device sensor and environmental sensor；And control actuating Device is to realize desired movement by the muscle for enhancing user.

In example 2, the theme of example 1 includes, wherein sensing data includes to movement, object, gesture, voice, removes The measurement result of audible sound, position or at least one of the degree of approach except voice.

In example 3, the theme of example 1-2 includes, wherein in order to control actuator with by enhance user muscle come Realize desired movement, processing circuit modifies control based on operation mode, wherein operation mode is one in the following terms: Passively-reaction pattern, active-reaction pattern, active-prediction mode, override mode or exercise mode.

In example 4, the theme of example 3 includes, wherein passive-reaction pattern mitigates unintended movements, active-reaction The user of mode auxiliary loss strength, active-prediction mode predicts desired movement, and exercise mode is for promoting strength and spirit Activity retains and monitors the current ability of user.

In example 5, the theme of example 1-4 includes, wherein environmental sensor includes at least one of the following: Microphone, accelerometer, gyroscope, global positioning system (GPS) sensor, proximity sensor, position sensor, guide Needle, camera or biosensor.

In example 6, the theme of example 5 includes, wherein contextual information is provided to housebroken machine learning model To infer the desired movement of user.

In example 7, the theme of example 1-6 includes, wherein for the expected fortune for handling sensing data to infer user Dynamic, sensing data is converted to contextual information by processing circuit, which includes at least one of the following: High probability situation context, high probability operation context or high probability move context.

In example 8, the theme of example 7 includes, wherein in response to being made by user response in the control of actuator Audible order, the override mode of system circuit processed realize that override mode makes processing circuit execute following operation: modification pair The control of actuator is to defer to audible order；And using from sensing data current context and it is audible order come again Training machine learning model, to improve following deduction.

In example 9, the theme of example 5-8 includes: memory, for storing physical object configuration file, physical object Configuration file includes normal modal configuration file, assists at least one of mode configuration file or customization mode configuration file, Physical object configuration file circuit processed is for creating high probability operation context.

In example 10, the theme of example 8-9 includes: communication component, for include customization mode physical object The corresponding object of configuration file sends operation requests, and the operation requests are in response to desired movement corresponding with the operation of object And it is sent.

In example 11, the theme of example 1-10 includes, wherein in order to control actuator, processing circuit is based on user's Current ability controls to modify to adjust strength subsidiary level.

In example 12, the theme of example 1-11 includes, wherein for the expection for handling sensing data to infer user Movement, processing circuit are realized: desired movement estimator, desired movement estimator use context derived from sensing data Desired movement is generated, which includes one or more movements.

In example 13, the theme of example 12 includes, wherein desired movement estimator includes the operation mode for system Multiple level of accuracy, wherein level of accuracy depend on current context in available sensor data and in history under Data available in text, and wherein, the analysis of available sensor data and historical context is distributed in ancillary equipment Including the first processing circuit and far from ancillary equipment second processing circuit between, wherein the first processing circuit Internet access Memory including object profile known to user, and second processing circuit Internet access includes being unfamiliar with for user Object object profile and historical context data memory, wherein the first processing circuit is arranged in and the Two processing circuits, which disconnect Shi Yiyu relatively low level of accuracy when being communicatively connected to second processing circuit, to be come Infer desired movement.

Example 14 is a kind of for mitigating the method for neuromuscular disorders, this method comprises: using the equipment of ancillary equipment Sensor measurement wear movement, pressure or the contraction of the muscle of the user of ancillary equipment and in loosening at least one of；Processing passes For sensor data to infer the desired movement of user, sensing data is received from ancillary device sensor and environmental sensor； And the actuator of control ancillary equipment is to realize desired movement by the muscle for enhancing user.

In example 15, the theme of example 14 includes, wherein sensing data include to movement, object, gesture, voice, The measurement result of at least one of audible sound, position or the degree of approach than speech.

In example 16, the theme of example 14-15 includes, wherein control actuator with by enhance user muscle come Realize that desired movement includes: to modify control based on operation mode, wherein operation mode is one in the following terms: passive- Reaction pattern, active-reaction pattern, active-prediction mode, override mode or exercise mode.

In example 17, the theme of example 16 includes, wherein and passive-reaction pattern mitigates unintended movements, actively-anti- Answer mode auxiliary loss strength user, active-prediction mode predict desired movement, and exercise mode for promote strength with Flexibility retains and monitors the current ability of user.

In example 18, the theme of example 14-17 includes, wherein environmental sensor includes at least one in the following terms : microphone, gyroscope, global positioning system (GPS) sensor, proximity sensor, position sensor, refers to accelerometer Compass, camera or biosensor.

In example 19, the theme of example 18 includes, wherein contextual information is provided to housebroken machine learning mould Type is to infer the desired movement of user.

In example 20, the theme of example 14-19 includes, wherein processing sensing data is to infer that the expected of user is transported Dynamic includes: that sensing data is converted to contextual information, which includes at least one of the following: high general Rate situation context, high probability operation context or high probability move context.

In example 21, the theme of example 20 include: in response to by user response in the control of actuator make can Order is listened, realizes override mode, which includes: the control of modification degree actuator to defer to audible command；And it uses Current context and audible order from sensing data carry out re -training machine learning model, to improve following deduction.

In example 22, the theme of example 18-21 includes: to store physical object configuration file, physics pair on a memory As configuration file include normal modal configuration file, auxiliary mode configuration file or customize mode configuration file at least one It is a；And high probability is created according to physical object configuration file and operates context.

In example 23, the theme of example 21-22 includes: to opposite with including the customization physical object configuration file of mode The object answered sends operation requests, which is sent in response to desired movement corresponding with the operation of object.

In example 24, the theme of example 14-23 includes, wherein control actuator includes: the current ability based on user To modify control to adjust strength subsidiary level.

In example 25, the theme of example 14-24 includes, wherein processing sensing data is to infer that the expected of user is transported It is dynamic include: using derived from sensing data context generate desired movement, which includes one or more move Make.

In example 26, the theme of example 25 includes, wherein using derived from sensing data context it is pre- to generate Phase movement is performed using the technology with multiple level of accuracy for different operation modes, wherein accuracy water Data available in flat available sensor data and historical context depending in current context, and wherein, to available The analysis of sensing data and historical context is distributed on the first processing circuit for including in ancillary equipment and sets far from auxiliary Between standby second processing circuit, wherein the first processing circuit Internet access includes depositing for object profile known to user Reservoir, and second processing circuit Internet access include for the unfamiliar object of user object profile and in history under The memory of literary data, wherein the first processing circuit, which is arranged in, to be disconnected Shi Yiyu with second processing circuit and led to Letter when being connected to second processing circuit relatively low level of accuracy infer desired movement.

Example 27 is at least one non-transitory machine-readable media, should including the instruction for mitigating neuromuscular disorders Instruction by processing circuit when being executed, so that processing circuit executes the operation including the following terms: using the equipment of ancillary equipment Sensor measurement wear movement, pressure or the contraction of the muscle of the user of ancillary equipment and in loosening at least one of；Processing passes For sensor data to infer the desired movement of user, sensing data is received from ancillary device sensor and environmental sensor； And the actuator of control ancillary equipment is to realize desired movement by the muscle for enhancing user.

In example 28, the theme of example 27 includes, wherein sensing data include to movement, object, gesture, voice, The measurement result of at least one of audible sound, position or the degree of approach than speech.

In example 29, the theme of example 27-28 includes, wherein control actuator with by enhance user muscle come Realize that desired movement includes: to modify control based on operation mode, wherein operation mode is one in the following terms: passive- Reaction pattern, active-reaction pattern, active-prediction mode, override mode or exercise mode.

In example 30, the theme of example 29 includes, wherein and passive-reaction pattern mitigates unintended movements, actively-anti- Answer mode auxiliary loss strength user, active-prediction mode predict desired movement, and exercise mode for promote strength with Flexibility retains and monitors the current ability of user.

In example 31, the theme of example 27-30 includes, wherein environmental sensor includes at least one in the following terms : microphone, gyroscope, global positioning system (GPS) sensor, proximity sensor, position sensor, refers to accelerometer Compass, camera or biosensor.

In example 32, the theme of example 31 includes, wherein contextual information is provided to housebroken machine learning mould Type is to infer the desired movement of user.

In example 33, the theme of example 27-32 includes, wherein processing sensing data is to infer that the expected of user is transported Dynamic includes: that sensing data is converted to contextual information, which includes at least one of the following: high general Rate situation context, high probability operation context or high probability move context.

In example 34, the theme of example 33 includes, wherein operation include: in response to by user response in actuator The audible order for controlling and making, realizes override mode, the override mode include: modification to the control of actuator to defer to the sense of hearing Order；And using from sensing data current context and it is audible order come re -training machine learning model, to change Into following deduction.

In example 35, the theme of example 31-34 includes, wherein operation includes: to store physical object on a memory to match File is set, physical object configuration file includes normal modal configuration file, auxiliary mode configuration file or customization mode configuration text At least one of part；And high probability is created according to physical object configuration file and operates context.

In example 36, the theme of example 34-35 includes, wherein operation include: to include customization mode physics pair As the corresponding object transmission operation requests of configuration file, the operation requests are in response to expected fortune corresponding with the operation of object It moves and is sent.

In example 37, the theme of example 27-36 includes, wherein control actuator includes: the current ability based on user To modify control to adjust strength subsidiary level.

In example 38, the theme of example 27-37 includes, wherein processing sensing data is to infer that the expected of user is transported It is dynamic include: using derived from sensing data context generate desired movement, which includes one or more move Make.

In example 39, the theme of example 38 includes, wherein using derived from sensing data context it is pre- to generate Phase movement is performed using the technology with multiple level of accuracy for different operation modes, wherein accuracy water Data available in flat available sensor data and historical context depending in current context, and wherein, to available The analysis of sensing data and historical context is distributed on the second processing circuit for including in ancillary equipment and sets far from auxiliary Between standby third processing circuit, wherein second processing circuit Internet access includes depositing for object profile known to user Reservoir, and third processing circuit Internet access include for the unfamiliar object of user object profile and in history under The memory of literary data, wherein second processing circuit, which is arranged in, to be disconnected Shi Yiyu with third processing circuit and led to Letter when being connected to third processing circuit relatively low level of accuracy infer desired movement.

Example 40 is a kind of for mitigating the system of neuromuscular disorders, which includes: for using ancillary equipment The device of device sensor measurement at least one of the following: the muscle of movement, pressure or contraction and the user loosened auxiliary are set It is standby；The desired movement of user is inferred for handling sensing data, it is received from ancillary device sensor and environmental sensor The device of sensing data；With the actuator for controlling ancillary equipment to realize expected fortune by the muscle for enhancing user Dynamic device.

In example 41, the theme of example 40 includes, wherein sensing data include to movement, object, gesture, voice, The measurement of at least one of audible sound, position or the degree of approach than speech.

In example 42, the theme of example 40-41 includes, wherein for controlling actuator by the flesh of enhancing user Meat includes: for the device based on operation mode modification control the device of realizing desired movement, wherein operation mode is following One in items: passive-reaction pattern, active-reaction pattern, active-prediction mode, override mode or exercise mode.

In example 43, the theme of example 42 includes, wherein and passive-reaction pattern mitigates unintended movements, actively-anti- Answer mode auxiliary loss strength user, active-prediction mode predict desired movement, and exercise mode for promote strength with Flexibility retains and monitors the current ability of user.

In example 44, the theme of example 40-43 includes, wherein environmental sensor includes at least one in the following terms : microphone, gyroscope, global positioning system (GPS) sensor, proximity sensor, position sensor, refers to accelerometer Compass, camera or biosensor.

In example 45, the theme of example 44 includes, wherein contextual information is provided to housebroken machine learning mould Type is to infer that the expected of user acts.

In example 46, the theme of example 40-45 includes, wherein for handling sensing data to infer that user's is pre- The device of phase movement includes: the device for sensing data to be converted to contextual information, which includes following At least one of in items: high probability situation context, high probability operation context or high probability move context.

In example 47, the theme of example 46 include: in response to by user response in the control of actuator make can Order is listened, for realizing the device of override mode, comprising: for modifying the control to actuator to defer to the dress of audible command It sets；And for using current context from sensing data and audible order to carry out re -training machine learning model to change Into the device of following deduction.

In example 48, the theme of example 44-47 includes: the dress for storing physical object configuration file on a memory It sets, physical object configuration file includes in normal modal configuration file, auxiliary mode configuration file or customization mode configuration file At least one；And the device for creating high probability operation context according to physical object configuration file.

In example 49, the theme of example 47-48 include: for include customization mode physical object configuration file Corresponding object sends the device of operation requests, the operation requests in response to desired movement corresponding with the operation of object and It is sent.

In example 50, the theme of example 40-49 includes, wherein the device for controlling actuator includes: for being based on The current ability of user controls to modify to adjust the device of strength subsidiary level.

In example 51, the theme of example 40-50 includes, wherein for handling sensing data to infer that user's is pre- The device of phase movement includes: the device for generating desired movement for using derived from sensing data context, the expection Movement includes one or more movements.

In example 52, the theme of example 51 includes, wherein is used for use context next life derived from sensing data Device at desired movement is performed using the technology with multiple level of accuracy for different operation modes, In, level of accuracy depends on the data available in available sensor data and historical context in current context, and Wherein, the analysis of available sensor data and historical context is distributed on the first processing circuit for including in ancillary equipment and Between second processing circuit far from ancillary equipment, wherein the first processing circuit Internet access includes that object known to user is matched The memory of file is set, and second processing circuit Internet access includes the object profile for the unfamiliar object of user With the memory of historical context data, wherein the first processing circuit is arranged in when disconnecting with second processing circuit To infer desired movement with level of accuracy relatively low when being communicatively connected to second processing circuit.

Example 53 is at least one machine readable media, including instruction, and the instruction by processing circuit when being executed, so that place It manages circuit and executes operation with any one of implementation example 1-52.

Example 54 is a kind of device, including the device for realizing any one of example 1-52.

Example 55 is a kind of system, for realizing any one of example 1-52.

Example 56 is a kind of method, for realizing any one of example 1-52.

Technique described herein is not limited to any specific hardware or software configuration；They can be adapted for any calculating, disappear Power-consuming son or processing environment.These technologies can be realized with hardware, software, firmware or combinations thereof, to generate support herein The logic or circuit of execution or the realization of the embodiment of description.

For simulation, program code can indicate using hardware description language or another functional description language (its substantially Provide the expected model how to execute of hardware of design) hardware.Program code can be assembler language or machine language or The data that can be compiled or interpreted.In addition, in the art, refer to software take movement or generation with some form the result is that Common.Such expression is only to illustrate that (this makes processor execution act or produce by processing system execution program code Raw result) shorthand way.

Each program can with high level procedural, declaratively or the programming language of object-oriented is realized, with processing system System is communicated.However, if it is desired to which program can be realized with assembler language or machine language.Under any circumstance, language It can be compiled or interpreted.

Program instruction can be used to so that executing behaviour described herein with the general or specialized processing system of instruction programming Make.Alternatively, operation can be executed for executing the specialized hardware components of the firmware hardwired logic operated by including, or by warp knit Any combination of the computer module of journey and custom hardware components executes.Method described herein may be provided as computer Program product (is also been described as computer or machine-accessible or readable medium), may include that one or more machines can visit It asks storage medium, is stored with instruction on it, which can be used to be programmed processing system or other electronic equipments To execute method.

Program code or instruction can be stored in for example volatibility or nonvolatile memory (for example, storage equipment or Correlation machine is readable or machine accessible medium, including solid-state memory, hard disk drive, floppy disk, optical storage, tape, flash memory, Memory stick, digital video disks, digital versatile disc (DVD) etc.) and more peculiar medium (machine-accessible biological state guarantors Storage) in.Machine readable media may include any machine for storing, sending or receiving information in machine readable form System, and medium may include tangible medium, by the tangible medium can transmit it is electronics, optical, sound or other The coding of form has transmitting signal or carrier wave, such as antenna, optical fiber, communication interface of program code etc..Program code can be with The form of grouping, serial data, parallel data, transmitting signal etc. is transmitted, and can be come with compression or encryption format using.

Program code can realize with the program executed on programmable machine, for example, mobile or fixed computer, a Personal digital assistant, smart phone, mobile internet device, set-top box, cellular phone and pager, consumer-elcetronics devices (including DVD player, personal video recorder, personal video player, satellite receiver, stereo receiver, cable television receiver Device) and other electronic equipments, respectively include the readable volatibility or nonvolatile memory of processor, processor, at least one Input equipment or one or more output equipments.Program code can be applied to the data inputted using input equipment, to execute Described embodiment and generate output information.Output information can be applied to one or more output equipments.This field is general Logical technical staff is appreciated that the embodiment of disclosed theme can be implemented with various computer system configurations, including more Processor or multi-core processor system, minicomputer, mainframe computer and it can be embedded into general in almost arbitrary equipment And or microcomputer or processor.The embodiment of disclosed theme can also be in distributed computing environment, cloud environment, equity Or implement in network micro services, wherein task or part thereof can be held by the remote processing devices being linked through a communication network Row.

Processor subsystem can be used to execute the instruction in machine readable or machine accessible medium.Processor subsystem System may include one or more processors, and each processor has one or more cores.In addition, processor subsystem can be by It is arranged on one or more physical equipments.Processor subsystem may include one or more application specific processors, for example, figure Processing unit (GPU), digital signal processor (DSP), field programmable gate array (FPGA) or fixed function processor.

Although operation can be described as order process, some operations can actually concurrently, simultaneously or It is performed in distributed environment, and wherein program code locally or long-range is being stored for uniprocessor or multiprocessor machine Device accesses.In addition, in some embodiments, essence of the sequence without departing from disclosed theme of operation can be rearranged Mind.Program code can be used by embedded controller or is used in combination with embedded controller.

As described herein, example may include circuit, logic or multiple components, module or mechanism, or can circuit, It is operated in logic or multiple components, module or mechanism.Module can be communicably coupled to one or more processors from And execute the hardware, software or firmware of operation described herein.It should be appreciated that module or logic with hardware component or can be set The standby, software that runs on the one or more processors or firmware, or combinations thereof realize.Module can be through shared or biography Delivery data and integrated different and independent components or module can be the sub-component of individual module, or be divided into multiple Module.Component can be in single calculate node run or realize process, or can be distributed across concurrently, simultaneously, The process between multiple calculate nodes sequentially or in combination run, is such as described more fully in conjunction with the flow chart in attached drawing 's.In this way, module can be hardware module, and module is considered the tangible entity for being able to carry out specified operation, and And it can be configured or be arranged in a specific way.In this example, can in a specific way by circuit (for example, internally or External entity relative to such as other circuits etc) it is arranged as module.In this example, one or more computer system (examples Such as, independent client or server computer system) or all or part of of one or more hardware processor can be by Firmware or software (for example, instruction, application obscure portions or application) is configured to operation to execute the module of specified operation.In this example, Software may reside on machine readable media.In this example, software holds hardware in the bottom hardware execution by module The specified operation of row.Therefore, term hardware module is understood to include tangible entity, which is physically to be constructed, specially Door ground is configured (for example, hardwired) or provisionally (for example, temporarily) is configured (for example, being programmed) to be with specific mode Operated or executed the entity of some or all of any operation described herein.Consider that wherein module is provisionally configured Example does not need to instantiate each module at any one time.For example, including by using software configuration, arrangement or tune in module In the case where whole common hardware processor, common hardware processor can be configured as corresponding different mould in different time Block.Therefore, software can configure hardware processor to for example constitute particular module a moment, and at different times Constitute different modules.Module is also possible to software module or firmware module, operates for executing method described herein.

In this document, as common in patent document, term "a" or "an" be used to include one or one More than a, this is independently of any other "at least one" or the example or usage of " one or more ".In this document, unless separately Have an instruction, term "or" be used to refer to nonexcludability or, i.e. " A or B " include " A rather than B ", " B rather than A " and " A and B".In the appended claims, term " includes " and the popular English for " wherein " being used as corresponding term "comprising" He " wherein " Equivalent in language.In addition, in the following claims, it includes in addition to that term " includes " and "comprising", which are open, The system, equipment of the element except this kind of term element listed below, article or processing are still considered as falling into the power a bit In the protection scope that benefit requires.In addition, in the following claims, term " first ", " second " and " third " etc. are used only as marking Number, and be not intended to force the numerical order requirement to their object.

Although describing this theme by reference to illustrative embodiments, which is not intended to limit or constrain Meaning explains.For example, above-mentioned example (or in terms of one or more) can be used with other example combinations.It can make With other embodiments, for example, what those of ordinary skill in the art will be understood that when reading this disclosure.Abstract is for permitting Perhaps essence disclosed in reader's fast discovery technology.It should be appreciated that abstract is not used in explanation or limitation power when however, submitting abstract The range or meaning that benefit requires.