阅读说明：本技术 车辆能量管理系统和方法 (Vehicle energy management system and method ) 是由 王兴池 D·普弗洛姆 于 2020-05-12 设计创作，主要内容包括：提出了一种车辆能量管理系统。所述系统包括：接收单元,其用于接收与车辆的运动状态相关的数据；以及确定单元,其用于利用机器学习模型基于所述数据确定等效燃料因子。所述机器学习模型是基于已知的与车辆的运动状态相关的数据和相对应的等效燃料因子训练得到的。通过利用与车辆的运动状态相关的数据确定等效燃料因子,能够无需知道驾驶工况而得到等效燃料因子,这简化了系统并且提高了确定等效燃料因子的准确性。(A vehicle energy management system is presented. The system comprises: a receiving unit for receiving data relating to a motion state of a vehicle; and a determination unit for determining an equivalent fuel factor based on the data using a machine learning model. The machine learning model is trained based on known data relating to the motion state of the vehicle and a corresponding equivalent fuel factor. By determining the equivalent fuel factor using data relating to the motion state of the vehicle, the equivalent fuel factor can be obtained without knowing the driving conditions, which simplifies the system and improves the accuracy of determining the equivalent fuel factor.)

1. A vehicle energy management system, comprising:

a receiving unit for receiving data relating to a motion state of a vehicle; and

a determination unit for determining an equivalent fuel factor based on the data using a machine learning model;

wherein the machine learning model is trained based on known data relating to a state of motion of the vehicle and a corresponding equivalent fuel factor.

2. The vehicle energy management system of claim 1,

the determination unit is further configured to derive at least one feature data based on the data; and determining the equivalent fuel factor based on the at least one characterizing data.

3. The vehicle energy management system of claim 2,

wherein the at least one item of feature data comprises the following feature data within a predetermined time period: maximum velocity, average velocity, standard deviation of velocity, zero velocity time scale, average positive acceleration, and average negative acceleration.

4. The vehicle energy management system of any of claims 1-3,

wherein the machine learning model is further adaptively trained based on historical driving data of the vehicle, the historical driving data including data relating to a moving state of the vehicle recorded in past driving of the vehicle and an equivalent fuel factor calculated from fuel consumption of the past driving.

5. The vehicle energy management system of any of claims 1-3,

wherein the receiving unit is configured to receive the data for a predetermined time period at a predetermined frequency; and the processing unit is further configured to update the determined equivalent fuel factor at the predetermined frequency based on the data for the predetermined time period.

6. The vehicle energy management system of any of claims 1-3, further comprising:

a prediction unit for predicting a travel speed and/or a route of the vehicle; and

an adjusting unit for adjusting the equivalent fuel factor determined by the determining unit based on the predicted travel speed and/or route.

7. A vehicle energy management method, comprising:

receiving data relating to a state of motion of a vehicle; and

determining an equivalent fuel factor based on the data using a machine learning model;

wherein the machine learning model is trained based on known data relating to a state of motion of the vehicle and a corresponding equivalent fuel factor.

8. The vehicle energy management method of claim 7, further comprising:

deriving at least one item of feature data based on the data; and is

Determining the equivalent fuel factor based on the at least one characteristic data.

9. The vehicle energy management method of claim 8,

the at least one item of feature data comprises the following feature data within a predetermined time period: maximum velocity, average velocity, standard deviation of velocity, zero velocity time scale, average positive acceleration, and average negative acceleration.

10. The vehicle energy management method according to any one of claims 7-9, further comprising:

adaptively training the machine learning model further based on historical driving data of the vehicle, the historical driving data including data relating to a moving state of the vehicle recorded in past driving of the vehicle and an equivalent fuel factor calculated from fuel consumption of the past driving.

11. The vehicle energy management method according to any one of claims 7-9, further comprising:

receiving the data at a predetermined frequency for a predetermined period of time; and

updating the determined equivalent fuel factor at the predetermined frequency based on the data for the predetermined period of time.

12. The vehicle energy management method according to any one of claims 7-9, further comprising:

predicting a travel speed and/or route of the vehicle; and

adjusting the equivalent fuel factor determined by the determination unit based on the predicted travel speed and/or route.

13. A vehicle energy management apparatus comprising:

a memory having computer readable instructions stored thereon; and

a processor that, when executed by the processor, causes the processor to perform the vehicle energy management method of any of claims 7-12.

14. A computer readable medium having computer readable instructions stored thereon, which, when executed by a processor, cause the processor to perform the method of any one of claims 7-12.

Technical Field

The present disclosure relates to vehicles and, more particularly, to managing vehicle energy.

Background

Currently, hybrid vehicles using new energy have been widely developed and applied for the purpose of environmental protection and energy saving, including Hybrid Electric Vehicles (HEV), Fuel Cell Electric Vehicles (FCEV), Fuel Cell Hybrid Electric Vehicles (FCHEV), and the like. The energy management strategy is a key technology for the hybrid electric vehicle, and plays an important role in saving cost and increasing endurance mileage.

One energy management strategy that has been proposed is the equivalent fuel consumption minimization strategy (ECMS), of which determining the optimal equivalent fuel factor is a key technology by which fuel consumption minimization in hybrid vehicles can be achieved. Currently, the equivalent fuel factor is determined based on an identification of the vehicle driving condition. The driving conditions may include: urban driving, rural driving, high-speed driving and the like. After the driving conditions are identified, based on the determined current driving conditions, the equivalent fuel factor corresponding to the current driving conditions is found out by utilizing a lookup table in which various driving conditions and corresponding equivalent fuel factors are stored.

In the actual driving process, the working condition of the vehicle is complex, and the corresponding equivalent fuel factor is determined by looking up a table only according to the preset driving working condition, so that the requirement of the complex working condition can be difficult to meet. For example, while city driving involves complex driving situations, it is typically assigned only one specific equivalent fuel factor. If the requirements of complex working conditions are met, the storage resources occupied by the lookup table are considerable under the condition of various driving working conditions. This is particularly problematic for on-board systems that have limited storage capacity.

Accordingly, it is desirable to provide improved management of vehicle energy.

Disclosure of Invention

Improved vehicle energy management systems and methods are provided that enable a direct derivation of an equivalent fuel factor from data relating to the state of motion of the vehicle without prior knowledge of the vehicle driving conditions.

According to one aspect of the present invention, a vehicle energy management system is provided. The system comprises: a receiving unit for receiving data relating to a motion state of a vehicle; and a determination unit for determining an equivalent fuel factor based on the data using a machine learning model; wherein the machine learning model is trained based on known data relating to a state of motion of the vehicle and a corresponding equivalent fuel factor.

In accordance with another aspect of the present invention, a vehicle energy management method is provided. The method comprises the following steps: receiving data relating to a state of motion of a vehicle; and determining an equivalent fuel factor based on the data using a machine learning model; wherein the machine learning model is trained based on known data relating to a state of motion of the vehicle and a corresponding equivalent fuel factor.

According to another aspect of the present invention, a vehicle energy management apparatus is provided. The apparatus comprises: a memory having computer readable instructions stored thereon; and a processor, which when executed by the processor, causes the processor to perform the method according to various embodiments of the invention.

According to another aspect of the present invention, there is provided a computer storage medium having computer-readable instructions stored thereon, which, when executed by a processor, cause the processor to perform a method according to various embodiments of the present invention.

According to various embodiments of aspects of the present invention, it is possible to directly derive the equivalent fuel factor from data related to the motion state of the vehicle without knowing the driving condition of the vehicle in advance. Therefore, a lookup table comprising specific vehicle driving conditions and corresponding equivalent fuel factors does not need to be stored, thereby saving the storage space of the system. Furthermore, determining the equivalent fuel factor based on data relating to the motion state of the vehicle enables the determined equivalent fuel factor to be more suitable for the current actual driving situation, thereby improving the accuracy of determining the equivalent fuel factor.

Still further advantages of the present invention will be appreciated to those of ordinary skill in the art upon reading and understand the following detailed description.

Drawings

FIG. 1 shows a block diagram of a vehicle energy management system according to one embodiment of the invention.

FIG. 2 illustrates an exemplary neural network model for determining an equivalent fuel factor, according to one embodiment of the invention.

FIG. 3 shows a flow diagram of a vehicle energy management method according to one embodiment of the invention.

Detailed Description

FIG. 1 shows a block diagram of a vehicle energy management system 100 according to one embodiment of the invention. The system 100 comprises at least a receiving unit 101 and a determining unit 102. The receiving unit 101 receives data related to the motion state of the vehicle, in particular data related to the speed, acceleration and zero speed time of the vehicle. The determination unit 102 determines the equivalent fuel factor based on the data received by the reception unit 10 using the machine learning model 10 as shown in fig. 2. The machine learning model 10 is, for example, a forward neural network regression model.

FIG. 2 illustrates an exemplary machine learning model 10 that receives data D relating to a motion state of a vehicle, according to one embodiment of the invention_inAs a model input, and outputs an equivalent fuel factor E. Thus, it will be appreciated that the machine learning model 10 is trained by machine learning methods using a large amount of known data relating to the state of motion of the vehicle and the corresponding optimal equivalent fuel factor. Those skilled in the art will appreciate that the hyper-parameters and the implicit elements of the model (e.g., number of iterations, expected error, etc.) may be adjusted prior to training to minimize over-fitting and under-fitting to meet the demands on model accuracy.

For example, for a particular configuration of vehicle, the equivalent fuel factor at which fuel consumption is minimal can be determined for particular data relating to the state of motion of the vehicle as the known optimal equivalent fuel factor. Finally, the machine learning model 10 is trained using a large amount of specific data with corresponding known optimal equivalent fuel factors.

The above-mentioned data relating to the motion state of the vehicle can be obtained by respective sensors 20, the sensors 20 including, but not limited to: a speed sensor, an acceleration sensor, and a timer.

It can be appreciated that the sensor 20 can be included in the system 100 as part of the system 100. Alternatively, the sensor 20 may be separate from the components of the system 100, for example, the sensor 20 may be an in-vehicle sensor, or the sensor 20 may be integrated on a smart mobile device. The sensor 20, which is independent of the system 100, is able to transmit the obtained data to the receiving unit 101 of the system 100 using a communication module.

In a preferred embodiment, the determination unit 102 first derives at least one feature data based on data related to the motion state of the vehicle, e.g. derives a feature data related to a statistic of velocity based on the received velocity data, and derives a feature data related to a statistic of acceleration based on the received acceleration data, and then determines the equivalent fuel factor based on the at least one feature data using a machine learning model.

The inventors of the present invention have discovered that preferred characteristic data that can be received as input by the machine learning model 10 to accurately determine the equivalent fuel factor is included over a predetermined period of time, such as the following characteristic data over a driving cycle: maximum velocity V_MAXAverage velocity V_avgStandard deviation of velocity V_stdZero velocity-time ratio Idle, average forward acceleration a⁺ _avgAnd average negative acceleration a^- _avg. Wherein the maximum speed V_MAXIs determined as the speed at which the vehicle is traveling fastest throughout the driving cycle, the average speed V_avgIs determined as the average of the vehicle's travel speed, the standard deviation V of the speed, over the entire driving cycle_stdIs determined as the standard deviation of the vehicle's travel speed over the entire driving cycle, and the zero speed time proportion Idle is determined as the proportion of the time the vehicle is stopped (e.g., due to a red waiting light) over the entire driving cycle, the average forward acceleration a⁺ _avgIs determined to be greater than 0.1m/s throughout the driving cycle²Average value of acceleration values of (a), and average negative acceleration a^- _avgIs determined to be less than 0.1m/s throughout the driving cycle²Is measured.

In other embodiments, one skilled in the art will appreciate that one or more of the above features may be reduced, depending on different requirements for determination accuracy; or other features may be added such as the standard deviation of acceleration or the proportion of time that the speed is within a particular range, etc.

In one embodiment, at least one of the characteristic data is available to a processing unit other than the system 100, and the receiving unit 101 directly receives at least one of the characteristic data as data relating to the motion state of the vehicle.

After one embodiment, the machine learning model 10 may be trained in real-time continuously during use of the vehicle. For example, the machine learning model 10 can be further trained based on historical driving data of the vehicle. For example, in a certain historical driving, the vehicle consumes less fuel, so a better equivalent fuel factor can be calculated from the fuel consumption. The machine learning model 10 can be further adaptively trained by using the data related to the motion state of the vehicle recorded in the driving and the calculated better equivalent fuel factor, so that the accuracy of the machine learning model 10 can be gradually improved along with the advancement of the service time of the vehicle.

During the running of the vehicle, the receiving unit 101 of the system 100 may receive data relating to the state of motion of the vehicle for a predetermined period of time, for example from the sensors 20; in response to receiving the data for the predetermined period of time, the determination unit 102 updates the equivalent fuel factor based on the received data for the predetermined period of time. In a preferred embodiment, the receiving unit receives the data for the predetermined period of time at a predetermined frequency, for example, the predetermined frequency may be 1 second and the predetermined period of time may be 1 minute. The receiving unit thus receives data relating to the state of motion of the vehicle within 1 minute before it every second for use by the determining unit. The determination unit updates the equivalent fuel factor at a corresponding predetermined frequency based on the data for the predetermined period of time received from the receiving unit.

The method and the device are different from the prior art that the equivalent fuel factor is determined according to a driving condition, and the equivalent fuel factor can be updated in real time. For example, even if the vehicle is used in such a condition that the vehicle is always driven in a city, the equivalent fuel factor can be updated in real time at a predetermined frequency and adapted to the current changing conditions of the vehicle according to the present invention. This enables a finer adjustment of the equivalent fuel factor than would be the case if the equivalent fuel factor were updated only when the driving conditions were changed, thereby enabling improved management of vehicle energy consumption.

One skilled in the art will appreciate that the predetermined frequency and/or the predetermined time period may be set manually. Alternatively, the predetermined frequency and/or the predetermined time period may be automatically adjusted based on data sensed by the sensor 20. For example, the predetermined frequency may be adjusted lower in the case where the speed sensor data and/or the acceleration sensor data indicate that the driving condition of the vehicle has changed little over a long period of time. Whereas in the case of, for example, complex city driving, frequent changes in vehicle speed and/or acceleration may occur, the predetermined frequency may be adjusted higher and the predetermined time period set relatively short to allow the determined equivalent fuel factor to accommodate frequent changes.

In one embodiment, as shown in fig. 1, the system 100 further comprises a prediction unit 103 for predicting a next driving speed and/or route of the vehicle based on data related to the driving of the vehicle.

In one simplest example, data related to vehicle driving may include navigation data for guiding driving. For example, the prediction unit 103 may receive navigation data set by a user on a car navigator and predict the next travel speed and/or route of the vehicle according to the destination position indicated by the navigation data and the current traffic condition.

Alternatively, the data relating to vehicle driving may include data collected by the sensors 20, or feature data derived from data collected by the sensors 20 as noted above. The prediction unit 103 can predict the next travel speed and/or route of the vehicle based on the sensor data or the characteristic data described above. For example, for a specific vehicle, it can exhibit similar characteristic data on the same route that is frequently traveled, and therefore, the prediction unit 103 can predict, based on the specific characteristic data, a route that the vehicle will travel frequently at a frequently used travel speed next.

In this embodiment, the system 100 further comprises an adjustment unit 104. The adjustment unit 104 is able to adjust or correct the equivalent fuel factor determined by the determination unit 102 based on the predicted travel speed and/or route.

In the case where the determined equivalent fuel factor does not significantly correspond to the predicted driving situation, the adjusting unit 104 may adjust the determined equivalent fuel factor into a range that corresponds to the predicted driving situation. This is because the predicted travel speed and/or course of the prediction unit 103 can actually reflect the next driving condition of the vehicle. Although the model of the invention can determine the equivalent fuel factor under the condition of not knowing the driving condition, under the condition that the determined equivalent fuel factor has deviation with the predicted working condition, the predicted working condition can be used as a basis for adjusting the equivalent fuel factor, so that the accuracy of the finally determined equivalent fuel factor is maximized.

Those skilled in the art will appreciate that although the vehicle energy management system is described with reference to the receiving unit 101, the determining unit 102, the predicting unit 103, and the adjusting unit 104 described above, these units are merely illustrative and not restrictive, and they may be combined/divided/partially combined to achieve the corresponding functions.

In addition, one or more of the receiving unit 101, the determining unit 102, the predicting unit 103, and the adjusting unit 104 may also be implemented as computer readable instructions stored on a computer readable medium. Alternatively, corresponding computer readable instructions can be stored in a memory, and the functions of the above-described respective units can be implemented by a processor executing the corresponding instructions. The vehicle energy management device may be constituted by such a memory and a processor. The vehicle energy management device may be part of a vehicle control system.

Those skilled in the art will appreciate that the system 100 according to one or more embodiments may be integrated within a vehicle to form a portion of an on-board system. Or the system 100 according to one or more embodiments may be a stand-alone system capable of communicating with an in-vehicle system through, for example, a data interface.

It is also conceivable that the functions of the units of the vehicle energy management system are implemented at a server side, and the server receives data from the vehicle, determines the corresponding equivalent fuel factor, and issues the equivalent fuel factor to the corresponding vehicle. A server may monitor multiple vehicles simultaneously.

FIG. 3 shows a flow diagram of a vehicle energy management method 300 according to one embodiment of the invention.

At 310, data relating to a motion state of the vehicle is obtained using one or more sensors.

At 320, data relating to a motion state of the vehicle is received.

At 330, an equivalent fuel factor is determined based on the data using a machine learning model, wherein the machine learning model is trained based on known data related to a state of motion of the vehicle and a corresponding equivalent fuel factor.

At 340, the data is received at a predetermined frequency for a predetermined period of time.

At 350, the determined equivalent fuel factor is updated at the predetermined frequency based on the data for the predetermined period of time.

It is understood that the method according to the present application has the same or similar embodiments as the system according to the present application.

The processing of the method shown in fig. 3 can be implemented by a processor executing corresponding instructions. The instructions can be stored on any suitable computer readable medium.

While the method of the present invention has been described above only with reference to the embodiment shown in fig. 3, it is to be understood that the operations included in the above embodiment are not restrictive, and may be deleted, combined, changed, split and/or recombined as necessary to add/modify/delete the corresponding functions.

The systems and methods of the present invention have been described above with reference to various embodiments, which may include a particular feature, structure, or characteristic, but not every embodiment necessarily includes the particular feature, structure, or characteristic. In addition, some embodiments may have some, all, or none of the features described for other embodiments.

Various features of different embodiments or examples may be combined in various ways with some features included and others excluded to accommodate various different applications. The figures and the foregoing description give examples of embodiments. Those skilled in the art will appreciate that one or more of the described elements may be combined into a single functional element. Alternatively, some elements may be divided into multiple functional elements. Elements from one embodiment may be added to another embodiment. For example, the order of the processes described herein may be changed and is not limited to the manner described herein. Moreover, the operations of any flow diagram need not be implemented in the order shown; nor does it necessarily require all operations to be performed. Further, those operations that are not dependent on other operations may be performed in parallel with the other operations. The scope of the embodiments is in no way limited by these specific examples. Many variations, such as differences in the order of operations, product compositions, and structures, are possible, whether or not explicitly set forth in the specification.