阅读说明：本技术 基于神经网络的mems传感器的检测方法、装置、设备及介质 (MEMS sensor detection method, device, equipment and medium based on neural network ) 是由 王小平 曹万 熊波 于 2021-11-29 设计创作，主要内容包括：本发明实施例公开了一种基于神经网络的MEMS传感器的检测方法、装置、设备及介质,该方法包括：接收预置的MEMS传感器的检测请求；根据预置的第一生成式对抗网络生成多组所述MEMS传感器的第一输入信号；将多组所述第一输入信号输入至所述MEMS传感器中,得到每组所述第一输入信号的第一输出信号；将每组所述第一输出信号输入至预置的第二生成式对抗网络中,得到每组所述第一输出信号的特征图；将所述特征图输入至预置的VGG神经网络中,得到所述MEMS传感器的检测结果。本发明基于神经网络技术,不仅提升了MEMS传感器检测的可靠性和准确性,同时还提高了对MEMS传感器的多样性。(The embodiment of the invention discloses a detection method, a device, equipment and a medium of an MEMS sensor based on a neural network, wherein the method comprises the following steps: receiving a detection request of a preset MEMS sensor; generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network; inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals; inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals; and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor. The invention is based on the neural network technology, thereby not only improving the reliability and the accuracy of the detection of the MEMS sensor, but also improving the diversity of the MEMS sensor.)

1. A detection method of a MEMS sensor based on a neural network is characterized by comprising the following steps:

receiving a detection request of a preset MEMS sensor;

generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network;

inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals;

inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals;

and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

2. The method for detecting a neural network-based MEMS sensor according to claim 1, further comprising, before receiving a detection request of a preset MEMS sensor:

acquiring multiple groups of second output signals output by a preset normal MEMS sensor and a preset abnormal MEMS sensor;

and training the first generative confrontation network according to the second output signal to obtain the trained first generative confrontation network.

3. The method for detecting a MEMS sensor based on a neural network of claim 2, wherein before acquiring the plurality of sets of second output signals outputted from the preset normal MEMS sensor and the preset abnormal MEMS sensor, the method further comprises:

inputting multiple preset groups of pressure signals into any MEMS sensor to obtain multiple groups of first signals;

and screening the multiple groups of first signals according to a preset screening rule to obtain multiple groups of screening signals input into the normal MEMS sensor and the abnormal MEMS sensor.

4. The method of claim 2, wherein training the first generative confrontation network according to the second output signal to obtain the trained first generative confrontation network comprises:

preprocessing the second output signal to obtain a preprocessed second output signal;

and training the first generative confrontation network according to the preprocessed second output signal to obtain the trained first generative confrontation network.

5. The method for detecting a neural network-based MEMS sensor according to claim 4, wherein preprocessing the second output signal to obtain a preprocessed second output signal comprises:

acquiring multiple groups of second signals output by the normal MEMS sensor and multiple groups of second signals output by the abnormal MEMS sensor;

combining multiple groups of second signals output by the normal MEMS sensor according to a preset first combination rule to obtain a combined first signal;

and combining multiple groups of second signals output by the abnormal MEMS sensor according to a preset second combination rule to obtain combined second signals.

6. The method of claim 2, wherein the training of the first generative confrontation network according to the second output signal further comprises:

acquiring a plurality of groups of second input signals output during the training of the first generative confrontation network;

inputting a plurality of groups of second input signals into any MEMS sensor to obtain a plurality of groups of third output signals;

and training the second generative confrontation network according to the plurality of groups of third output signals to obtain the trained second generative confrontation network.

7. The method as claimed in claim 6, wherein the training of the second generative countermeasure network according to the plurality of sets of the third output signals, and after obtaining the trained second generative countermeasure network, further comprises:

and training the VGG neural network according to the characteristic diagram output during the training of the second generation type confrontation network to obtain the trained VGG neural network.

8. A detection device of a MEMS sensor based on a neural network is characterized by comprising:

the receiving unit is used for receiving a detection request of a preset MEMS sensor;

the first generation unit is used for generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network;

the first input unit is used for inputting a plurality of groups of first input signals into the MEMS sensor to obtain a first output signal of each group of first input signals;

the second input unit is used for inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals;

and the third input unit is used for inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of detection of a neural network-based MEMS sensor according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, causes the processor to execute the neural network-based MEMS sensor detection method according to any one of claims 1 to 7.

Technical Field

The invention belongs to the technical field of neural networks, and particularly relates to a detection method, a detection device, detection equipment and a detection medium of an MEMS sensor based on a neural network.

Background

The MEMS (Micro-Electro-Mechanical System) sensor is a novel sensor manufactured by using Micro-electronics and Micro-machining technologies, and compared with the conventional sensor, the MEMS sensor has the characteristics of small volume, light weight, low cost, low power consumption, high reliability, easiness in integration and mass production, and the like, and meanwhile, because the MEMS sensor has a characteristic dimension of micron order, the MEMS sensor can complete functions which cannot be realized by some conventional sensors.

In the prior art, when detecting whether an MEMS sensor is abnormal, a single input signal is usually used to obtain an output signal, and then the output signal and a value of an expected output signal are subjected to error analysis, and if the error is within an acceptable range, it is determined that the MEMS sensor is not abnormal, but when detecting whether the MEMS sensor is abnormal by using the method, the output value of the MEMS sensor may be within an error allowable range due to the influence of environmental factors, and further, the accuracy is not high when detecting the abnormality of the MEMS sensor.

Disclosure of Invention

The embodiment of the invention provides a detection method, a detection device, detection equipment and a detection medium of an MEMS sensor based on a neural network, and aims to solve the problem of low accuracy in the process of carrying out abnormity detection on the MEMS sensor in the prior art.

In a first aspect, an embodiment of the present invention provides a detection method for a neural network-based MEMS sensor, which includes:

receiving a detection request of a preset MEMS sensor;

generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network;

inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals;

inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals;

and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

In a second aspect, an embodiment of the present invention provides a detection apparatus for a MEMS sensor based on a neural network, which includes:

the receiving unit is used for receiving a detection request of a preset MEMS sensor;

the first generation unit is used for generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network;

the first input unit is used for inputting a plurality of groups of first input signals into the MEMS sensor to obtain a first output signal of each group of first input signals;

the second input unit is used for inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals;

and the third input unit is used for inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

In a third aspect, an embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the method for detecting a neural network-based MEMS sensor according to the first aspect.

In a fourth aspect, the present invention further provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and the computer program, when executed by a processor, causes the processor to execute the method for detecting a neural network-based MEMS sensor according to the first aspect.

The MEMS sensor detection method based on the neural network adopts the trained generative confrontation network to generate signals with larger output difference of the MEMS sensor in advance, then the signals are input into the MEMS sensor to be detected, the signals output by the MEMS sensor to be detected are input into the other trained generative confrontation network to obtain the corresponding characteristic diagram, and then the characteristic diagram is classified and identified to realize the classification of the MEMS sensor to be detected, so that the reliability and the accuracy of the MEMS sensor detection are improved, and meanwhile, the diversity of the MEMS sensor is also improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic flow chart of a detection method of a neural network-based MEMS sensor according to an embodiment of the present invention.

Fig. 2 is another schematic flow chart of a detection method of a neural network-based MEMS sensor according to an embodiment of the present invention.

Fig. 3 is another schematic flow chart of a detection method of a neural network-based MEMS sensor according to an embodiment of the present invention.

Fig. 4 is a sub-flow diagram of a detection method of a MEMS sensor based on a neural network according to an embodiment of the present invention.

Fig. 5 is another sub-flow diagram of a detection method of a MEMS sensor based on a neural network according to an embodiment of the present invention.

Fig. 6 is another schematic flow chart of a detection method of a neural network-based MEMS sensor according to an embodiment of the present invention.

Fig. 7 is another schematic flow chart of a detection method of a neural network-based MEMS sensor according to an embodiment of the present invention.

Fig. 8 is a schematic block diagram of units of a detection device of a neural network-based MEMS sensor according to an embodiment of the present invention.

FIG. 9 is a schematic block diagram of a computer device provided by an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

Referring to fig. 1, fig. 1 is a schematic flow chart of a detection method of a MEMS (Micro-Electro-Mechanical System) sensor based on a neural network according to an embodiment of the present invention. The detection method of the MEMS sensor based on the neural network is applied to terminal equipment, and is executed through application software installed in the terminal equipment. The terminal device is a terminal device with an internet access function, such as a desktop computer, a notebook computer, a tablet computer or a mobile phone.

The detection method of the MEMS sensor based on the neural network will be described in detail below. As shown in FIG. 1, the method includes the following steps S110 to S150.

And S110, receiving a detection request of a preset MEMS sensor.

Specifically, the detection request is instruction information that the terminal device receives an abnormality detection request for the MEMS sensor, and the terminal device can execute the step of detecting whether the MEMS sensor is in an abnormal state after receiving the detection request.

In other embodiments of the present invention, as shown in fig. 2, step S110 further includes: s210 and S220.

S210, acquiring multiple groups of second output signals output by the preset normal MEMS sensor and the preset abnormal MEMS sensor.

In this embodiment, the normal MEMS sensor is a sensor with normal signal output, the abnormal MEMS sensor is a sensor with abnormal signal output, and because the difference between the signal output by the normal MEMS sensor and the signal output by the abnormal MEMS sensor is large, the training of the generative countermeasure network can be completed quickly by inputting two signals with large difference into the generative countermeasure network.

In other embodiments of the present invention, as shown in fig. 3, step S210 further includes, before: s310 and S320.

S310, inputting multiple preset groups of pressure signals into any MEMS sensor to obtain multiple groups of first signals.

In this embodiment, any MEMS sensor is a normal MEMS sensor, multiple sets of pressure signals received by the MEMS sensor are generated by a pressure generating device, and multiple sets of pressure signals generated by the pressure generating device are input into any normal MEMS sensor, so that multiple sets of first signals output by the MEMS sensor can be obtained, where after the multiple sets of first signals are input into the normal MEMS sensor and the abnormal MEMS sensor, there are both signals with large differences and signals with small differences.

S320, screening the multiple groups of first signals according to a preset screening rule to obtain multiple groups of screening signals input into the normal MEMS sensor and the abnormal MEMS sensor.

Specifically, the screening rule is rule information for screening a plurality of groups of the first signals so that the normal MEMS sensor and the abnormal MEMS sensor output signals with a large difference. In this embodiment, when multiple groups of first signals are screened, all the first signals are input to the abnormal MEMS sensor and the normal MEMS sensor respectively, and then, the differences of the outputs of the same first signal in the abnormal MEMS sensor and the normal MEMS sensor are compared to determine whether the differences reach a preset threshold, and if the differences reach the preset threshold, the first signal can be retained, and if the differences do not reach the preset threshold, the first signal is rejected.

S220, training the first generative confrontation network according to the second output signal to obtain the trained first generative confrontation network.

The Generative confrontation network (GAN) is a deep learning model, and the Generative confrontation network generates output through mutual game learning of a generator and a discriminator in the network, that is, the generator and the discriminator are simultaneously enhanced through mutual competition in a training process. The generator models the joint probability and describes the generation of data according to the distribution condition of the data from the statistical angle, the convergence speed is high, the generator comprises naive Bayes, GDA, HMM and other models, the discriminator models the condition probability P (Y | X), and the discriminator mainly searches for the optimal classification surface among different categories, including LR, SVM and other models. In this embodiment, when the second output signal trains the first generative confrontation network, the second output signal is first input into the generator in the first generative confrontation network, then the signal generated by the generator is discriminated according to the discriminator in the first generative confrontation network, and finally the parameter of the discriminator is updated according to the loss function of the discriminator, and meanwhile, the generator is updated according to the loss function of the generator until the generative confrontation network converges, that is, a back propagation algorithm is adopted, and the minimum loss function is used to optimize the parameters in the network.

In other inventive embodiments, as shown in FIG. 4, step S220 includes sub-steps S221 and S222.

S221, preprocessing the second output signal to obtain a preprocessed second output signal.

In this embodiment, the preprocessing of the second output signal includes correspondingly summarizing the second output signal, that is, summarizing the signals output by the normal MEMS sensors into one group, summarizing the signals output by the abnormal MEMS sensors into another group, and then performing corresponding combination to obtain the preprocessed second output signal. Since the frame of preprocessing the second output signal is constructed based on the tensrflow (the tensrflow is a symbolic mathematical system based on data flow programming), the second output signal needs to be converted into a 32-bit floating point type signal and then input into the corresponding MEMS sensor.

In other inventive embodiments, as shown in fig. 5, step S221 includes sub-steps S2211, S2212, and S2213.

S2211, acquiring multiple groups of second signals output by the normal MEMS sensor and multiple groups of second signals output by the abnormal MEMS sensor;

s2212, combining multiple groups of second signals output by the normal MEMS sensor according to a preset first combination rule to obtain a combined first signal;

and S2213, combining the plurality of groups of second signals output by the abnormal MEMS sensor according to a preset second combination rule to obtain a combined second signal.

In this embodiment, the first combination rule is rule information for combining any second signal of the plurality of sets of second signals output by the normal MEMS sensor, and the second combination rule is rule information for combining the plurality of sets of second signals output by the abnormal MEMS sensor, where the plurality of sets of second signals output by the normal MEMS sensor and the plurality of sets of second signals output by the abnormal MEMS sensor need to keep the sizes of the signals consistent during the combination.

S222, training the first generative confrontation network according to the preprocessed second output signal to obtain the trained first generative confrontation network.

In other embodiments of the present invention, as shown in fig. 6, after step S220, steps S230, S240, and S250 are further included.

And S230, acquiring multiple groups of second input signals output during the training of the first generative confrontation network.

S240, inputting the multiple groups of second input signals into any MEMS sensor to obtain multiple groups of third output signals.

And S250, training the second generative confrontation network according to the plurality of groups of third output signals to obtain the trained second generative confrontation network.

In this embodiment, after the plurality of sets of the second input signals are generated by the first generative confrontation network during the training process, and the plurality of sets of the third output signals are generated by the plurality of sets of the second input signals passing through any one of the normal MEMS sensors, the sensors output the data sets, and the data sets are used for training the second generative confrontation network. Wherein the second generative countermeasure network is in the same principle as the first generative countermeasure network, and is different from the first generative countermeasure network in that the first generative countermeasure network generates a plurality of sets of the second input signals using random noise data, and the second generative countermeasure network is a profile obtained by passing output values of normal MEMS sensors using data generated by the first generative countermeasure network.

In other embodiments of the present invention, as shown in fig. 6, after step S250, step S260 is further included.

And S260, training the VGG neural network (the VGG neural network is named after the VGG neural network is provided by a Visual Geometry Group of Oxford university) according to the feature diagram output during the training of the second generation type confrontation network, and obtaining the trained VGG neural network.

In this embodiment, the data used in the VGG neural network training is a feature map generated in the network training process by the second generating equation. Because the difference between each input signal in the training process of the second generation type antagonistic network is large, the difference between the output signals in the training process of the second generation type antagonistic network is also large, so that the VGG neural network model can be converged quickly, and the training time of the VGG neural network model is further saved.

And S120, generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generation countermeasure network.

Specifically, the first input signal is used for being input into the MEMS sensor to perform abnormality detection on the MEMS sensor, the first generative countermeasure network is a network trained in advance and capable of generating an input signal of the MEMS sensor, in the process of detecting the MEMS sensor, an output signal of the first generative countermeasure network model includes a signal output by the abnormal MEMS sensor and also includes a signal output by the normal MEMS sensor, the two types of signals are simultaneously input into the MEMS sensor to be detected, and then a characteristic diagram is generated according to the signal output by the MEMS sensor to be detected and classified and identified, so that whether the MEMS sensor to be detected is abnormal or not can be identified.

S130, inputting the multiple groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals.

Specifically, the first output signal is a signal output by the MEMS sensor after the first input signal is input into the MEMS sensor, and whether the MEMS sensor is in an abnormal state can be detected by performing corresponding detection classification on the first output signal.

And S140, inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals.

Specifically, the generative countermeasure network is a feature map for generating the first output signal output by the MEMS sensor, and then the feature map is classified and identified, so that whether the MEMS sensor is in an abnormal state can be detected.

S150, inputting the characteristic diagram into a preset VGG neural network to obtain a detection result of the MEMS sensor.

The VGG neural network is a deep convolution neural network, convolution kernels in the VGG neural network are small-size convolution kernels and continuous small-size convolution kernels, the continuous small-size convolution kernels are separated by a pooling layer, after deep convolution is carried out on the feature map by the VGG neural network and pooling operation is carried out, the classification result of the feature map can be output through a full connection layer in the VGG neural network. In this embodiment, a VGG19 neural network is used to classify and identify the feature map, the VGG19 neural network includes 19 hidden layers, wherein 16 convolutional layers and 3 fully-connected layers are included in the 19 hidden layers, and the VGG19 neural network has the characteristic of higher stability and higher accuracy in a plurality of neural networks, and can be combined with a generative countermeasure network to effectively improve the stability and reliability of detection of the MEMS sensor.

In the detection method of the MEMS sensor based on the neural network provided by the embodiment of the invention, a preset detection request of the MEMS sensor is received; generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network; inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals; inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals; and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor. The MEMS sensor can output signals with large difference by adopting the trained generative countermeasure network in advance, then the signals are input into the MEMS sensor to be detected, the signals output by the MEMS sensor to be detected are input into the other trained generative countermeasure network to obtain a corresponding characteristic diagram, and then the characteristic diagram is classified and recognized to realize the classification of the MEMS sensor to be detected, so that the reliability and the accuracy of the detection of the MEMS sensor are improved, and meanwhile, the diversity of the MEMS sensor is also improved.

The embodiment of the invention also provides a detection device 100 of the MEMS sensor based on the neural network, which is used for executing any embodiment of the detection method of the MEMS sensor based on the neural network. Specifically, referring to fig. 8, fig. 8 is a schematic block diagram of a detection apparatus 100 of a MEMS sensor based on a neural network according to an embodiment of the present invention.

As shown in fig. 8, the detection apparatus 100 for the neural network-based MEMS sensor includes a receiving unit 110, a first generating unit 120, a first input unit 130, a second input unit 140, and a third input unit 150.

A receiving unit 110 for receiving a detection request of a preset MEMS sensor;

a first generating unit 120, configured to generate a plurality of groups of first input signals of the MEMS sensors according to a preset first generative countermeasure network.

A first input unit 130, configured to input multiple sets of the first input signals into the MEMS sensor, so as to obtain a first output signal of each set of the first input signals.

The second input unit 140 is configured to input each group of the first output signals into a preset second generative countermeasure network, so as to obtain a characteristic diagram of each group of the first output signals.

And a third input unit 150, configured to input the feature map into a preset VGG neural network, so as to obtain a detection result of the MEMS sensor.

In another embodiment, the detecting apparatus 100 for a neural network based MEMS sensor further includes: the device comprises a first acquisition unit and a first training unit.

And the first acquisition unit is used for acquiring multiple groups of second output signals output by the preset normal MEMS sensor and the preset abnormal MEMS sensor.

And the first training unit is used for training the first generative confrontation network according to the second output signal to obtain the trained first generative confrontation network.

In another embodiment, the detecting apparatus 100 for a neural network based MEMS sensor further includes: a fourth input unit and a screening unit.

And the fourth input unit is used for inputting a plurality of preset groups of pressure signals into any MEMS sensor to obtain a plurality of groups of first signals.

And the screening unit is used for screening the multiple groups of first signals according to a preset screening rule to obtain multiple groups of screening signals input into the normal MEMS sensor and the abnormal MEMS sensor.

In another embodiment, the first training unit comprises: a preprocessing unit and a second training unit.

The preprocessing unit is used for preprocessing the second output signal to obtain a preprocessed second output signal; and the second training unit is used for training the first generative confrontation network according to the preprocessed second output signal to obtain the trained first generative confrontation network.

In another embodiment, the preprocessing unit includes: the device comprises a second acquisition unit, a first combination unit and a second combination unit.

The second acquisition unit is used for acquiring a plurality of groups of second signals output by the normal MEMS sensor and a plurality of groups of second signals output by the abnormal MEMS sensor; the first combination unit is used for combining a plurality of groups of second signals output by the normal MEMS sensor according to a preset first combination rule to obtain a combined first signal; and the second combination unit is used for combining a plurality of groups of second signals output by the abnormal MEMS sensor according to a preset second combination rule to obtain a combined second signal.

In another embodiment, the detecting apparatus 100 for a neural network based MEMS sensor further includes: a third acquisition unit, a fifth input unit and a third training unit.

And the third acquisition unit is used for acquiring a plurality of groups of second input signals output during the training of the first generative confrontation network.

And the fifth input unit is used for inputting the multiple groups of second input signals into any MEMS sensor to obtain multiple groups of third output signals.

And the third training unit is used for training the second generative confrontation network according to the plurality of groups of third output signals to obtain the trained second generative confrontation network.

In another embodiment, the detecting apparatus 100 for a neural network based MEMS sensor further includes: and a fourth training unit.

And the fourth training unit is used for training the VGG neural network according to the feature diagram output during the training of the second generation type confrontation network to obtain the trained VGG neural network.

The detection device 100 of the MEMS sensor based on the neural network according to the embodiment of the present invention is configured to execute the above-mentioned detection request for receiving the preset MEMS sensor; generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network; inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals; inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals; and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

Referring to fig. 9, fig. 9 is a schematic block diagram of a computer device according to an embodiment of the present invention.

Referring to fig. 9, the device 500 includes a processor 502, memory, and a network interface 505 connected by a system bus 501, where the memory may include a non-volatile storage medium 503 and an internal memory 504.

The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032, when executed, may cause the processor 502 to perform a neural network based detection method for a MEMS sensor.

The processor 502 is used to provide computing and control capabilities that support the operation of the overall device 500.

The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503, and when the computer program 5032 is executed by the processor 502, the processor 502 may be enabled to execute the neural network-based MEMS sensor detection method.

The network interface 505 is used for network communication, such as providing transmission of data information. Those skilled in the art will appreciate that the configuration shown in fig. 9 is a block diagram of only a portion of the configuration associated with aspects of the present invention and does not constitute a limitation of the apparatus 500 to which aspects of the present invention may be applied, and that a particular apparatus 500 may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

Wherein the processor 502 is configured to run the computer program 5032 stored in the memory to implement the following functions: receiving a detection request of a preset MEMS sensor; generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network; inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals; inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals; and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

Those skilled in the art will appreciate that the embodiment of the apparatus 500 shown in fig. 9 does not constitute a limitation on the specific construction of the apparatus 500, and in other embodiments, the apparatus 500 may include more or fewer components than shown, or some components may be combined, or a different arrangement of components. For example, in some embodiments, the apparatus 500 may only include the memory and the processor 502, and in such embodiments, the structure and function of the memory and the processor 502 are the same as those of the embodiment shown in fig. 9, and are not repeated herein.

It should be understood that in the present embodiment, the Processor 502 may be a Central Processing Unit (CPU), and the Processor 502 may also be other general-purpose processors 502, a Digital Signal Processor 502 (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, discrete hardware component, etc. The general-purpose processor 502 may be a microprocessor 502 or the processor 502 may be any conventional processor 502 or the like.

In another embodiment of the present invention, a computer storage medium is provided. The storage medium may be a non-volatile computer-readable storage medium. The storage medium stores a computer program 5032, wherein the computer program 5032 when executed by the processor 502 performs the steps of: receiving a detection request of a preset MEMS sensor; generating a plurality of groups of first input signals of the MEMS sensor according to a preset first generative countermeasure network; inputting a plurality of groups of first input signals into the MEMS sensor to obtain first output signals of each group of first input signals; inputting each group of first output signals into a preset second generative countermeasure network to obtain a characteristic diagram of each group of first output signals; and inputting the characteristic diagram into a preset VGG neural network to obtain the detection result of the MEMS sensor.

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described apparatuses, devices and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described in a functional general in the foregoing description for the purpose of illustrating clearly the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

In the embodiments provided by the present invention, it should be understood that the disclosed apparatus, device and method can be implemented in other ways. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only a logical division, and there may be other divisions when the actual implementation is performed, or units having the same function may be grouped into one unit, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may also be an electric, mechanical or other form of connection.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment of the present invention.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a storage medium. Based on such understanding, the technical solution of the present invention essentially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product stored in a storage medium and including instructions for causing a device 500 (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disk.

While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.