An electric vehicle dynamic working condition electromagnetic anti-interference and function stability evaluation method based on deep learning and entropy algorithm

Abstract

The application provides an electric vehicle dynamic working condition electromagnetic anti-interference and function stability evaluation method based on deep learning and entropy algorithm, a controllable change electromagnetic field is generated on an integrated test platform where an electric vehicle is located, the electric vehicle runs under the influence of the electromagnetic field, visual collection is performed on the electric vehicle, dynamic working condition images of a plurality of target components are obtained, a controllable and changeable electromagnetic interference environment is provided for testing, and dynamic and continuous visual monitoring of different target components is realized; then, the dynamic working condition images are recognized, dynamic operation state information of the target components is obtained, time domain curves of the operation states of the target components are obtained, the operation states of the target components with time under electromagnetic interference are continuously identified, data basis for subsequent deep learning processing is provided; and then, the operation state time domain curves are processed based on an entropy algorithm, time correlation stability evaluation results of the target components are obtained, the electromagnetic anti-interference degree information of the target components is determined, the electromagnetic anti-interference of the electric vehicle is integrated, dynamic and comprehensive evaluation is performed, and the effectiveness and reliability of the electromagnetic compatibility test of the electric vehicle are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the field of electric vehicles, in particular to an electric vehicle dynamic working condition electromagnetic anti-interference and function stability evaluation method based on deep learning and an entropy algorithm. BACKGROUND

[0002] Electric vehicles play an increasingly important role in modern transportation. With the popularity of electric vehicles, the electromagnetic compatibility problem of electric vehicles has become increasingly prominent. Electric vehicles contain a large number of electronic components, which generate electromagnetic radiation when working, which may cause electromagnetic interference to other electronic systems of the electric vehicle, the external environment and other vehicles. Therefore, improving the electromagnetic anti-interference performance of electric vehicles to ensure their reliable operation in complex electromagnetic environments is the key to ensuring traffic safety and normal operation of electric vehicles. The electromagnetic compatibility test of existing electric vehicles is usually carried out in a laboratory simulation environment, which lacks testing means under real working conditions, and the corresponding calculation work of electromagnetic compatibility test is still limited to the data processing level of a conventional computer, which cannot conduct integrated, dynamic and comprehensive evaluation of the electromagnetic anti-interference of electric vehicles, cannot promote the unified standardization of the electromagnetic compatibility test technology of electric vehicles, and reduces the effectiveness and reliability of the electromagnetic compatibility test of electric vehicles. SUMMARY

[0003] The purpose of the present application is to provide an electric vehicle dynamic working condition electromagnetic anti-interference and function stability evaluation method based on deep learning and an entropy algorithm, which generates a controllable variable electromagnetic field on an integrated test platform where the electric vehicle is located, so that the electric vehicle operates under the influence of the electromagnetic field, and the electric vehicle is visually collected to obtain dynamic working condition images of each target component, to provide a controllable and variable electromagnetic interference environment for testing and to realize dynamic and continuous visual monitoring of different target components. The dynamic working condition images are then identified to obtain dynamic operation state information of the target components, from which operation state time domain curves of the target components are obtained, and the operation state of the target components over time under electromagnetic interference is continuously identified to provide a data basis for subsequent deep learning processing. The operation state time domain curves are also processed based on an entropy algorithm to obtain time-related stability evaluation results of the target components, which are used to determine the electromagnetic anti-interference degree information of the target components, to conduct integrated, dynamic and comprehensive evaluation of the electromagnetic anti-interference of the electric vehicle, and to improve the effectiveness and reliability of the electromagnetic compatibility test of the electric vehicle.

[0004] The present application is achieved by the following technical solutions:

[0005] An electric vehicle dynamic working condition electromagnetic anti-interference and function stability evaluation method based on deep learning and an entropy algorithm, comprising:

[0006] A controllable variable electromagnetic field is generated on an integrated test platform where an electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field; visual collection is performed on the electric vehicle in a running state, so as to obtain dynamic working condition images of each target component of the electric vehicle;

[0007] The dynamic working condition images are identified, so as to obtain dynamic operation state information of the target component; based on the dynamic operation state information, an operation state time domain curve of the target component is obtained.

[0008] Based on the operation state time domain curve, an entropy value algorithm processing is performed, so as to obtain a time correlation stability evaluation result of the target component; based on the time correlation stability evaluation result, electromagnetic anti-interference degree information of the target component is determined.

[0009] Optionally, the controllable variable electromagnetic field is generated on the integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field, and the method comprises the following steps.

[0010] A frequency and intensity controllable variable electromagnetic field is generated to different types of antennas inside the integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field covering.

[0011] The electromagnetic field intensity data inside the integrated test platform are also collected by a field strength probe inside the integrated test platform, and all types of antennas are feedback adjusted based on the electromagnetic field intensity data.

[0012] Optionally, the integrated test platform is further provided with a movable antenna, and the movable antenna can be used for auxiliary adjustment, which comprises the following steps.

[0013] Step A1, the movable antenna is controlled to perform auxiliary adjustment according to the electromagnetic field intensity data inside the integrated test platform by using the following formula (1),

[0014]

[0015] In the above formula (1), P represents the enabling control value of the movable antenna for auxiliary adjustment; {X, Y, Z} represents the position coordinate set of the electromagnetic field strength collected at the field strength probe distribution position inside the integrated test platform that does not satisfy the preset electromagnetic field strength; E(x, y, z) represents the electromagnetic field strength value collected by the field strength probe at the (x, y, z) coordinate point position inside the integrated test platform; E0 represents the preset electromagnetic field strength threshold; {(x, y, z) | [E(x, y, z) - E0]≤0} represents the (x, y, z) coordinate point position corresponding to the formula [E(x, y, z) - E0]≤0 selected from all the field strength probe distribution positions inside the integrated test platform; Q{X, Y, Z} represents the total number of coordinate points in the coordinate set {X, Y, Z}; N represents the total number of all the field strength probe distribution positions inside the integrated test platform; L({X, Y, Z}2) represents the distance value obtained by taking two from the coordinate points in the coordinate set {X, Y, Z}; F{L({X, Y, Z}2)≤L0} represents the total number of the distance values obtained by taking two from the coordinate points in the coordinate set {X, Y, Z} that are less than L0; and represents the logical and, and outputs the value 1 if the formulas on both sides of and are both true, and outputs the value 0 otherwise.

[0016] If P = 1, the movable antenna for auxiliary adjustment is enabled to be turned on.

[0017] If P = 0, the movable antenna for auxiliary adjustment is enabled to be turned off.

[0018] Step A2, according to the distribution of the electromagnetic field strength data inside the integrated test platform, the position of the movable antenna for auxiliary adjustment is controlled by using the following formula (2),

[0019]

[0020] In the above formula (2), (x0, y0, z0) represents the position coordinate point of the movable antenna for auxiliary adjustment; (L({X, Y, Z}2)≤L0)(a) represents the a-th coordinate point that satisfies the distance value obtained by taking two from the coordinate points in the coordinate set {X, Y, Z} that is less than L0; represents the average of the horizontal coordinate, the vertical coordinate, and the vertical coordinate of all the coordinate points that satisfy the distance value obtained by taking two from the coordinate points in the coordinate set {X, Y, Z} that is less than L0;

[0021] Step A3, according to the position of the movable antenna for auxiliary adjustment and the state of the electromagnetic field strength data around the position, the generated electromagnetic field strength of the movable antenna is controlled by using the following formula (3),

[0022]

[0023] In the above formula (3), E D represents the generated electromagnetic field strength of the movable antenna; l0 represents a unit distance value; e0 represents the minimum generated electromagnetic field strength of the antenna at the coverage distance; Lmax({X,Y,Z}2) represents the maximum value of the distance values obtained by taking two coordinates in the coordinate set {X,Y,Z}.

[0024] Optionally, the electric vehicle in the running state is visually collected to obtain dynamic working condition images of each target component of the electric vehicle, including:

[0025] The wiper area, the light area, the airbag and the steering wheel area of the electric vehicle in the running state are independently visually collected to obtain dynamic working condition images of the wiper, the turn signal, the low beam, the high beam, the airbag and the steering wheel of the electric vehicle.

[0026] Optionally, the wiper area, the light area, the airbag and the steering wheel area of the electric vehicle in the running state are independently visually collected to obtain dynamic working condition images of the wiper, the turn signal, the low beam, the high beam, the airbag and the steering wheel of the electric vehicle, including:

[0027] The wiper area, the airbag and the steering wheel area are binocularly visually collected, and the light area is high-speed visually collected to obtain corresponding binocular dynamic images of the wiper area, the airbag and the steering wheel area, and high-speed dynamic images of the light area.

[0028] The binocular dynamic images of the wiper area, the binocular dynamic images of the airbag and the steering wheel area, and the high-speed dynamic images of the light area are respectively denoised and segmented for preprocessing to extract dynamic working condition images of the wiper, the turn signal, the low beam, the high beam, the airbag and the steering wheel of the electric vehicle.

[0029] Optionally, the dynamic working condition images are identified to obtain dynamic operation state information of the target components, including:

[0030] The dynamic working condition images are target detected and identified to obtain swing state information of the wiper, flicker state information of the turn signal, on-off action state information of the low beam and the high beam, ejection action state information of the airbag, and rotation action state information of the steering wheel.

[0031] Optionally, based on the dynamic operation state information, operation state time domain curves of the target components are obtained, including:

[0032] The wobble state information of the wiper, the flicker state information of the turn signal, the switch action state information of the low beam and the high beam, the pop action state information of the airbag, and the rotation action state information of the steering wheel are respectively subjected to convolutional neural network processing to obtain a time-domain curve of the wiper wobble angle changing over time, a time-domain curve of the turn signal flicker switching state changing over time, a time-domain curve of the low beam and the high beam switch action switching state changing over time, a time-domain curve of the airbag pop action amplitude changing over time, and a time-domain curve of the steering wheel rotation angle changing over time.

[0033] Optionally, the time-dependent stability evaluation result of the target component is obtained by performing an entropy value algorithm processing based on the operation state time-domain curve, including:

[0034] Based on the sample entropy algorithm and the multi-scale entropy algorithm, the operation state time-domain curve is analyzed to obtain the stability evaluation result of the target component at different time scales;

[0035] Based on the Renyi entropy algorithm and the waveform entropy algorithm, the operation state time-domain curve is analyzed to obtain the time-domain stability evaluation result of the target component;

[0036] The stability evaluation results at different time scales and the time-domain stability evaluation result are integrated to obtain the time-dependent stability evaluation result of the target component; wherein the time-dependent stability evaluation result includes the stability evaluation result corresponding to each of the target component at different lengths of working duration and different lengths of cumulative working time.

[0037] Optionally, based on the time-dependent stability evaluation result, the electromagnetic anti-interference degree information of the target component is determined, including:

[0038] Based on the time-dependent stability evaluation result, the maximum time length information corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies is determined; based on the maximum time length information, the electromagnetic anti-interference degree information of the target component is determined; wherein the electromagnetic anti-interference program information includes the electromagnetic interference intensity information that the target component can resist.

[0039] Optionally, the maximum time length information is determined, including:

[0040] Step S1, assuming that the electromagnetic field intensity of the target electromagnetic field is I, and the frequency of the electromagnetic field is F, then the rate of change of the energy density of the target electromagnetic field over time is:

[0041] E = IF (4)

[0042] In the above formula (4), E is the rate of change of the energy density of the target electromagnetic field with time;

[0043] Step S2, set the stability threshold of the target component as T, then the stability coefficient of the target component in the target electromagnetic field is:

[0044]

[0045] In the above formula (5), f T is the stability coefficient of the target component in the target electromagnetic field;

[0046] Step S3, according to the calculation results of the above steps S1 and S2, determine the maximum time length information corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies,

[0047]

[0048] In the above formula (6), H is the maximum time length corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies, R is the interval time of the target component responding to the target electromagnetic interference and starting to show unstable signs, T D is the rate of decay of the stability of the target component with time, which is related to the material of the target component, and e is a natural constant.

[0049] Compared with the prior art, the present application has the following beneficial effects:

[0050] The electric vehicle dynamic working condition electromagnetic anti-interference and function stability evaluation method based on deep learning and entropy algorithm provided in the present application generates a controllable variable electromagnetic field on an integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field, and the electric vehicle is visually collected to obtain dynamic working condition images of each target component, thereby providing a controllable and variable electromagnetic interference environment for testing and realizing dynamic and continuous visual monitoring of different target components. Then, the dynamic working condition images are identified to obtain dynamic operation state information of the target component, so as to obtain an operation state time domain curve of the target component, and the operation state of the target component with time under electromagnetic interference is continuously identified to provide a data basis for subsequent deep learning processing. Furthermore, the operation state time domain curve is processed based on an entropy algorithm to obtain a time-related stability evaluation result of the target component, so as to determine electromagnetic anti-interference degree information of the target component, integrate the electromagnetic anti-interference of the electric vehicle, dynamically and comprehensively evaluate the electromagnetic anti-interference of the electric vehicle, and improve the effectiveness and reliability of the electromagnetic compatibility test of the electric vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only represent some of the embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative effort based on these drawings. Among them:

[0052] Figure 1 A flowchart of an electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method based on deep learning and entropy algorithm provided by the present application.

[0053] Figure 2 A structural block diagram of an integrated test platform used by an electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method based on deep learning and entropy algorithm provided by the present application. DETAILED DESCRIPTION

[0054] In order to make the above objectives, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings. It can be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. In addition, it should be noted that, for the purpose of description, only the parts related to the present application are shown in the drawings, but not all the structures. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0055] The terms "comprising" and "having" and any variations thereof in the present application are intended to cover non-exclusive inclusion. For example, a process, method, method, product or device including a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0056] In this paper, the "embodiment" means that the specific features, structures or characteristics described in conjunction with the embodiment can be included in at least one embodiment of the present application. The phrase appears at various places in the specification does not necessarily refer to the same embodiment, nor is it independent or alternative to other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0057] Please refer to Figure 1As shown, an embodiment of the present application provides an electromagnetic anti-interference and function stability evaluation method for dynamic working conditions of an electric vehicle based on deep learning and an entropy algorithm. The electromagnetic anti-interference and function stability evaluation method for dynamic working conditions of an electric vehicle based on deep learning and an entropy algorithm comprises:

[0058] A controllable and changeable electromagnetic field is generated on an integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field; visual collection is performed on the electric vehicle in a running state to obtain dynamic working condition images of a plurality of target components of the electric vehicle;

[0059] The dynamic working condition images are recognized to obtain dynamic operation state information of the target components; and based on the dynamic operation state information, operation state time-domain curves of the target components are obtained.

[0060] Based on the operation state time-domain curves, entropy algorithm processing is performed to obtain time-related stability evaluation results of the target components; and based on the time-related stability evaluation results, electromagnetic anti-interference degree information of the target components is determined.

[0061] The above embodiment has the beneficial effect that the electromagnetic anti-interference and function stability evaluation method for dynamic working conditions of an electric vehicle based on deep learning and an entropy algorithm generates a controllable and changeable electromagnetic field on an integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field, visual collection is performed on the electric vehicle to obtain dynamic working condition images of a plurality of target components of the electric vehicle, a controllable and changeable electromagnetic interference environment is provided for testing, and dynamic and continuous visual monitoring of different target components is realized; then the dynamic working condition images are recognized to obtain dynamic operation state information of the target components, so as to obtain operation state time-domain curves of the target components, and the operation state of the target components over time under electromagnetic interference is continuously identified, thereby providing a data basis for subsequent deep learning processing; further, entropy algorithm processing is performed based on the operation state time-domain curves to obtain time-related stability evaluation results of the target components, so as to determine electromagnetic anti-interference degree information of the target components, and integrated, dynamic and comprehensive evaluation of electromagnetic anti-interference of the electric vehicle is performed, thereby improving the effectiveness and reliability of electromagnetic compatibility testing of the electric vehicle.

[0062] In another embodiment, a controllable and changeable electromagnetic field is generated on an integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field, comprising:

[0063] A controllable and changeable electromagnetic field is generated on an integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field, comprising:

[0064] The electromagnetic field intensity data inside the integrated test platform is collected by the field strength probe inside the integrated test platform, and all types of antennas are feedback adjusted based on the electromagnetic field intensity data.

[0065] The above embodiment has the beneficial effect that the electromagnetic field generated by the electronic components inside the electric vehicle during actual driving is not fixed, but varies in terms of electromagnetic field intensity and frequency, and the circuit structures of different target components inside the electric vehicle are not the same, and the anti-interference performance of different target components to the electromagnetic field is also not the same. In order to comprehensively and effectively test the electromagnetic anti-interference of all target components inside the electric vehicle, different types of antennas such as double-cone antennas, logarithmic-periodic antennas, and horn antennas are arranged in the semi-anechoic chamber of the integrated test platform. The above different types of antennas can generate different distributed electromagnetic fields, and all antennas are also connected with the signal generator, power meter, switching unit and power amplifier in the control room of the integrated test platform. In this way, all antennas can be controlled individually, so as to independently change the frequency and intensity of the electromagnetic field generated by each antenna. In addition, field strength probes are arranged inside the semi-anechoic chamber for detecting the electromagnetic field intensity inside the semi-anechoic chamber, and the field strength probes are also connected with the field strength meter and the host computer in the control room. The host computer is also connected with the signal generator, so as to form a closed-loop control of the antennas. The electromagnetic field intensity data inside the integrated test platform is collected by the field strength probe, and all types of antennas are feedback adjusted based on the electromagnetic field intensity data, so as to improve the accuracy of antenna adjustment. In addition, a chassis dynamometer and a camera are arranged in the semi-anechoic chamber for chassis testing and visual acquisition of the electric vehicle, and the chassis dynamometer and the camera are connected with the host computer through the industrial computer and the controller respectively. The structure of the integrated test platform is shown in Figure 2 , which will not be described in detail here.

[0066] In another embodiment, a movable antenna is also arranged inside the integrated test platform, which can be used for auxiliary adjustment, including:

[0067] Step A1, according to the electromagnetic field intensity data inside the integrated test platform, the movable antenna is controlled to enable auxiliary adjustment by using the following formula (1),

[0068]

[0069] In the above formula (1), P represents the enabling control value of the movable antenna for auxiliary adjustment; {X, Y, Z} represents the position coordinate set of the electromagnetic field strength collected at the field strength probe distribution position inside the integrated test platform that does not satisfy the preset electromagnetic field strength; E(x, y, z) represents the electromagnetic field strength value collected by the field strength probe at the (x, y, z) coordinate point position inside the integrated test platform; E0 represents the preset electromagnetic field strength threshold; {(x, y, z) | [E(x, y, z) - E0]≤0} represents the (x, y, z) coordinate point position corresponding to the formula [E(x, y, z) - E0]≤0 selected from all the field strength probe distribution positions inside the integrated test platform; Q{X, Y, Z} represents the total number of coordinate points in the coordinate set {X, Y, Z}; N represents the total number of all the field strength probe distribution positions inside the integrated test platform; L({X, Y, Z}2) represents the distance value obtained by taking two from the coordinate points in the coordinate set {X, Y, Z}; F{L({X, Y, Z}2)≤L0} represents the total number of the distance values obtained by taking two from the coordinate points in the coordinate set {X, Y, Z} that are less than L0; and represents the logical and, and outputs the value 1 if the formulas on both sides of and are both true, and outputs the value 0 otherwise.

[0070] If P = 1, the movable antenna for auxiliary adjustment is enabled to be turned on.

[0071] If P = 0, the movable antenna for auxiliary adjustment is enabled to be turned off.

[0072] Step A2, according to the distribution of the electromagnetic field strength data inside the integrated test platform, the position of the movable antenna for auxiliary adjustment is controlled by using the following formula (2),

[0073]

[0074] In the above formula (2), (x0, y0, z0) represents the position coordinate point of the movable antenna for auxiliary adjustment; (L({X, Y, Z}2)≤L0)(a) represents the a-th coordinate point that satisfies the distance value obtained by taking two from the coordinate points in the coordinate set {X, Y, Z} that is less than L0; represents the average of the horizontal coordinate, the vertical coordinate, and the vertical coordinate of all the coordinate points that satisfy the distance value obtained by taking two from the coordinate points in the coordinate set {X, Y, Z} that is less than L0;

[0075] Step A3, according to the position of the movable antenna for auxiliary adjustment and the state of the electromagnetic field strength data around the position, the generated electromagnetic field strength of the movable antenna is controlled by using the following formula (3),

[0076]

[0077] In the above formula (3), E D represents the generated electromagnetic field strength of the movable antenna; l0 represents a unit distance value; e0 represents the minimum generated electromagnetic field strength of the antenna at the coverage distance; and Lmax({X,Y,Z}2) represents the maximum value of the distance values obtained by taking two coordinates in the coordinate set {X,Y,Z}.

[0078] The above embodiment has the following beneficial effects. According to the electromagnetic field strength data inside the integrated test platform, the movable antenna is controlled to enable auxiliary adjustment by using the above formula (1), thereby ensuring the reliability of the auxiliary adjustment of the movable antenna. According to the distribution of the electromagnetic field strength data inside the integrated test platform, the position of the movable antenna for auxiliary adjustment is controlled by using the above formula (2), thereby automatically calculating the position of the movable antenna for automatic movement, which embodies the intelligent feature of the system. According to the position of the movable antenna for auxiliary adjustment and the state of the electromagnetic field strength data around the position, the generated electromagnetic field strength of the movable antenna is controlled by using the above formula (3), thereby automatically and intelligently adjusting the electromagnetic field strength inside the integrated test platform by using the movable antenna, which can improve the adjustment efficiency of the system.

[0079] In another embodiment, a running electric vehicle is visually collected to obtain dynamic working condition images of each target component of the electric vehicle, including:

[0080] The wiper area, the light area, the airbag and the steering wheel area of the running electric vehicle are independently visually collected to obtain dynamic working condition images of the wiper, the steering light, the low beam light, the high beam light, the airbag and the steering wheel of the electric vehicle.

[0081] The beneficial effects of the above embodiments, the wiper, turn signal, low beam, high beam, airbag, steering wheel and other target components of the electric vehicle have complex circuit structure inside, so that these target components will inevitably be disturbed by electromagnetic field during operation and cause work disorder, such as wiper cannot swing at uniform speed, turn signal cannot normally flicker, low beam and high beam cannot normally open or close, airbag cannot pop out in time, steering wheel cannot normally rotate and turn direction, etc. In order to obtain the dynamic working conditions of all target components comprehensively and accurately during electromagnetic anti-interference test of electric vehicle, a plurality of cameras are arranged to perform real-time visual collection on different areas of electric vehicle, such as different cameras can be arranged in front of electric vehicle to perform independent visual shooting on wiper area and lamp area of electric vehicle, and a camera is arranged above the driver's seat inside the electric vehicle to aim at the airbag and steering wheel area, so that continuous and complete image recording of the action working condition of the corresponding target component during the whole electromagnetic anti-interference test can be ensured, and reliable image data support is provided for subsequent analysis of whether abnormal conditions caused by electromagnetic interference occur in the target component.

[0082] In another embodiment, the wiper area, lamp area, airbag and steering wheel area of the electric vehicle in operation are independently visually collected, and the dynamic working condition images of the wiper, turn signal, low beam, high beam, airbag and steering wheel of the electric vehicle are obtained, including:

[0083] The wiper area, the airbag and the steering wheel area are respectively collected by binocular vision, and the high-speed vision of the lamp area is collected, and the corresponding binocular dynamic image of the wiper area, the binocular dynamic image of the airbag and the steering wheel area, and the high-speed dynamic image of the lamp area are obtained.

[0084] The binocular dynamic image of the wiper area, the binocular dynamic image of the airbag and the steering wheel area, and the high-speed dynamic image of the lamp area are respectively denoised and segmented for preprocessing, and the dynamic working condition images of the wiper, turn signal, low beam, high beam, airbag and steering wheel of the electric vehicle are extracted.

[0085] The beneficial effects of the above embodiment are that the operating actions of the wiper, airbag and steering wheel all occur in three-dimensional space. In order to accurately collect the dynamic working conditions of the wiper, airbag and steering wheel, a binocular camera can be used to collect the wiper area, the airbag and the steering wheel area respectively by binocular vision to obtain the corresponding wiper area binocular dynamic image, airbag and steering wheel area binocular dynamic image. The turn signal, low beam and high beam all work by emitting light of a specific shape. In order to accurately capture the light emitted by the turn signal, low beam and high beam, a high-speed industrial camera can be used to collect the high-speed vision of the vehicle lamp area to obtain the high-speed dynamic image of the vehicle lamp area, which can improve the reliability of the visual collection of the dynamic working condition. In addition, the wiper area binocular dynamic image, the airbag and the steering wheel area binocular dynamic image, and the high-speed dynamic image of the vehicle lamp area are respectively denoised and segmented for preprocessing, and the dynamic working condition images of the wiper, turn signal, low beam, high beam, airbag, steering wheel of the electric vehicle are extracted, so that each dynamic working condition image corresponds to only one target component, effectively improving the reliability of the dynamic working condition image and reducing the noise interference of the image.

[0086] In another embodiment, the dynamic working condition image is identified to obtain dynamic operation state information of the target component, including:

[0087] The dynamic working condition image is detected and identified to obtain the swing state information of the wiper, the flashing state information of the turn signal, the on-off action state information of the low beam and the high beam, the ejection action state information of the airbag, and the rotation action state information of the steering wheel.

[0088] The beneficial effects of the above embodiment are that in actual image processing, the corresponding deep learning neural network model can be used to detect and identify the corresponding target component of the dynamic working condition image to obtain the swing state information of the wiper (such as the swing angular velocity and swing amplitude of the wiper), the flashing state information of the turn signal (such as the flashing frequency of the turn signal), the on-off action state information of the low beam and the high beam (such as the opening illumination brightness and on-off response speed of the low beam and the high beam), the ejection action state information of the airbag (such as the ejection response speed and ejection coverage of the airbag), and the rotation action state information of the steering wheel (such as the rotation smoothness of the steering wheel), which can accurately identify the dynamic working condition of all target components with respect to time, and provide a reliable basis for subsequent judgment of whether the target component fails due to electromagnetic interference.

[0089] In another embodiment, based on the dynamic operation state information, an operation state time domain curve of the target component is obtained, including:

[0090] The wobble state information of the wiper, the flicker state information of the turn signal, the switch action state information of the low beam and the high beam, the deployment action state information of the airbag, and the rotation action state information of the steering wheel are respectively subjected to convolutional neural network processing to obtain a time-domain curve of the wobble angle of the wiper changing over time, a time-domain curve of the flicker switching state of the turn signal changing over time, a time-domain curve of the switch action switching state of the low beam and the high beam changing over time, a time-domain curve of the deployment action amplitude of the airbag changing over time, and a time-domain curve of the rotation angle of the steering wheel changing over time.

[0091] The wobble state information of the wiper, the flicker state information of the turn signal, the switch action state information of the low beam and the high beam, the deployment action state information of the airbag, and the rotation action state information of the steering wheel are respectively subjected to convolutional neural network processing to obtain a time-domain curve of the wobble angle of the wiper changing over time, a time-domain curve of the flicker switching state of the turn signal changing over time, a time-domain curve of the switch action switching state of the low beam and the high beam changing over time, a time-domain curve of the deployment action amplitude of the airbag changing over time, and a time-domain curve of the rotation angle of the steering wheel changing over time.

[0092] In another embodiment, the operation state time-domain curve is subjected to entropy value algorithm processing to obtain a time-related stability evaluation result of the target component, including:

[0093] Based on the sample entropy algorithm and the multi-scale entropy algorithm, the operation state time-domain curve is analyzed to obtain a stability evaluation result of the target component at different time scales;

[0094] Based on the Renyi entropy algorithm and the waveform entropy algorithm, the operation state time-domain curve is analyzed to obtain a time-domain stability evaluation result of the target component;

[0095] The stability evaluation result at different time scales and the time-domain stability evaluation result are integrated to obtain a time-related stability evaluation result of the target component; wherein the time-related stability evaluation result includes a stability evaluation result corresponding to each of different lengths of working duration and different lengths of cumulative working time of the target component.

[0096] The beneficial effects of the above embodiments are that, in order to refine the analysis of the correlation between the operation state of the target component and electromagnetic interference, the operation state time domain curve is analyzed based on a sample entropy algorithm and a multi-scale entropy algorithm to obtain stability evaluation results of the target component at different time scales, i.e., corresponding stability evaluation results of the target component at different time lengths during the electromagnetic interference process; and the operation state time domain curve is analyzed based on a Renyi entropy algorithm and a waveform entropy algorithm to obtain a time domain stability evaluation result of the target component, i.e., a stability evaluation result of the target component corresponding to the change in the cumulative working time of the target component itself during the electromagnetic interference process. The sample entropy algorithm, the multi-scale entropy algorithm, the Renyi entropy algorithm, and the waveform entropy algorithm all belong to conventional algorithms in the field, and will not be described in detail here. In addition, the stability evaluation results at different time scales and the time domain stability evaluation result are integrated to obtain stability evaluation results corresponding to different lengths of working duration and different lengths of cumulative working time of the target component, respectively, so that the performance of the target component that can maintain a normal working state under electromagnetic interference can be accurately evaluated.

[0097] In another embodiment, based on the time-dependent stability evaluation result, the electromagnetic interference resistance degree information of the target component is determined, including:

[0098] Based on the time-dependent stability evaluation result, the maximum time length information corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies is determined; based on the maximum time length information, the electromagnetic interference resistance degree information of the target component is determined; wherein the electromagnetic interference resistance program information includes electromagnetic interference intensity information that the target component can resist.

[0099] The beneficial effects of the above embodiments are that, based on the time-dependent stability evaluation result, the maximum time length information corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies is determined, i.e., the maximum continuous duration length corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies is determined; then, based on the maximum time length information, the performance of the target component is comprehensively evaluated to determine the electromagnetic interference intensity information that the target component can resist, thereby quantitatively determining the electromagnetic interference resistance performance of the target component, wherein the comprehensive evaluation belongs to a conventional technical means in the field, which can be realized by a convolutional neural network, and will not be described in detail here.

[0100] In another embodiment, the maximum time length information is determined, including:

[0101] In step S1, the electromagnetic field intensity of the target electromagnetic field is I, the frequency of the electromagnetic field is F, and the rate of change of the energy density of the target electromagnetic field with time is:

[0102] E = IF (4)

[0103] In the above formula (4), E is the rate of change of the energy density of the target electromagnetic field with time;

[0104] Step S2, set the stability threshold of the target component T, then the stability coefficient of the target component in the target electromagnetic field is:

[0105]

[0106] In the above formula (5), f T is the stability coefficient of the target component in the target electromagnetic field;

[0107] Step S3, according to the calculation results of the above steps S1 and S2, determine the maximum time length information corresponding to each of the target component under the action of electromagnetic field with different intensity and frequency to maintain stable working state,

[0108]

[0109] In the above formula (6), H is the maximum time length corresponding to each of the target component under the action of electromagnetic field with different intensity and frequency to maintain stable working state, R is the interval time of the target component responding to the target electromagnetic interference and starting to show unstable signs, T D is the rate of decay of the stability of the target component with time, which is related to the material of the target component, and e is the natural constant.

[0110] The beneficial effects of the above embodiment, the maximum time length information is to measure the maximum time length of the target component maintaining stable work under external electromagnetic interference, and the accuracy of its value is related to the safety of the electric vehicle, so the determination of the maximum time length information, especially the determination of the maximum time length information under the action of electromagnetic field with different intensity and frequency, is very important; According to the related information of the target electromagnetic field and the stability of the target component, the maximum time length of the target component maintaining stable work under the electromagnetic interference of the target electromagnetic field is accurately calculated, which provides important guarantee for the safety of the electric vehicle.

[0111] Overall, the dynamic working condition electromagnetic anti-interference and functional stability evaluation method of electric vehicles based on deep learning and entropy algorithm generates controllable variable electromagnetic fields on the integrated test platform where the electric vehicles are located, so that the electric vehicles run under the influence of the electromagnetic fields, the electric vehicles are visually collected, the dynamic working condition images of the subordinate target components are obtained, a controllable and variable electromagnetic interference environment is provided for the test, and dynamic and continuous visual monitoring of different target components is realized; the dynamic working condition images are identified, the dynamic operation state information of the target components is obtained, the operation state time domain curve of the target components is obtained, the operation state of the target components with time under electromagnetic interference is continuously identified, and a data basis for subsequent deep learning processing is provided; and the operation state time domain curve is processed based on the entropy algorithm, the time-related stability evaluation result of the target components is obtained, the electromagnetic anti-interference degree information of the target components is determined, the electromagnetic anti-interference of the electric vehicle is integrated, dynamically and comprehensively evaluated, and the effectiveness and reliability of the electromagnetic compatibility test of the electric vehicle are improved.

[0112] The above is only one specific embodiment of the present application, and any improvement made on the basis of the concept of the present application is considered to be within the protection scope of the present application.

Claims

1. A method for evaluating the electromagnetic anti-interference and functional stability of an electric vehicle under dynamic conditions based on deep learning and entropy algorithms, characterized in that, The method comprises the following steps: Generating a controllable variable electromagnetic field on an integrated test platform where an electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field; Collecting images of the electric vehicle in operation, to obtain dynamic working condition images of each target component of the electric vehicle; Identifying the dynamic working condition images to obtain dynamic operation state information of the target component, and obtaining an operation state time-domain curve of the target component based on the dynamic operation state information; Processing the operation state time-domain curve based on an entropy value algorithm to obtain a time-dependent stability evaluation result of the target component, and determining electromagnetic anti-interference degree information of the target component based on the time-dependent stability evaluation result; Generating a controllable variable electromagnetic field on an integrated test platform where an electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field, comprising: Generating an electromagnetic field with individually controllable frequency and intensity to different types of antennas inside the integrated test platform where the electric vehicle is located, so that the electric vehicle runs under the influence of the electromagnetic field to cover the corresponding target components; Further collecting electromagnetic field intensity data inside the integrated test platform through field strength probes inside the integrated test platform, and feeding back and adjusting all types of antennas based on the electromagnetic field intensity data; The integrated test platform further comprises a movable antenna, which can be used for auxiliary adjustment, comprising: Step A1, controlling the movable antenna to enable auxiliary adjustment according to the electromagnetic field intensity data inside the integrated test platform by using the following formula (1), In the above formula (1), P represents the enable control value of the movable antenna for auxiliary adjustment; {X, Y, Z} represents a set of position coordinates of the electromagnetic field intensity that does not meet the preset electromagnetic field intensity among the distribution positions of the field strength probes inside the integrated test platform; E(x, y, z) represents the electromagnetic field intensity value collected by the field strength probe at the (x, y, z) coordinate point position inside the integrated test platform; E0 represents the preset electromagnetic field intensity threshold; {(x, y, z)|[E(x, y, z)-E0]≤0} represents the (x, y, z) coordinate point position selected from all the field strength probe distribution positions inside the integrated test platform that satisfies the algorithm [E(x, y, z)-E0]≤0; Q{X, Y, Z} represents the total number of coordinate points in the coordinate set {X, Y, Z}; N represents the total number of all field strength probe distribution positions inside the integrated test platform; L({X, Y, Z}2) represents the distance value obtained by taking two coordinates in the coordinate set {X, Y, Z}; F{L({X, Y, Z}2)≤L0} represents the total number of distance values obtained by taking two coordinates in the coordinate set {X, Y, Z} that are less than L0; and represents logical and, which outputs the value 1 if both sides of the algorithm are true, and outputs the value 0 otherwise; If P=1, the enablement of the movable antenna for auxiliary adjustment is turned on; If P=0, the enablement of the movable antenna for auxiliary adjustment is turned off; Step A2, according to the distribution of electromagnetic field intensity data inside the integrated test platform, the movable antenna is controlled to assist adjustment of the position by using the following formula (2), In the above formula (2), (x0, y0, z0) represents a position coordinate point of the movable antenna for auxiliary adjustment; (L({X, Y, Z}2)≤L0)(a) represents the a-th coordinate point satisfying the condition that the distance value obtained by taking two coordinate points in the coordinate set {X, Y, Z} is smaller than L0; represents the average of the horizontal coordinates, the vertical coordinates and the vertical coordinates of all coordinate points satisfying the condition that the distance value obtained by taking two coordinate points in the coordinate set {X, Y, Z} is smaller than L0. Step A3, according to the position of the movable antenna assisted adjustment and the state of the electromagnetic field intensity data around the position, the electromagnetic field intensity generated by the movable antenna is controlled by using the following formula (3), In the above equation (3), E D represents the generated electromagnetic field strength of the movable antenna; l0represents a unit distance value; e0represents the minimum generated electromagnetic field strength of the antenna at the coverage distance; and Lmax({X,Y,Z}2) represents the maximum value among the distance values obtained by taking two at a time from the coordinate points within the coordinate set {X,Y,Z}.

2. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method of claim 1, wherein: the electric vehicle in a running state is visually collected to obtain dynamic working condition images of each target component of the electric vehicle, including: the wiper area, the light area, the airbag area, and the steering wheel area of the electric vehicle in a running state are independently visually collected to obtain dynamic working condition images of the wiper, the turn signal, the low beam, the high beam, the airbag, and the steering wheel of the electric vehicle.

3. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method of claim 2, wherein: the wiper area, the light area, the airbag area, and the steering wheel area of the electric vehicle in a running state are independently visually collected to obtain dynamic working condition images of the wiper, the turn signal, the low beam, the high beam, the airbag, and the steering wheel of the electric vehicle, including: the wiper area, the airbag area, and the steering wheel area are binocularly visually collected, and the light area is high-speed visually collected to obtain corresponding binocular dynamic images of the wiper area, the airbag area, and the steering wheel area, and high-speed dynamic images of the light area; the binocular dynamic images of the wiper area, the airbag area, and the steering wheel area, and the high-speed dynamic images of the light area are respectively denoised and segmented for preprocessing to extract the dynamic working condition images of the wiper, the turn signal, the low beam, the high beam, the airbag, and the steering wheel of the electric vehicle.

4. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method of claim 3, wherein: the dynamic working condition images are recognized to obtain dynamic operation state information of the target components, including: the dynamic working condition images are target-detected and recognized to obtain swing state information of the wiper, flicker state information of the turn signal, on-off action state information of the low beam and the high beam, ejection action state information of the airbag, and rotation action state information of the steering wheel.

5. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method of claim 4, wherein: based on the dynamic operation state information, operation state time-domain curves of the target components are obtained, including: ​ ​ ​ ​ ​ ​ ​ ​ The wobble state information of the wiper, the flicker state information of the turn signal, the switch action state information of the low beam and the high beam, the pop action state information of the airbag, and the rotation action state information of the steering wheel are respectively subjected to convolutional neural network processing to obtain a time-domain curve of the wiper wobble angle changing over time, a time-domain curve of the turn signal flicker switching state changing over time, a time-domain curve of the low beam and the high beam switch action switching state changing over time, a time-domain curve of the airbag pop action amplitude changing over time, and a time-domain curve of the steering wheel rotation angle changing over time.

6. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method according to claim 1, characterized in that: Based on the operation state time-domain curve, entropy algorithm processing is performed to obtain the time-dependent stability evaluation result of the target component, including: Based on the sample entropy algorithm and the multi-scale entropy algorithm, the operation state time-domain curve is analyzed to obtain the stability evaluation result of the target component at different time scales; Based on the Renyi entropy algorithm and the waveform entropy algorithm, the operation state time-domain curve is analyzed to obtain the time-domain stability evaluation result of the target component; The stability evaluation results at different time scales and the time-domain stability evaluation result are integrated to obtain the time-dependent stability evaluation result of the target component; wherein the time-dependent stability evaluation result includes the stability evaluation result corresponding to each of the target component at different lengths of working duration and different lengths of cumulative working time.

7. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method according to claim 6, characterized in that: Based on the time-dependent stability evaluation result, the electromagnetic anti-interference degree information of the target component is determined, including: Based on the time-dependent stability evaluation result, the maximum time length information corresponding to the stable working state of the target component under the action of electromagnetic fields of different intensities and frequencies is determined; based on the maximum time length information, the electromagnetic anti-interference degree information of the target component is determined; wherein the electromagnetic anti-interference degree information includes the electromagnetic interference intensity information that the target component can resist.

8. The deep learning and entropy algorithm-based electric vehicle dynamic working condition electromagnetic anti-interference and functional stability evaluation method according to claim 7, characterized in that: Determining the maximum time length information includes: Step S1, assuming that the electromagnetic field intensity of the target electromagnetic field is I and the frequency of the electromagnetic field is F, then the energy density of the target electromagnetic field changes over time at a rate of: E = IF (4), In the above formula (4), E is the rate of change of the energy density of the target electromagnetic field over time; Step S2, assuming that the stability threshold of the target component is T, then the stability coefficient of the target component in the target electromagnetic field is: In the above equation (5), f T is a stability coefficient of the target component in the target electromagnetic field; Step S3, according to the calculation results of steps S1 and S2, determine the target component in different intensity and frequency of electromagnetic field under the action of the respective maximum length of time information corresponding to the stable working state, In the above formula (6), H is the maximum length of time for the target component to remain in a stable operating state under the action of electromagnetic fields of different intensities and frequencies, R is the interval of time for the target component to respond to the target electromagnetic interference and begin to show signs of instability, T D is the rate of decay of the stability of the target component over time, which is related to the material of the target component, and e is a natural constant.

Images