Training method and apparatus, vehicle safety function control method and apparatus, and vehicle

Abstract

The present disclosure relates to the field of control, and provides a training method and apparatus, a vehicle safety function control method and apparatus, and a vehicle. The training method comprises: acquiring a signal sample image of a vehicle, wherein the signal sample image comprises a safety function normal image and a safety function failure image; using the signal sample image to train a deep neural network model, and using the trained deep neural network model to perform data augmentation processing on the signal sample image to obtain training sample images; and using the training sample images to train a safety function failure identification classifier, wherein the safety function failure identification classifier is used for identifying a safety function state during vehicle operation, and the safety function state includes a safety function normal state or a safety function failure state.

Description

Training method, vehicle safety function control method, device and vehicle

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is based on the application with CN application number 202410323642.9 and application date March 20, 2024, and claims its priority. The disclosed content of the CN application is hereby introduced as a whole into this application. Technical Field

[0003] The present disclosure relates to the field of control, and in particular to a training method, a vehicle safety function control method, a device, and a vehicle. Background Art

[0004] In recent years, with the emergence and continuous progress of high-tech technologies such as electronic technology, computer technology, and sensor technology, the technical level of construction machinery has continued to improve, and the construction machinery industry has provided solid support for the stability of infrastructure construction. At present, with the development of intelligentization in the construction machinery industry, electronic, electrical, and programmable devices have been widely used. Especially in achieving functional safety, the failure of electronic and electrical systems will lead to failure of control functions, thereby causing accidents and even personal injury. Previously, the European Union has included the ISO 13849 standard for the functional safety of construction machinery into the CE certification, and the ISO 19014 standard will be officially included in the EU EN standard in 2023. This Class C standard provides detailed regulations and systematic evaluations for the functional safety of earth-moving machinery. As people's awareness of functional safety has gradually become stronger, this has also put forward more stringent requirements for the technological development of the construction machinery industry. Therefore, more and more experts and scholars have begun to study the functional safety-related technologies of construction machinery products. Summary of the Invention

[0005] According to one aspect of the present disclosure, a method for training a safety function failure recognition classifier is proposed, comprising: obtaining signal sample images of a vehicle, wherein the signal sample images include a normal safety function image and a failed safety function image; training a deep neural network model using the signal sample images, performing data augmentation processing on the signal sample images using the trained deep neural network model to obtain training sample images; and training a safety function failure recognition classifier using the training sample images, wherein the safety function failure recognition classifier is used to identify the safety function status of the vehicle during operation, wherein the safety function status includes a normal safety function status or a failed safety function status.

[0006] In some embodiments, the vehicle is a loader, the safety function failure image includes a speed safety function failure image and a deceleration-stop safety function failure image, and the safety function status includes a safety function normal state, a speed safety function failure state, or a deceleration-stop safety function failure state.

[0007] In some embodiments, obtaining a signal sample image of a vehicle includes: obtaining data sample signals of the vehicle under normal operation, failure of a speed safety function, and failure of a deceleration-stop safety function, respectively, wherein the data sample signals include an accelerator pedal sample signal, a brake pedal sample signal, and a torque feedback sample signal; and fusing the data sample signals into a signal sample image.

[0008] In some embodiments, acquiring the signal sample image of the vehicle further includes: acquiring the generation time of the data sample signal, wherein converting the data sample signal into the signal sample image includes: generating the signal sample image according to the data sample signal and the generation time.

[0009] In some embodiments, the deep neural network model includes a generative adversarial network model, and training the deep neural network model using signal sample images includes: inputting random noise into the generator of the generative adversarial network model to generate a forged sample image; inputting the signal sample image and the forged sample image into the discriminator of the generative adversarial network model to obtain the sample authenticity; and training the discriminator using a first loss function and training the generator using a second loss function, wherein the value of the first loss function is negatively correlated with the accuracy of the sample authenticity, and the value of the second loss function is positively correlated with the accuracy of the sample authenticity.

[0010] According to another aspect of the present disclosure, a vehicle safety function control method is also proposed, including: acquiring vehicle operation data; fusing the vehicle operation data into a data image; inputting the data image into a safety function failure recognition classifier to obtain the vehicle's safety function status, the safety function status including a safety function normal state or a safety function failure state, wherein the safety function failure recognition classifier is obtained based on the above-mentioned training method; and controlling the vehicle's safety function according to the vehicle's safety function status.

[0011] In some embodiments, the vehicle includes a loader, the vehicle operation data includes an accelerator pedal signal, a brake pedal signal, and a torque feedback signal, and the safety function failure state includes a speed safety function failure state or a deceleration-stop safety function failure state.

[0012] In some embodiments, controlling the safety function of the vehicle according to the safety function status of the vehicle includes: in response to the vehicle being in a speed safety function failure state, controlling the vehicle to enter a safe state and determining whether the vehicle is in a driving state; in response to the vehicle being in a driving state, controlling the speed of the vehicle according to a predetermined accelerator pedal instruction; and in response to the vehicle being in a starting state, controlling the vehicle to stop running.

[0013] In some embodiments, controlling the safety function of the vehicle according to the safety function status of the vehicle includes: in response to the vehicle being in a deceleration-stop safety function failure state, controlling the vehicle to enter a safe state and determining whether the vehicle is in a driving state; in response to the vehicle being in a driving state, controlling the speed of the vehicle according to a predetermined brake pedal instruction; and in response to the vehicle being in a starting state, controlling the vehicle to stop running.

[0014] In some embodiments, controlling the safety function of the vehicle according to the safety function status of the vehicle includes: in response to the vehicle being in a normal safety function status, controlling the vehicle driving state to remain unchanged.

[0015] In some embodiments, a shift switch signal and a direction feedback signal of the vehicle are obtained; based on the shift switch signal and the direction feedback signal, it is identified whether the vehicle is in a state of failure of the reversing safety function; in response to the vehicle being in a state of failure of the reversing safety function, the vehicle is controlled to enter a safe state, and it is determined whether the vehicle is in a driving state; in response to the vehicle being in a driving state, the direction of the vehicle is controlled according to predetermined instructions; and in response to the vehicle being in a starting state, the vehicle is controlled to stop running.

[0016] In some embodiments, a brake switch signal of a vehicle is obtained; based on the brake switch signal and the torque feedback signal, it is identified whether the vehicle is in a brake safety function failure state; in response to the vehicle being in a brake safety function failure state, the vehicle is controlled to enter a safe state, and it is determined whether the vehicle brake is activated; in response to the vehicle brake being activated, the output torque of the vehicle is controlled to be 0; and in response to the vehicle brake not being activated, the output torque of the vehicle is controlled to remain unchanged.

[0017] In some embodiments, an emergency stop switch signal of the vehicle is obtained; based on the emergency stop switch signal, it is identified whether the vehicle is in an emergency stop switch safety function state; and in response to the vehicle being in the emergency stop switch safety function state, the vehicle is controlled to enter a safe state and the vehicle is controlled to stop running.

[0018] According to another aspect of the present disclosure, a training device for a safety function failure identification classifier is proposed, including: a first data acquisition module, configured to acquire signal sample images of a vehicle, wherein the signal sample images include normal safety function images and failed safety function images; a first training module, configured to train a deep neural network model using the signal sample images, so as to perform data augmentation processing on the signal sample images using the trained deep neural network model to obtain training sample images; and a second training module, configured to train a safety function failure identification classifier using the training sample images, wherein the safety function failure identification classifier is used to identify the safety function status of the vehicle during operation, wherein the safety function status includes a normal safety function status or a failed safety function status.

[0019] According to another aspect of the present disclosure, a vehicle safety function control device is also proposed, including: a second data acquisition module, configured to acquire vehicle operation data of the vehicle and fuse the vehicle operation data into a signal image; a state determination module, configured to input the data image into a safety function failure recognition classifier to obtain the safety function status of the vehicle, the safety function status including the normal state of the safety function or the failure state of the safety function, wherein the safety function failure recognition classifier is obtained based on the above training method; and a control module, configured to control the safety function of the vehicle according to the safety function status of the vehicle.

[0020] According to another aspect of the present disclosure, an electronic device is also proposed, including: a memory; and a processor connected to the memory, the processor being configured to execute the above-mentioned training method, or the above-mentioned vehicle safety function control method, based on instructions stored in the memory.

[0021] According to another aspect of the present disclosure, a vehicle is further provided, comprising: the training device for the safety function failure identification classifier, the vehicle safety function control device, and at least one of the electronic devices.

[0022] According to another aspect of the present disclosure, a computer-readable storage medium is further provided, on which computer program instructions are stored. When the instructions are executed by a processor, the training method described above, or the vehicle safety function control method described above, is implemented.

[0023] According to another aspect of the present disclosure, a computer program product is also proposed, including a computer program or instructions, which implements the above-mentioned training method, or the above-mentioned vehicle safety function control method when executed by a processor.

[0024] Other features and advantages of the present disclosure will become apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0026] The present disclosure can be more clearly understood from the following detailed description with reference to the accompanying drawings, in which:

[0027] FIG1 is a flowchart of some embodiments of a method for training a safety function failure identification classifier disclosed herein;

[0028] FIG2 is a block diagram of the implementation principle of a generative adversarial network model according to some embodiments of the present disclosure;

[0029] FIG3 is a flowchart of other embodiments of the method for training a safety function failure identification classifier disclosed herein;

[0030] FIG4 is a flow chart of some embodiments of the vehicle safety function control method disclosed herein;

[0031] FIG5 is a flow chart of other embodiments of the vehicle safety function control method disclosed herein;

[0032] FIG6 is a flow chart of other embodiments of the vehicle safety function control method disclosed herein;

[0033] FIG7 is a flow chart of other embodiments of the vehicle safety function control method disclosed herein;

[0034] FIG8 is a flow chart of other embodiments of the vehicle safety function control method disclosed herein;

[0035] FIG9 is a flow chart of other embodiments of the vehicle safety function control method disclosed herein;

[0036] FIG10 is a schematic structural diagram of some embodiments of a training device for a safety function failure identification classifier disclosed herein;

[0037] FIG11 is a schematic structural diagram of some embodiments of the vehicle safety function control device disclosed herein;

[0038] FIG12 is a schematic structural diagram of some embodiments of the electronic device disclosed herein; and

[0039] FIG13 is a schematic structural diagram of some embodiments of the vehicle disclosed herein. DETAILED DESCRIPTION

[0040] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure.

[0041] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0042] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0043] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0044] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0045] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0046] In order to make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0047] FIG1 is a flow chart of some embodiments of the method for training a safety function failure identification classifier disclosed herein, including steps S11 to S13.

[0048] In step S11 , a signal sample image of a vehicle is acquired, wherein the signal sample image includes a normal safety function image and a disabled safety function image.

[0049] In some embodiments, the vehicle is a loader, which is a type of earth-moving engineering machinery among engineering machinery. It is a self-propelled loading and unloading machinery for earth and stone in construction projects such as roads, railways, buildings, hydropower, ports, and mines.

[0050] In some embodiments, the signal sample images of the loader include: a normal speed and deceleration-stop safety function image, a speed safety function failure image, and a deceleration-stop safety function failure image.

[0051] In some embodiments, data sample signals of the loader are obtained respectively when it is in normal operation, when the speed safety function fails, and when the deceleration-stop safety function fails, wherein the data sample signals include an accelerator pedal sample signal, a brake pedal sample signal, and a torque feedback sample signal; and the data sample signals are fused into a signal sample image.

[0052] For example, a loader is equipped with multiple sensors. Simultaneously, the accelerator pedal sensor collects the accelerator pedal's opening, the brake pedal sensor collects the brake pedal's opening, and the torque feedback sensor collects the motor's output torque. The accelerator pedal, brake pedal, and torque feedback sample signals are fused into a single image to generate a signal sample image.

[0053] In some embodiments, data sample signals of the loader in normal operation, failure of the speed safety function, and failure of the deceleration-stop safety function, as well as the generation time of the data sample signals, are obtained respectively; and a signal sample image is generated based on the data sample signals and the generation time.

[0054] For example, according to the time of signal acquisition, the accelerator pedal sample signal, the brake pedal sample signal and the torque feedback sample signal are fused into one graph to obtain a signal sample image.

[0055] In some embodiments, the signal sample image can be a time domain image or a time-frequency image. The time-frequency image can provide more features, which is conducive to feature extraction and classification by deep convolutional neural networks.

[0056] In this step, the numerical signal is converted into an image to facilitate subsequent model training.

[0057] In some embodiments, after feature extraction is performed on the signal sample images, image classification processing is performed. For example, the signal sample images are classified into images showing normal loader speed and deceleration-stop safety functions, images showing loader speed safety functions failing, and images showing loader deceleration-stop safety functions failing. This step can be performed using a classifier or manually calibrated in advance.

[0058] Those skilled in the art should understand that when the number of safety function items of a loader changes or when it is applied to other construction machinery / earthmoving machinery, the type of sensor signal will also change, and thus the type of sample image after classification processing will also change.

[0059] In step S12, the signal sample image is used to train a deep neural network model, so as to perform data augmentation processing on the signal sample image using the trained deep neural network model to obtain a training sample image.

[0060] In some embodiments, the deep neural network model includes a generative adversarial network model. Training the deep neural network model using the signal sample image includes: inputting random noise into a generator of the generative adversarial network model to generate a forged sample image; inputting the signal sample image and the forged sample image into a discriminator of the generative adversarial network model to obtain sample authenticity; and training the discriminator using a first loss function and training the generator using a second loss function, wherein the value of the first loss function is negatively correlated with the accuracy of the sample authenticity, and the value of the second loss function is positively correlated with the accuracy of the sample authenticity.

[0061] For example, as shown in FIG2 , step S12 is explained. FIG2 is a block diagram of the implementation principle of the generative adversarial network model of some embodiments of the present disclosure.

[0062] In step S121 , a real sample data set, ie, a sample of a real sample, is obtained according to the signals of each sensor.

[0063] In step S122, the generator receives a random noise signal and generates a forged sample. The random noise signal can be uniformly distributed or Gaussian distributed.

[0064] In step S123, the discriminator receives the samples generated by the generator and the samples from the real samples, and determines the authenticity of the samples.

[0065] For real samples, the probability is assigned as high as possible; for forged samples, the probability is assigned as low as possible. The higher the probability, the more likely the sample is real. The generator continuously improves its capabilities, making the generated samples closer and closer to real samples, which means that the discriminator becomes increasingly unable to distinguish whether the samples are real or not. This process is repeated until the discriminator cannot distinguish whether the received samples are real or generated.

[0066] Those skilled in the art should understand that the generative adversarial network model here is only used as an example, and other types of deep neural network models with the same or similar functions as the generative adversarial network model, such as the DCGAN (Deep Convolutional Generative Adversarial Networks) model, can also be used to perform data augmentation processing on sample data.

[0067] In step S13 , a safety function failure recognition classifier is trained using training sample images, wherein the safety function failure recognition classifier is used to recognize the safety function status of the vehicle during operation, where the safety function status includes a normal safety function status or a safety function failure status.

[0068] In some embodiments, the safety function status of the loader includes a speed and deceleration-stop safety function normal state, a speed safety function failure state, or a deceleration-stop safety function failure state.

[0069] In some embodiments, a deep convolutional neural network is used to construct a safety function failure recognition classifier, which is trained using a large number of generated training sample images. The output of the safety function failure recognition classifier is compared with the labeling information of the training sample images to determine whether the comparison result meets the requirements of the loss function for constructing the safety function failure recognition classifier. The parameters of the safety function failure recognition classifier are optimized and adjusted repeatedly to ensure that the comparison result ultimately meets the requirements of the loss function for constructing the safety function failure recognition classifier, and the safety function failure recognition classifier is saved.

[0070] The data images collected when the vehicle is working are input into the safety function failure recognition classifier, so that the safety function status of the vehicle can be obtained, and then the safety function of the vehicle can be controlled according to the safety function status of the vehicle.

[0071] In the above embodiment, a deep neural network model is constructed and trained, and the deep neural network model is used to perform data augmentation processing on the signal sample image to obtain a large number of training sample images, which are then used to train a safety function failure recognition classifier, so that the safety function failure recognition classifier can be used to identify the safety function status of the vehicle, which can improve the accuracy and speed of vehicle safety function status recognition.

[0072] As shown in FIG3 , a vehicle loader is taken as an example to introduce the training method of the safety function failure identification classifier disclosed in the present invention.

[0073] In step S31, the accelerator pedal sample signal, brake pedal sample signal and torque feedback sample signal and the time information of the signal generation are respectively obtained when the loader is in normal operation, when the speed safety function fails and when the deceleration-stop safety function fails.

[0074] In step S32, a signal sample image is generated by fusing the signal and the signal corresponding time.

[0075] In step S33, the signal sample images are classified to obtain a normal image of the loader speed and deceleration-stop safety function, a disabled image of the loader speed safety function, and a disabled image of the loader deceleration-stop safety function.

[0076] In step S341, a first generative adversarial network model is trained using normal images of the loader speed and the deceleration-stop safety function.

[0077] In step S351, a large number of normal training sample images of speed and deceleration-stop safety functions are generated using a first generative adversarial network model.

[0078] In step S342, the second generative adversarial network model is trained using the loader speed safety function failure image.

[0079] In step S352, a large number of speed safety function failure training sample images are generated using the second generative adversarial network model.

[0080] In step S343, the third generative adversarial network model is trained using the loader deceleration-stop safety function failure image.

[0081] In step S353, a large number of deceleration-stop safety function failure training sample images are generated using the third generative adversarial network model.

[0082] Those skilled in the art will appreciate that, while three generative adversarial network models are trained for three different types of images to generate training sample images, only one generative adversarial network model can be trained to generate different types of training sample images.

[0083] In some embodiments, the algorithm training process of the generative adversarial network model is as follows:

[0084] For each type of sample image data, the input parameters include: the time-frequency dataset distribution P(x) generated by fusion of real sensor signals, the number of adversarial training iterations T, the number of training iterations K for each discriminator network, the number of mini-batch samples M, the generator network G(z, v), the discriminator network D(x, r), and the noise distribution N(z). The output parameters include: the parameters v of the generator network and the parameters r of the discriminator network.

[0085] Randomly initialize parameters v, r;

[0086] For t←1to T do

[0087] / / Train the discriminator network D(x,r)

[0088] For k←1to K do

[0089] Collect M samples {x m},1≤m≤M;

[0090] Collect M samples from the noise distribution N(z) {z m},1≤m≤M;

[0091] Use stochastic gradient ascent to update the parameter r with the gradient:

[0092] End

[0093] / / Train generator network G(z,v)

[0094] Collect M samples from the noise distribution N(z) {z m},1≤m≤M;

[0095] Use stochastic gradient ascent to update the parameter v with the gradient:

[0096] End

[0097] In step S36, a safety function failure recognition classifier is trained using the training sample images. The safety function failure recognition classifier is used to extract and classify features from the images. The features extracted from the images may be, for example, one or more of geometric features, shape features, amplitude features, and histogram features.

[0098] In some embodiments, the safety function failure identification classifier is constructed using a deep convolutional neural network, such as GoogLeNet, AlexNet, and ResNet.

[0099] In the above embodiment, the training of the two models facilitates the subsequent judgment of the safety function status of the loader during operation, effectively reduces the probability of safety function failure of the loader, and thus effectively improves the safety of the loader during operation.

[0100] FIG4 is a flowchart of some embodiments of the vehicle safety function control method disclosed herein, which includes steps S41 - S44 .

[0101] In step S41, vehicle operation data is acquired.

[0102] In some embodiments, the vehicle includes a loader, and the vehicle operation data includes an accelerator pedal signal, a brake pedal signal, and a torque feedback signal.

[0103] In step S42, the vehicle operation data is fused into a data image.

[0104] In step S43, the data image is input into a safety function failure recognition classifier to obtain the safety function status of the vehicle, which includes a safety function normal state or a safety function failure state. The safety function failure recognition classifier is obtained based on the training method of the above embodiment.

[0105] In some embodiments, the safety function failure state includes a speed safety function failure state or a deceleration-stop safety function failure state.

[0106] In step S44, the safety function of the vehicle is controlled according to the state of the safety function of the vehicle.

[0107] In the above embodiment, the trained safety function failure recognition classifier has the ability to recognize safety function failures. By using the trained safety function failure recognition classifier, the current safety function status of the vehicle can be more accurately recognized, thereby improving the operational safety of the vehicle.

[0108] In some embodiments, step S44 is described by taking FIG. 5 as an example. FIG. 5 is a flowchart of other embodiments of the vehicle safety function control method disclosed herein.

[0109] In step S4411, in response to the vehicle being in a speed safety function failure state, the vehicle is controlled to enter a safe state.

[0110] For example, by monitoring the accelerator pedal signal and torque feedback signal, if unexpected torque output occurs, the safety function failure identification classifier will output that the loader is in a speed safety function failure state. In this case, a command is immediately sent to control the vehicle to enter a safe state. For example, by pressing the safe state button, the vehicle enters a safe state, ensuring safe operation.

[0111] In step S4412, determine whether the vehicle is in the driving process. If so, execute step S4413; otherwise, execute step S4414.

[0112] In step S4413, in response to the vehicle being in a driving state, the vehicle speed is controlled according to a predetermined accelerator pedal instruction.

[0113] For example, when the loader is in motion, the speed is forced to be controlled according to the expected instructions, such as controlling the loader to slowly decelerate until it stops.

[0114] In step S4414, in response to the vehicle being in the start-up state, the vehicle is controlled to stop running.

[0115] For example, if the loader is not in the driving process, that is, is in the starting period, the movement of the loader is stopped.

[0116] In the above embodiment, when it is identified that the vehicle is in a speed safety function failure state, the speed safety function of the vehicle can be controlled in a timely manner, thereby improving the operational safety of the vehicle, clarifying the path for the application of vehicle functional safety technology, and thus improving the safety of the vehicle.

[0117] In some embodiments, step S44 is described by taking FIG. 6 as an example. FIG. 6 is a flowchart of other embodiments of the vehicle safety function control method disclosed herein.

[0118] In step S4421, in response to the vehicle being in a deceleration-stop safety function failure state, the vehicle is controlled to enter a safe state.

[0119] For example, by monitoring the brake pedal signal and torque feedback signal, if unexpected torque output occurs, the safety function failure identification classifier will indicate that the loader has failed the deceleration-stop safety function. In this case, a command is immediately sent to control the vehicle to a safe state. For example, by pressing a safe state button, the vehicle enters a safe state, ensuring safe operation.

[0120] In step S4422, determine whether the vehicle is in the driving process. If so, execute step S4423; otherwise, execute step S4424.

[0121] In step S4423, in response to the vehicle being in a driving state, the vehicle speed is controlled according to a predetermined brake pedal instruction.

[0122] For example, when the loader is in motion, the speed is forced to be controlled according to the expected instructions, such as controlling the loader to slowly decelerate until it stops.

[0123] In step S4424, in response to the vehicle being in the start-up state, the vehicle is controlled to stop running.

[0124] For example, if the loader is not in the driving process, that is, is in the starting period, the movement of the loader is stopped.

[0125] In the above embodiment, when it is identified that the vehicle is in a deceleration-stop safety function failure state, the deceleration-stop safety function of the vehicle can be controlled in a timely manner, thereby improving the operational safety of the vehicle, clarifying the path for the application of vehicle functional safety technology, and thus improving the safety of the vehicle.

[0126] In some embodiments, in response to the vehicle being in a normal safety function state, the vehicle driving state is controlled to remain unchanged. For example, when the loader speed and deceleration / stop safety functions are normal, the loader maintains a normal driving state.

[0127] FIG7 is a flowchart of another embodiment of the vehicle safety function control method disclosed herein, which includes steps S71 - S76 .

[0128] In step S71 , a shift switch signal and a direction feedback signal of the vehicle are acquired.

[0129] In some embodiments, the shift switch signal and the direction feedback signal of the motor controller are monitored in real time by the vehicle safety controller.

[0130] In step S72, based on the shift switch signal and the direction feedback signal, it is determined whether the vehicle is in a reversing safety function failure state. If so, step S73 is executed; otherwise, the vehicle maintains a normal driving state.

[0131] In some embodiments, the shift switch signal and the direction feedback signal of the motor controller are monitored. If an unexpected direction output occurs, it indicates that a commutation abnormality has occurred in the loader. In this case, the loader needs to be controlled to enter a safe state.

[0132] In step S73 , in response to the vehicle being in the direction-reversing safety function failure state, the vehicle is controlled to enter a safe state.

[0133] For example, an instruction is immediately issued to control the loader to enter a safe state.

[0134] In step S74, it is determined whether the vehicle is in motion. If so, step S75 is executed; otherwise, step S76 is executed.

[0135] In step S75 , in response to the vehicle being in a driving state, the direction of the vehicle is controlled according to a predetermined instruction.

[0136] For example, if the loader is in motion, the loader is forced to be steered according to the expected instructions, for example, the loader is controlled to slowly decelerate in the predetermined direction until it stops.

[0137] In step S76 , in response to the vehicle being in the start state, the vehicle is controlled to stop running.

[0138] For example, if the loader is in the starting stage, the loader movement is stopped.

[0139] In the above embodiment, when the vehicle is in a state where the reversing safety function fails, the safety function of the vehicle can be controlled in a timely manner, thereby improving the operational safety of the vehicle.

[0140] FIG8 is a flowchart of another embodiment of the vehicle safety function control method disclosed herein, which includes steps S81 - S86 .

[0141] In step S81 , a brake switch signal and a torque feedback signal of the vehicle are acquired.

[0142] In some embodiments, the brake switch signal is detected by a brake sensor, and the torque feedback signal is detected by a torque feedback sensor.

[0143] In step S82, based on the brake switch signal and the torque feedback signal, it is identified whether the vehicle is in a brake safety function failure state. If so, step S83 is executed; otherwise, the vehicle maintains a normal driving state.

[0144] For example, if the loader's parking brake fails to activate as expected, it indicates that the loader's parking brake is abnormal. For example, if the brake switch is pressed and the torque feedback signal is 0, it indicates that the brake safety function is normal. If the torque feedback signal is not 0, it indicates that the brake safety function has failed.

[0145] In step S83 , in response to the vehicle being in a brake safety function failure state, the vehicle is controlled to enter a safe state.

[0146] For example, in the event of an abnormal parking brake condition on a loader, an instruction is immediately issued to control the loader to enter a safe state.

[0147] In step S84 , it is determined whether the vehicle's brake is activated. If so, step S85 is executed; otherwise, step S86 is executed.

[0148] In step S85 , in response to the vehicle brake being activated, the output torque of the vehicle is controlled to be zero.

[0149] For example, when the brake of the loader is activated, a control command is issued to set the loader output torque to 0.

[0150] In step S86 , in response to the vehicle brake being inactivated, the output torque of the vehicle is controlled to remain unchanged.

[0151] For example, in a case where the brakes of the loader are not activated, the loader maintains the current output torque.

[0152] In the above embodiment, when the vehicle is in a brake safety function failure state, the vehicle's safety function can be controlled in a timely manner, thereby improving the operational safety of the vehicle.

[0153] FIG9 is a flowchart of another embodiment of the vehicle safety function control method disclosed herein, which includes steps S91 - S94 .

[0154] In step S91, the emergency stop switch signal of the vehicle is acquired.

[0155] In step S92, based on the emergency stop switch signal, it is determined whether the vehicle is in the emergency stop switch safety function state. If so, step S93 is executed; otherwise, the vehicle maintains the normal driving state.

[0156] For example, if the emergency stop switch is pressed, it means that the loader is in the emergency stop switch safety function state.

[0157] In step S93 , in response to the vehicle being in the emergency stop switch safety function state, the vehicle is controlled to enter a safe state.

[0158] In step S94, the vehicle is controlled to stop running.

[0159] For example, sending an operation instruction to control the loader to stop moving.

[0160] In the above embodiment, when the vehicle is in the emergency stop switch safety function state, the vehicle's safety function can be controlled in a timely manner, thereby improving the vehicle's operational safety.

[0161] In some embodiments of the present disclosure, various safety functions of earthmoving machinery loaders are defined according to the ISO 19014 functional safety standard, including a loader speed safety function, a loader reversing safety function, a loader deceleration-stop safety function, a loader parking brake safety function, and a loader emergency stop safety function. The loader's accelerator pedal signal, brake pedal signal, torque feedback signal, direction feedback signal, brake switch signal, and emergency stop switch signal are acquired. The accelerator pedal signal, brake pedal signal, and torque feedback signal are fused into a time-frequency image. A trained safety function failure recognition classifier is then used to identify whether the loader is in a speed safety function failure state, a deceleration-stop safety function failure state, or a normal speed or deceleration-stop safety function state. Furthermore, based on the torque feedback signal, direction feedback signal, brake switch signal, and emergency stop switch signal, it is directly determined whether the loader is in a reversing safety function failure state, a brake safety function failure state, or an emergency stop switch safety function state, thereby controlling various safety functions of the loader. This control method can effectively reduce the probability of safety function failure in the loader and effectively improve the safety of the loader during operation.

[0162] The steps of FIG. 4 to FIG. 9 are executed by a vehicle safety function control device, which may be, for example, a vehicle safety controller of a loader, or a controller or control unit module that is only responsible for the vehicle safety function.

[0163] FIG10 is a schematic structural diagram of some embodiments of a training device for a safety function failure identification classifier disclosed herein. The training device includes: a first data acquisition module 1010 , a first training module 1020 , and a second training module 1030 .

[0164] The first data acquisition module 1010 is configured to acquire a signal sample image of the vehicle, wherein the signal sample image includes a safety function normal image and a safety function failure image.

[0165] In some embodiments, the vehicle is a loader, and the safety function failure image includes a speed safety function failure image and a deceleration-stop safety function failure image.

[0166] The first data acquisition module 1010 respectively acquires data sample signals of the vehicle under normal operation, failure of the speed safety function, and failure of the deceleration-stop safety function, wherein the data sample signals include an accelerator pedal sample signal, a brake pedal sample signal, and a torque feedback sample signal; and fuses the data sample signals into a signal sample image.

[0167] For example, the generation time of the data sample signal is obtained, and a signal sample image is generated according to the data sample signal and the generation time. The signal sample image is, for example, a time domain image or a time-frequency image.

[0168] The first training module 1020 is configured to train a deep neural network model using the signal sample image, so as to perform data augmentation processing on the signal sample image using the trained deep neural network model to obtain a training sample image.

[0169] In some embodiments, the deep neural network model includes a generative adversarial network model. The first training module 1020 inputs random noise into the generator of the generative adversarial network model to generate a forged sample image; inputs the signal sample image and the forged sample image into the discriminator of the generative adversarial network model to obtain the sample authenticity; and trains the discriminator using a first loss function and trains the generator using a second loss function, wherein the value of the first loss function is negatively correlated with the accuracy of the sample authenticity, and the value of the second loss function is positively correlated with the accuracy of the sample authenticity.

[0170] The second training module 1030 is configured to train a safety function failure recognition classifier using training sample images, wherein the safety function failure recognition classifier is used to identify the safety function status of the vehicle during operation, and the safety function status includes a normal safety function status or a failed safety function status.

[0171] In some embodiments, the safety function status includes a safety function normal state, a speed safety function failure state, or a deceleration-stop safety function failure state.

[0172] In the above embodiment, a deep neural network model is constructed and trained, and the deep neural network model is used to perform data augmentation processing on the signal sample image to obtain a large number of training sample images, which are then used to train a safety function failure recognition classifier, so that the safety function failure recognition classifier can be used to identify the safety function status of the vehicle, which can improve the accuracy and speed of vehicle safety function status recognition.

[0173] FIG11 is a schematic structural diagram of some embodiments of a vehicle safety function control device disclosed herein, which includes a second data acquisition module 1110 , a state determination module 1120 , and a control module 1130 .

[0174] The second data acquisition module 1110 is configured to acquire vehicle operation data of the vehicle and fuse the vehicle operation data into a signal image.

[0175] In some embodiments, the vehicle includes a loader, and the vehicle operation data includes an accelerator pedal signal, a brake pedal signal, and a torque feedback signal.

[0176] The state determination module 1120 is configured to input the data image into the safety function failure recognition classifier to obtain the safety function state of the vehicle, where the safety function state includes a normal safety function state or a safety function failure state, wherein the safety function failure recognition classifier is obtained based on the training method of the above embodiment.

[0177] In some embodiments, the safety function failure state includes a speed safety function failure state or a deceleration-stop safety function failure state.

[0178] The control module 1130 is configured to control the safety function of the vehicle according to the status of the safety function of the vehicle.

[0179] In some embodiments, in response to the vehicle being in a speed safety function failure state, the vehicle is controlled to enter a safe state; in response to the vehicle being in a driving state, the vehicle speed is controlled according to a predetermined accelerator pedal instruction; and in response to the vehicle being in a starting state, the vehicle is controlled to stop running.

[0180] In some embodiments, in response to the vehicle being in a deceleration-stop safety function failure state, the vehicle is controlled to enter a safe state; in response to the vehicle being in a driving state, the vehicle speed is controlled according to a predetermined brake pedal instruction; and in response to the vehicle being in a starting state, the vehicle is controlled to stop running.

[0181] In some embodiments, in response to the vehicle being in a normal safety function state, the vehicle's driving state is controlled to remain unchanged.

[0182] In the above embodiment, the trained safety function failure recognition classifier has the ability to recognize safety function failures. By using the trained safety function failure recognition classifier, the current safety function status of the vehicle can be more accurately recognized, thereby improving the operational safety of the vehicle.

[0183] In other embodiments of the present disclosure, the second data acquisition module 1110 is further configured to acquire a shift switch signal and a direction feedback signal from the vehicle. The state determination module 1120 is further configured to identify whether the vehicle is in a reverse safety function failure state based on the shift switch signal and the direction feedback signal. The control module 1130 is further configured to, in response to the vehicle being in a reverse safety function failure state, control the vehicle to enter a safe state and determine whether the vehicle is in a driving state; in response to the vehicle being in a driving state, control the vehicle's direction according to a predetermined instruction; and, in response to the vehicle being in a starting state, control the vehicle to stop running.

[0184] In the above embodiment, when the vehicle is in a state where the reversing safety function fails, the safety function of the vehicle can be controlled in a timely manner, thereby improving the operational safety of the vehicle.

[0185] In some embodiments, the second data acquisition module 1110 is further configured to acquire a brake switch signal from the vehicle. The state determination module 1120 is further configured to identify whether the vehicle is in a brake safety function failure state based on the brake switch signal and the torque feedback signal. The control module 1130 is further configured to, in response to the vehicle being in a brake safety function failure state, control the vehicle to enter a safe state and determine whether the vehicle brake is activated; in response to the vehicle brake being activated, control the vehicle output torque to zero; and in response to the vehicle brake not being activated, control the vehicle output torque to remain unchanged.

[0186] In the above embodiment, when the vehicle is in a brake safety function failure state, the vehicle's safety function can be controlled in a timely manner, thereby improving the operational safety of the vehicle.

[0187] In some embodiments, the second data acquisition module 1110 is further configured to acquire an emergency stop switch signal from the vehicle. The state determination module 1120 is further configured to identify whether the vehicle is in an emergency stop switch safety function state based on the emergency stop switch signal. The control module 1130 is further configured to control the vehicle to enter a safe state and stop operation in response to the vehicle being in the emergency stop switch safety function state.

[0188] In the above embodiment, when the vehicle is in the emergency stop switch safety function state, the vehicle's safety function can be controlled in a timely manner, thereby improving the vehicle's operational safety.

[0189] Figure 12 is a schematic diagram of the structure of some embodiments of the electronic device disclosed herein. The electronic device 1200 includes a memory 1210 and a processor 1220. The memory 1210 may be a disk, flash memory, or any other non-volatile storage medium. The memory is used to store instructions in the above embodiments. The processor 1220 is coupled to the memory 1210 and may be implemented as one or more integrated circuits, such as a microprocessor or microcontroller. The processor 1220 is used to execute instructions stored in the memory.

[0190] In some embodiments, the processor 1220 is coupled to the memory 1210 via a BUS 1230. The electronic device 1200 may also be connected to an external storage device 1250 via a storage interface 1240 to access external data, and may also be connected to a network or another computer system (not shown) via a network interface 1260. Detailed descriptions are omitted here.

[0191] In this embodiment, by storing data instructions in a memory and then processing the instructions through a processor, the probability of safety function failure in the vehicle can be effectively reduced, and the safety of the vehicle during operation can be effectively improved.

[0192] In other embodiments of the present disclosure, a vehicle, such as a loader, is protected, and includes a training device for the safety function failure identification classifier, a vehicle safety function control device, or an electronic device in the above-mentioned embodiment.

[0193] In some embodiments, as shown in FIG13 , which is a schematic structural diagram of some embodiments of the vehicle disclosed herein, the vehicle includes a sensor group 1310 , a vehicle safety function control device 1320 , a motor controller 1330 , a relay 1340 , a human-machine interface 1350 , a warning light 1360 , and a buzzer 1370 .

[0194] The sensor group 1310 includes multiple sensors for collecting various vehicle signals. The vehicle safety function control device 1320 is the vehicle's control center, responsible for receiving signals from the sensor group, running control algorithms, and outputting control instructions to actuators to control vehicle operation and safety functions. These actuators include a motor controller 1330, relays 1340, a human-machine interface 1350, a warning light 1360, and a buzzer 1370.

[0195] The motor controller 1330 actively controls the motor to operate according to a set direction, speed, angle, and response time. The relay 1340 acts as an "automatic switch" that uses a small current to control a large current, providing safety protection and switching functions. For example, when the emergency stop switch is pressed, the relay is de-energized, and the loader stops moving. The human-machine interface 1350 displays the vehicle's status information and related sensor values. The warning light 1360 generates an alarm signal if a vehicle safety feature fails. The buzzer 1370 emits an audible sound to alert the operator / user if a vehicle safety feature fails.

[0196] In other embodiments, a computer-readable storage medium stores computer program instructions thereon, which, when executed by a processor, implement the steps of the method in the above-described embodiment. Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, devices, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable non-transient storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0197] In some embodiments, a computer program product is protected, including a computer program or instructions, which implement the above-mentioned training method or vehicle safety function control method when executed by a processor. The computer program product includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by the CPU, the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.

[0198] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram and the combination of the processes and / or boxes in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate a device for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0199] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0200] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0201] In some embodiments of the present disclosure, a computer program is further provided, comprising: instructions, which, when executed by a processor, cause the processor to execute the training method as described above, or the vehicle safety function control method.

[0202] The present disclosure has been described in detail so far. To avoid obscuring the concept of the present disclosure, some details known in the art have not been described. Based on the above description, those skilled in the art can fully understand how to implement the technical solutions disclosed herein.

[0203] The methods and apparatus of the present disclosure may be implemented in many ways. For example, the methods and apparatus of the present disclosure may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present disclosure are not limited to the order specifically described above, unless otherwise specifically stated. In addition, in some embodiments, the present disclosure may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the methods according to the present disclosure. Therefore, the present disclosure also covers recording media that store programs for executing the methods according to the present disclosure.

[0204] Although some specific embodiments of the present disclosure have been described in detail by way of examples, those skilled in the art will appreciate that the above examples are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Those skilled in the art will appreciate that modifications may be made to the above embodiments without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A method for training a safety function failure identification classifier, comprising: Acquiring a signal sample image of the vehicle, wherein the signal sample image includes a normal safety function image and a failed safety function image; Using the signal sample images to train a deep neural network model, and using the trained deep neural network model to perform data augmentation processing on the signal sample images to obtain training sample images; and The training sample images are used to train a safety function failure recognition classifier, wherein the safety function failure recognition classifier is used to recognize a safety function status of the vehicle during operation, wherein the safety function status includes a safety function normal state or a safety function failure state.

2. The training method according to claim 1, wherein: The vehicle is a loader, the safety function failure image includes a speed safety function failure image and a deceleration-stop safety function failure image, and the safety function status includes a safety function normal state, a speed safety function failure state, or a deceleration-stop safety function failure state.

3. The training method according to claim 2, wherein: Acquiring the signal sample image of the vehicle includes: respectively acquiring data sample signals of the vehicle in normal operation, when the speed safety function fails, and when the deceleration-stop safety function fails, wherein the data sample signals include an accelerator pedal sample signal, a brake pedal sample signal, and a torque feedback sample signal; and The data sample signals are fused into the signal sample image.

4. The training method according to claim 3, wherein: Acquiring the signal sample image of the vehicle further includes: Acquiring a generation time of the data sample signal, wherein converting the data sample signal into the signal sample image includes: The signal sample image is generated according to the data sample signal and the generation time.

5. The training method according to any one of claims 1 to 4, wherein: The deep neural network model includes a generative adversarial network model, and training the deep neural network model using the signal sample image includes: Inputting random noise into the generator of the generative adversarial network model to generate a forged sample image; Inputting the signal sample image and the forged sample image into the discriminator of the generative adversarial network model to obtain the sample authenticity; and The discriminator is trained using a first loss function, and the generator is trained using a second loss function, wherein a value of the first loss function is negatively correlated with the accuracy of the sample authenticity, and a value of the second loss function is positively correlated with the accuracy of the sample authenticity.

6. A vehicle safety function control method comprising: Obtain vehicle operation data; fusing the vehicle operation data into a data image; Inputting the data image into a safety function failure recognition classifier to obtain a safety function status of the vehicle, wherein the safety function status includes a safety function normal state or a safety function failure state, wherein the safety function failure recognition classifier is obtained based on the training method according to any one of claims 1 to 5; as well as The safety function of the vehicle is controlled according to the safety function status of the vehicle.

7. The vehicle safety function control method according to claim 6, wherein: The vehicle includes a loader, the vehicle operation data includes an accelerator pedal signal, a brake pedal signal, and a torque feedback signal, and the safety function failure state includes a speed safety function failure state or a deceleration-stop safety function failure state.

8. The vehicle safety function control method according to claim 7, wherein: Controlling the safety function of the vehicle according to the safety function status of the vehicle includes: In response to the vehicle being in a speed safety function failure state, controlling the vehicle to enter a safe state and determining whether the vehicle is in a driving state; In response to the vehicle being in a driving state, controlling the speed of the vehicle according to a predetermined accelerator pedal command; and In response to the vehicle being in the start state, the vehicle is controlled to stop running.

9. The vehicle safety function control method according to claim 7 or 8, wherein: Controlling the safety function of the vehicle according to the safety function status of the vehicle includes: In response to the vehicle being in a deceleration-stop safety function failure state, controlling the vehicle to enter a safe state and determining whether the vehicle is in a driving state; In response to the vehicle being in a driving state, controlling the speed of the vehicle according to a predetermined brake pedal command; and In response to the vehicle being in the start state, the vehicle is controlled to stop running.

10. The vehicle safety function control method according to any one of claims 7 to 9, wherein: Controlling the safety function of the vehicle according to the safety function status of the vehicle includes: In response to the vehicle being in a normal safety function state, controlling the vehicle's driving state to remain unchanged.

11. The vehicle safety function control method according to any one of claims 6 to 10, further comprising: Obtaining a shift switch signal and a direction feedback signal of the vehicle; identifying whether the vehicle is in a reversing safety function failure state according to the shift switch signal and the direction feedback signal; In response to the vehicle being in a state where the direction-reversing safety function is disabled, controlling the vehicle to enter a safe state and determining whether the vehicle is in a driving state; In response to the vehicle being in a driving state, controlling the direction of the vehicle according to a predetermined instruction; as well as In response to the vehicle being in the start state, the vehicle is controlled to stop running.

12. The vehicle safety function control method according to any one of claims 6 to 11, further comprising: obtaining a brake switch signal of the vehicle; identifying whether the vehicle is in a brake safety function failure state according to the brake switch signal and the torque feedback signal; In response to the vehicle being in a brake safety function failure state, controlling the vehicle to enter a safe state and determining whether the vehicle brake is activated; In response to the vehicle brake being activated, controlling the output torque of the vehicle to be 0; as well as In response to the vehicle brakes being deactivated, the output torque of the vehicle is controlled to remain unchanged.

13. The vehicle safety function control method according to any one of claims 6 to 12, further comprising: Obtaining an emergency stop switch signal of the vehicle; Identifying whether the vehicle is in an emergency stop switch safety function state according to the emergency stop switch signal; as well as In response to the vehicle being in an emergency stop switch safety function state, the vehicle is controlled to enter a safe state and to stop running.

14. A training device for a safety function failure identification classifier, comprising: A first data acquisition module is configured to acquire a signal sample image of the vehicle, wherein the signal sample image includes a normal safety function image and a failed safety function image; a first training module configured to train a deep neural network model using the signal sample image, so as to perform data augmentation processing on the signal sample image using the trained deep neural network model to obtain a training sample image; and The second training module is configured to train a safety function failure recognition classifier using the training sample images, wherein the safety function failure recognition classifier is used to identify the safety function status of the vehicle during operation, and the safety function status includes a normal safety function status or a failed safety function status.

15. A vehicle safety function control device comprising: a second data acquisition module configured to acquire vehicle operation data of the vehicle and fuse the vehicle operation data into a signal image; a state determination module, configured to input the data image into a safety function failure recognition classifier to obtain a safety function state of the vehicle, wherein the safety function state includes a safety function normal state or a safety function failure state, wherein the safety function failure recognition classifier is obtained based on the training method according to any one of claims 1 to 5; as well as The control module is configured to control the safety function of the vehicle according to the safety function status of the vehicle.

16. An electronic device comprising: Memory; as well as A processor coupled to the memory, the processor being configured to execute the training method according to any one of claims 1 to 5, or the vehicle safety function control method according to any one of claims 6 to 13, based on instructions stored in the memory.

17. A vehicle comprising: At least one of the training device for the safety function failure identification classifier according to claim 14, the vehicle safety function control device according to claim 15, and the electronic device according to claim 16.

18. A computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the training method according to any one of claims 1 to 5, or the vehicle safety function control method according to any one of claims 6 to 13.

19. A computer program product, comprising a computer program or instructions, which, when executed by a processor, implements the training method described in any one of claims 1 to 5, or the vehicle safety function control method described in any one of claims 6 to 13.

20. A computer program comprising: Instructions, when executed by a processor, cause the processor to execute the training method according to any one of claims 1 to 5, or the vehicle safety function control method according to any one of claims 6 to 13.

Images