An accelerator anti-misstep method, system and storage medium

Abstract

This invention proposes a method, system, and storage medium for preventing accidental accelerator pedal presses, relating to the technical field of intelligent assisted driving. The method includes: real-time detection of data on the driver pressing the accelerator pedal; then, using a deep learning anomaly detection model to identify anomalies in the data on the driver pressing the accelerator pedal, and monitoring whether there are obstacles on the vehicle's travel route and whether there is a collision risk between the vehicle and the obstacles; further, based on the data anomalies and collision risks, determining whether the driver has accidentally pressed the accelerator pedal; if a confirmation signal of accidental accelerator pedal press is received, the accelerator pedal is suppressed and acceleration is not responded to; finally, when the driver completely releases the accelerator pedal or presses the brake pedal, the control function of the control execution module exits, achieving more intelligent and accurate identification of accidental accelerator pedal presses, avoiding or reducing the risk of collision accidents caused by accidental accelerator pedal presses.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent assisted driving, and in particular to a method, system and storage medium for preventing accidental accelerator pedal press. Background Technology

[0002] Traffic accidents caused by drivers mistaking the accelerator pedal for the brake pedal have been frequent, causing great harm to people's lives, health and property. Therefore, studying the problem of drivers accidentally pressing the accelerator pedal is an important safety issue that needs to be addressed in intelligent assisted driving systems.

[0003] Existing patent literature discloses a device and method for correcting and correcting accidental accelerator pedal use. The method determines that the driver has accidentally pressed the accelerator pedal when both the accelerator pedal opening and the pedal speed exceed a preset threshold, the driver is distracted, and there is a risk of collision. Otherwise, it determines that the driver has not accidentally pressed the accelerator pedal. However, this patent relies on monitoring the driver's concentration and distraction levels. Driver attention is uncertain; even during normal driving, drivers can be distracted, leading to misjudgments or missed detections of accidental accelerator pedal use. Furthermore, the method only compares the opening and speed with a set threshold for accelerator pedal parameter determination, without using machine learning to comprehensively judge based on historical data of normal accelerator pedal use. Therefore, the identification of accidental accelerator pedal use is highly random and uncertain. Existing patent documents also disclose a method, system, and storage medium for controlling accidental accelerator pedal presses. When the accelerator pedal opening increment reaches a preset threshold, and there is an obstacle in the direction of the vehicle's movement while the vehicle speed meets the preset speed requirement, the engine power output is stopped. However, this patent only identifies the presence of an obstacle in the direction of the vehicle's movement and does not distinguish whether there is a collision risk between the vehicle and the obstacle. Furthermore, it relies solely on the accelerator pedal opening increment to determine whether the accelerator pedal has been accidentally pressed, without fully considering the accelerator pedal pressure, accelerator pedal opening, and accelerator pedal opening change rate for a comprehensive judgment. In addition, it does not involve the processing logic for suppressing acceleration and the human-computer interaction design, so its impact on eliminating accidental accelerator pedal presses is relatively limited. Summary of the Invention

[0004] To address the shortcomings of the prior art, the present invention aims to solve the problem of how to more intelligently and accurately identify abnormal accelerator pedal misoperation caused by the driver mistaking the accelerator pedal for the brake pedal by using parameters of the driver pressing the accelerator pedal, thereby avoiding or reducing the risk of collision accidents caused by accidental accelerator pedal misoperation.

[0005] To achieve the above-mentioned technical effects, the technical solution of the present invention is as follows:

[0006] A method for preventing accidental accelerator pedal press includes the following steps:

[0007] The accelerator pedal parameter monitoring module monitors the data of the driver pressing the accelerator pedal in real time and transmits the monitored data to the parameter anomaly identification module. The parameter anomaly identification module uses a deep learning anomaly detection model to identify anomalies in the data of the driver pressing the accelerator pedal. The collision risk monitoring module monitors in real time whether there are obstacles on the vehicle's travel route and whether there is a collision risk between the vehicle and the obstacles.

[0008] The Accidental Accelerator Pedal Detection Module determines whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module then sends a signal confirming accidental accelerator pedal press to the control execution module.

[0009] The control execution module receives a confirmation signal that the accelerator pedal has been accidentally pressed and then suppresses the accelerator pedal, refraining from accelerating.

[0010] When the driver fully releases the accelerator or presses the brake pedal, the control function of the control execution module is deactivated.

[0011] Furthermore, it also includes the following steps:

[0012] The interactive warning module receives a confirmation signal that the accelerator pedal has been accidentally pressed and alerts the driver.

[0013] Furthermore, the data on the driver pressing the accelerator pedal includes the pressure applied to the accelerator pedal, the pedal opening, and the rate of change of the pedal opening.

[0014] Furthermore, the method for the parameter anomaly detection module to detect anomalies in the data of the driver pressing the accelerator pedal includes the following steps:

[0015] Obtain a dataset of historical normal acceleration using the accelerator pedal;

[0016] Preprocess the dataset of historical normal acceleration using the accelerator pedal;

[0017] Construct a deep learning anomaly detection model, train the constructed deep learning anomaly detection model using the preprocessed dataset, and obtain the trained deep learning anomaly detection model.

[0018] The trained deep learning anomaly detection model is used to perform deep learning anomaly detection on the data of a driver pressing the accelerator pedal once, and the anomaly identification results of the data of a driver pressing the accelerator pedal once are obtained.

[0019] Update the dataset with the results of identifying data anomalies in the accelerator pedal.

[0020] Furthermore, the historical dataset of normal acceleration using the accelerator pedal is preprocessed. The specific preprocessing process is as follows: digital feature extraction is performed on the historical dataset of normal acceleration using the accelerator pedal to obtain the digital dimension data of the dataset. The digital dimension data is then transformed to obtain multi-dimensional feature data of the accelerator pedal pressing behavior.

[0021] Furthermore, the digital dimension data undergoes feature transformation. The specific transformation process involves regularization, category mapping, and function transformation of the digital dimension data. The specific steps for function transformation of the digital dimension data are as follows: obtaining the distribution function of the digital dimension data, mapping the distribution function to a linear function, and obtaining the multi-dimensional feature data corresponding to the digital dimension data.

[0022] Furthermore, if the estimated collision time between the vehicle and the obstacle is less than the collision time threshold, then there is a risk of collision between the vehicle and the obstacle; otherwise, there is no risk of collision between the vehicle and the obstacle.

[0023] The present invention also proposes an accelerator pedal misoperation prevention system, which includes: an accelerator pedal parameter monitoring module, a parameter anomaly identification module, a collision risk monitoring module, a misoperation accelerator pedal judgment module, and a control execution module;

[0024] The accelerator pedal parameter monitoring module is used to monitor the data of the driver pressing the accelerator pedal in real time and transmit the monitored data of the driver pressing the accelerator pedal to the parameter anomaly identification module.

[0025] The parameter anomaly identification module is used to identify anomalies in the data of the driver pressing the accelerator pedal using a deep learning anomaly detection model.

[0026] The collision risk monitoring module is used to monitor in real time whether there are obstacles in the vehicle's travel path and whether there is a risk of collision between the vehicle and the obstacle;

[0027] The Accidental Accelerator Pedal Detection Module is used to determine whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module sends a signal confirming accidental accelerator pedal pressing to the control execution module.

[0028] The control execution module is used to receive a confirmation signal of accidental accelerator pedal press and suppress the accelerator pedal, and not respond to acceleration; when the driver completely releases the accelerator pedal or presses the brake pedal, the control function of the control execution module is deactivated.

[0029] Furthermore, the parameter anomaly identification module includes a data feature extraction unit, a data feature input unit, an anomaly detection model unit, and a model parameter adjustment unit;

[0030] The data feature extraction unit is used to extract multi-dimensional features from the acquired historical dataset of normal acceleration using the accelerator pedal, and obtain multi-dimensional feature data of the accelerator pedal pressing behavior.

[0031] The data feature input unit is used to input the multi-dimensional feature data obtained from the processed data into the constructed deep learning anomaly detection model;

[0032] The anomaly detection model unit is used to detect anomalies in multi-dimensional feature data and obtain the identification results of data anomalies in the accelerator pedal.

[0033] The model parameter tuning unit is used to adjust the network parameters of the deep learning anomaly detection model during the training process.

[0034] The present invention also proposes a computer-readable storage medium storing a throttle pedal misoperation prevention program. When the throttle pedal misoperation prevention program is executed by a processor, the operation of the throttle pedal misoperation prevention method is implemented.

[0035] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0036] This invention proposes a method, system, and storage medium for preventing accidental accelerator pedal presses. First, it detects real-time data on the driver's accelerator pedal presses. Then, it uses a deep learning anomaly detection model to identify anomalies in the accelerator pedal press data, avoiding the mechanical use of threshold settings to determine accelerator pedal parameter abnormalities. It also monitors the presence of obstacles on the vehicle's path and the potential collision risk between the vehicle and obstacles. Based on the anomalies in the accelerator pedal press data and the potential collision risk, it determines whether the driver has accidentally pressed the accelerator. If a confirmation signal of accidental accelerator press is received, the accelerator is suppressed, and acceleration is not initiated. Finally, when the driver completely releases the accelerator or presses the brake pedal, the control function of the control execution module exits. This achieves a more intelligent and accurate identification of accidental accelerator pedal presses caused by mistaking the accelerator for the brake, thus avoiding or reducing the risk of collisions due to accidental accelerator pedal presses. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a method for preventing accidental accelerator pedal presses proposed in an embodiment of the present invention.

[0038] Figure 2 This is a schematic diagram illustrating the steps for anomaly identification of data related to a driver pressing the accelerator pedal, as proposed in an embodiment of the present invention.

[0039] Figure 3 This is a flowchart illustrating the anomaly detection of data related to the driver pressing the accelerator pedal, as proposed in this embodiment of the invention.

[0040] Figure 4 This is a schematic diagram of the structure of an accelerator pedal anti-accidental pressing system proposed in an embodiment of the present invention;

[0041] Figure 5 This is a flowchart illustrating a throttle anti-accidental pressing system proposed in an embodiment of the present invention.

[0042] Figure 6 This is a schematic diagram illustrating the structure of an abnormal accelerator pedal parameter identification system proposed in an embodiment of the present invention. Detailed Implementation

[0043] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0044] To better illustrate this embodiment, some parts of the accompanying drawings may be omitted, enlarged, or reduced, and do not represent actual dimensions. The descriptions of directions such as "up" and "down" are not intended to limit this patent.

[0045] It is understandable to those skilled in the art that some well-known details may be omitted from the accompanying drawings;

[0046] The positional relationships depicted in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0047] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0048] like Figure 1 As shown, this embodiment proposes a method for preventing accidental accelerator pedal presses, which includes the following steps:

[0049] S1. The accelerator pedal parameter monitoring module monitors the data of the driver pressing the accelerator pedal in real time and transmits the monitored data to the parameter anomaly identification module. The parameter anomaly identification module identifies anomalies in the data of the driver pressing the accelerator pedal. The collision risk monitoring module monitors in real time whether there are obstacles on the vehicle's travel route and whether there is a collision risk between the vehicle and the obstacles.

[0050] In step S1, the data on the driver pressing the accelerator pedal includes the pressing force of the accelerator pedal, the pedal opening, and the rate of change of the pedal opening. The rate of change of the pedal opening refers to the derivative of the pedal opening with respect to time.

[0051] The parameter anomaly detection module identifies anomalies in the data related to the driver pressing the accelerator pedal. See [link to documentation]. Figure 2 and Figure 3 It includes the following steps:

[0052] S11. Obtain the dataset of historical normal acceleration using the accelerator pedal;

[0053] S12. Preprocess the dataset of historical normal acceleration using the accelerator pedal;

[0054] In step S12, the historical dataset of normal acceleration by the accelerator pedal is preprocessed. The specific preprocessing process is as follows: digital feature extraction is performed on the historical dataset of normal acceleration by the accelerator pedal to obtain the digital dimension data of the dataset. The digital dimension data is then transformed to obtain multi-dimensional feature data of the accelerator pedal behavior. The digital dimension data is then transformed. The specific transformation process is as follows: regularization, class mapping and function transformation are performed on the digital dimension data respectively.

[0055] The specific steps for performing function transformation on digital dimension data are as follows: obtain the distribution function of the digital dimension data, map the distribution function to a linear function, and obtain the multi-dimensional feature data corresponding to the digital dimension data.

[0056] S13. Construct a deep learning anomaly detection model. Train the constructed deep learning anomaly detection model using the preprocessed dataset to obtain the trained deep learning anomaly detection model.

[0057] In step S13, during the training of the deep learning anomaly detection model, gradient descent is used to update the network parameters of the deep learning anomaly detection model. The network parameters include the parameters of the loss function minimized in the accelerator pedal parameter anomaly recognition model and the probability standard for outputting anomalies. The specific steps for training and constructing the accelerator pedal parameter anomaly recognition model are as follows:

[0058] S131. Based on the dataset, given training data X1, X2, ..., X... i ..., Xn;

[0059] Among them, X i This represents the multi-dimensional feature data corresponding to the i-th record. The training data are matched with the corresponding anomaly labels: y1, y2, ..., y i ...,yn;

[0060] Among them, y i ∈{0,1},y i Indicates an abnormal state, y i =1 indicates an anomaly, y i =0 indicates normal operation. A mapping function is learned through supervised learning. The specific expression of the mapping function is:

[0061] f(X) = p(y = 1|X)

[0062] Where p represents the probability that a given feature data record X is identified as an anomalous sample;

[0063] S132. By optimizing the loss function to minimize the probability of outlier samples, we obtain...

[0064] min(loss(y, f(X)))

[0065] The loss function, also known as the loss measure, is used to measure the difference between two values.

[0066] S133. Considering the unique characteristics of abnormal data collection, different weights are assigned to normal and abnormal samples, and the loss function is redefined as follows:

[0067] loss=Cp*loss(yp, f(Xp))+Cn*loss(yn, f(Xn))

[0068] Where Cp represents the weight of abnormal samples and Cn represents the weight of normal samples. By giving abnormal samples a larger weight, the recognition accuracy of the accelerator pedal parameter anomaly recognition model can be improved.

[0069] S134. Validate a single record if the following conditions are met:

[0070] f(Xi)>P0

[0071] The detection is then judged as an anomaly. The initial default value for P0 is P0 = 0.6, but in practical applications, it is calibrated based on the training results of historical datasets and is not limited to 0.6. It should be noted that the loss function model mentioned above is only an implementation example and is not limited to the scope of the patent. Any equivalent structural and procedural transformations made using the inventive process and its accompanying drawings, or the direct or indirect use of similar or related technologies in related fields, are all included within the scope of patent protection of the inventive.

[0072] S14. Use the trained deep learning anomaly detection model to perform deep learning anomaly detection on the data of the driver pressing the accelerator pedal once, and obtain the identification result of the anomaly of the data of the driver pressing the accelerator pedal once.

[0073] In step S14, the data of the driver pressing the accelerator pedal once includes the pressing force, the pedal opening, and the rate of change of the pedal opening. The recognition result of the trained deep learning anomaly detection model is used as the recognition result of the accelerator pedal data anomaly.

[0074] S15. Update the dataset with the results of identifying data anomalies in the accelerator pedal.

[0075] S2. Accidental Accelerator Pedal Judgment Module determines whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module sends a signal confirming accidental accelerator pedal pressing to the control execution module.

[0076] In step S2, if the estimated collision time between the vehicle and the obstacle is less than the collision time threshold, then there is a risk of collision between the vehicle and the obstacle; otherwise, there is no risk of collision between the vehicle and the obstacle. The obstacle detection sensor uses LiDAR, millimeter-wave radar and camera, etc., of the vehicle perception system. The collision risk determination logic of the obstacle detection sensor is as follows: if the estimated collision time of the vehicle's current speed is less than 2STTC, then it is considered that there is a risk of collision between the vehicle and the obstacle; otherwise, there is no risk of collision between the vehicle and the obstacle.

[0077] If both abnormal accelerator pedal data and a risk of collision between the vehicle and an obstacle are met, it is considered that the accelerator pedal has been accidentally pressed.

[0078] S3. The control execution module receives a confirmation signal that the accelerator pedal has been accidentally pressed and suppresses the accelerator pedal, without responding to acceleration;

[0079] S4. When the driver fully releases the accelerator or presses the brake pedal, the control function of the control execution module is deactivated.

[0080] In step S4, after receiving a signal confirming that the accelerator pedal has been accidentally pressed, the accelerator is suppressed, and the power output of the vehicle is controlled so that the vehicle maintains the cruising speed before the accelerator pedal was accidentally pressed, preventing the vehicle from accelerating too much and causing an accident.

[0081] It also includes the following steps:

[0082] The interactive warning module receives a confirmation signal that the accelerator pedal has been accidentally pressed and alerts the driver.

[0083] When the interactive warning module receives a confirmation signal of accidental acceleration, it will alert the driver through sound, images, text, and tactile feedback. The warning content includes: HUD graphic alarm, displayed for at least 1 second, with continuous alerts while the alarm signal is active; HU graphic alarm, displayed for at least 1 second, with continuous alerts while the alarm signal is active; steering wheel vibration / seatbelt tightening (based on whether both hands are on the steering wheel); and sound alarm, lasting at least 1 second.

[0084] In this embodiment, three accelerator pedal parameters—depression force, pedal opening, and rate of change of pedal opening—are first detected. Then, a deep learning anomaly detection model is constructed to avoid mechanically judging whether the accelerator pedal parameters are abnormal by setting thresholds. The model also monitors whether there are obstacles on the vehicle's path and whether there is a collision risk between the vehicle and the obstacles. When both the abnormal accelerator pedal parameters and the collision risk on the vehicle's path are met, it is determined that the driver has mistakenly pressed the accelerator pedal. The interactive warning module and the control execution module then intervene. Upon receiving the signal of mistaken accelerator pedal pressing, the interactive warning module, in conjunction with various facilities in the cabin, alerts the driver through text images, sound alarms, seatbelt tightening, or steering wheel vibration. The control execution module simultaneously suppresses the accelerator, preventing acceleration. This achieves a more intelligent and accurate identification of abnormal accelerator pedal pressing caused by mistaking the accelerator pedal for the brake pedal, thus avoiding or reducing the risk of collisions caused by mistaken accelerator pedal pressing.

[0085] In this embodiment, an accelerator pedal misoperation prevention system is also proposed, see [link to relevant documentation]. Figure 4 and Figure 5 The system includes: an accelerator pedal parameter monitoring module, a parameter anomaly identification module, a collision risk monitoring module, a mis-accelerator pedal judgment module, and a control execution module;

[0086] The accelerator pedal parameter monitoring module is used to monitor the data of the driver pressing the accelerator pedal in real time and transmit the monitored data of the driver pressing the accelerator pedal to the parameter anomaly identification module.

[0087] In the accelerator pedal parameter monitoring module, the data of the driver pressing the accelerator pedal includes the pressing force, the pedal opening, and the rate of change of the pedal opening. The data of the driver pressing the accelerator pedal is obtained from the data of the driver pressing the accelerator pedal once. The rate of change of the pedal opening refers to the derivative of the pedal opening with respect to time.

[0088] The parameter anomaly identification module is used to identify anomalies in the data of the driver pressing the accelerator pedal using a deep learning anomaly detection model.

[0089] See Figure 6 In the parameter anomaly identification module, the parameter anomaly identification module includes a data feature extraction unit, a data feature input unit, an anomaly detection model unit, and a model parameter adjustment unit;

[0090] The data feature extraction unit performs multi-dimensional feature extraction on the acquired historical dataset of normal acceleration using the accelerator pedal to obtain multi-dimensional feature data of accelerator pedal behavior.

[0091] The data feature input unit inputs the multi-dimensional feature data obtained from the processed data into the constructed deep learning anomaly detection model;

[0092] The anomaly detection model unit performs anomaly detection on multi-dimensional feature data to obtain the identification results of data anomalies in the accelerator pedal;

[0093] The model parameter tuning unit is used to adjust the network parameters of the deep learning anomaly detection model during the training process. The network parameters include the parameters of the loss function that minimizes the output and the probability criterion for anomalies in the deep learning anomaly detection model.

[0094] The collision risk monitoring module is used to monitor in real time whether there are obstacles in the vehicle's travel path and whether there is a risk of collision between the vehicle and the obstacle;

[0095] The Accidental Accelerator Pedal Detection Module is used to determine whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module sends a signal confirming accidental accelerator pedal pressing to the control execution module.

[0096] The control execution module is used to receive a confirmation signal of accidental accelerator pedal press and suppress the accelerator pedal, and not respond to acceleration; when the driver completely releases the accelerator pedal or presses the brake pedal, the control function of the control execution module is deactivated.

[0097] It also includes an interactive warning module, which receives a confirmation signal of accidental acceleration and alerts the driver to the accidental acceleration.

[0098] This embodiment also proposes a computer-readable storage medium, which can be at least one of ROM, RAM, magnetic disk, and optical disk. The computer-readable storage medium stores an accelerator pedal misoperation prevention program. When the accelerator pedal misoperation prevention program is executed by the processor, it implements the operation of the accelerator pedal misoperation prevention method, as follows:

[0099] The accelerator pedal parameter monitoring module monitors the data of the driver pressing the accelerator pedal in real time and transmits the monitored data to the parameter anomaly identification module. The parameter anomaly identification module uses a deep learning anomaly detection model to identify anomalies in the data of the driver pressing the accelerator pedal. The collision risk monitoring module monitors in real time whether there are obstacles on the vehicle's travel route and whether there is a collision risk between the vehicle and the obstacles.

[0100] The Accidental Accelerator Pedal Detection Module determines whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module then sends a signal confirming accidental accelerator pedal press to the control execution module.

[0101] The control execution module receives a confirmation signal that the accelerator pedal has been accidentally pressed and then suppresses the accelerator pedal, refraining from accelerating.

[0102] When the driver fully releases the accelerator or presses the brake pedal, the control function of the control execution module is deactivated.

[0103] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all possible implementations here.

Claims

1. A method for preventing accidental accelerator pedal press, characterized in that, Includes the following steps: The accelerator pedal parameter monitoring module monitors the data of the driver pressing the accelerator pedal in real time and transmits the monitored data to the parameter anomaly identification module. The data of the driver pressing the accelerator pedal includes the pressing force, the pedal opening, and the rate of change of the pedal opening. The rate of change of the pedal opening refers to the derivative of the pedal opening with respect to time. The parameter anomaly identification module uses a deep learning anomaly detection model to identify anomalies in the data of the driver pressing the accelerator pedal, including the following steps: Obtain a dataset of historical normal acceleration using the accelerator pedal; The dataset of historical acceleration with normal accelerator pedal movement is preprocessed. The specific preprocessing process is as follows: digital feature extraction is performed on the dataset to obtain the digital dimension data of the dataset. The digital dimension data is then transformed. The specific transformation process is as follows: regularization, class mapping, and function transformation are performed on the digital dimension data. The specific steps of the function transformation of the digital dimension data are: obtaining the distribution function of the digital dimension data, mapping the distribution function to a linear function, and obtaining the multi-dimensional feature data corresponding to the digital dimension data. A deep learning anomaly detection model is constructed and trained using a preprocessed dataset to obtain a trained deep learning anomaly detection model. During the training process of the deep learning anomaly detection model, gradient descent is used to update the network parameters of the deep learning anomaly detection model. The trained deep learning anomaly detection model is used to perform deep learning anomaly detection on the data of a driver pressing the accelerator pedal once, and the anomaly identification results of the data of a driver pressing the accelerator pedal once are obtained. Update the dataset with the results of identifying anomalies in the accelerator pedal data; The collision risk monitoring module monitors in real time whether there are obstacles in the vehicle's path and whether there is a risk of collision between the vehicle and the obstacles; The Accidental Accelerator Pedal Detection Module determines whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module then sends a signal confirming accidental accelerator pedal press to the control execution module. The control execution module receives a confirmation signal that the accelerator pedal has been accidentally pressed and then suppresses the accelerator pedal, refraining from accelerating. When the driver fully releases the accelerator or presses the brake pedal, the control function of the control execution module is deactivated.

2. The method for preventing accidental accelerator pedal press according to claim 1, characterized in that, It also includes the following steps: The interactive warning module receives a confirmation signal that the accelerator pedal has been accidentally pressed and alerts the driver.

3. The method for preventing accidental accelerator pedal press according to claim 1, characterized in that, If the estimated collision time between the vehicle and the obstacle is less than the collision time threshold, then there is a risk of collision between the vehicle and the obstacle; otherwise, there is no risk of collision between the vehicle and the obstacle.

4. A system for preventing accidental accelerator pedal press, characterized in that, The system includes: an accelerator pedal parameter monitoring module, a parameter anomaly identification module, a collision risk monitoring module, an accidental accelerator pedal judgment module, and a control execution module; The accelerator pedal parameter monitoring module is used to monitor the data of the driver pressing the accelerator pedal in real time and transmit the monitored data of the driver pressing the accelerator pedal to the parameter anomaly identification module. The parameter anomaly identification module is used to identify anomalies in the data of the driver pressing the accelerator pedal using a deep learning anomaly detection model. The collision risk monitoring module is used to monitor in real time whether there are obstacles in the vehicle's travel path and whether there is a risk of collision between the vehicle and the obstacle; The Accidental Accelerator Pedal Detection Module is used to determine whether the driver has accidentally pressed the accelerator pedal based on abnormal data of the driver pressing the accelerator pedal and whether there is a risk of collision between the vehicle and obstacles. The module sends a signal confirming accidental accelerator pedal pressing to the control execution module. The control execution module is used to receive a confirmation signal of accidental accelerator pedal press and suppress the accelerator pedal, and not respond to acceleration; when the driver completely releases the accelerator pedal or presses the brake pedal, the control function of the control execution module is deactivated.

5. The accelerator pedal anti-accidental pressing system according to claim 4, characterized in that, The parameter anomaly identification module includes a data feature extraction unit, a data feature input unit, an anomaly detection model unit, and a model parameter adjustment unit; The data feature extraction unit is used to extract multi-dimensional features from the acquired historical dataset of normal acceleration using the accelerator pedal, and obtain multi-dimensional feature data of the accelerator pedal pressing behavior. The data feature input unit is used to input the multi-dimensional feature data obtained from the processed data into the constructed deep learning anomaly detection model; The anomaly detection model unit is used to detect anomalies in multi-dimensional feature data and obtain the identification results of data anomalies in the accelerator pedal. The model parameter tuning unit is used to adjust the network parameters of the deep learning anomaly detection model during the training process.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a throttle pedal misoperation prevention program. When the throttle pedal misoperation prevention program is executed by the processor, it implements the operation of the throttle pedal misoperation prevention method as described in any one of claims 1-4.

Images