A pedal jamming monitoring method for braking intention recognition based on multi-source data fusion

Abstract

The present invention discloses a pedal jamming monitoring method for identifying braking intention based on multi-source data fusion, belonging to the field of automotive braking. S1: Synchronously collect the pressure applied to the brake pedal, the displacement of the brake pedal under the applied pressure, the displacement of the accelerator pedal, and the vehicle acceleration, and obtain the current driving environment information; S2: Calculate the pedal position change rate, the pedal force change rate, the throttle opening change rate, and the relative distance ratio between the current vehicle and the vehicle in front; S3: Use a random forest to generate a preliminary identification intention; S4: Generate a final identification intention; S5: Determine whether there is jamming. By adopting the above-mentioned pedal jamming monitoring method for identifying braking intention based on multi-source data fusion, through fusing multi-source data such as brake pedal displacement, pressure, throttle state, and environmental parameters, and combining mechanisms such as dual-model decision-making, dynamic threshold adjustment, and multi-rule weighted judgment, the accuracy, real-time performance, and environmental adaptability of brake pedal jamming monitoring are significantly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile braking, and in particular to a pedal sticking monitoring method based on multi-source data fusion and identification of braking intention. Background Art

[0002] As intelligent driving technology matures and stabilizes, more and more vehicles are equipped with intelligent driving features. Detecting a stuck pedal is now implemented in a variety of ways. Currently, most pedal position sensors monitor pedal position changes in real time and calculate whether the pedal is stuck based on a threshold. This approach offers advantages such as simplicity and low cost, making it suitable for most vehicle applications. However, if the controller fails to detect a stuck pedal, it can negatively impact the vehicle's braking performance and user experience.

[0003] Furthermore, using position sensors to determine whether the pedal is stuck requires high precision, reliability, and stability. If the sensor is damaged or its signal is interfered with, it will cause abnormal brake pedal input, affecting the vehicle's braking performance and even endangering the driver's safety. Summary of the Invention

[0004] The purpose of the present invention is to provide a pedal sticking monitoring method based on multi-source data fusion to identify braking intention, so as to solve the above technical problems.

[0005] To achieve the above objectives, the present invention provides a pedal sticking monitoring method for identifying braking intention based on multi-source data fusion, comprising the following steps:

[0006] S1. Synchronously collects the pressure applied to the brake pedal, the displacement of the brake pedal under pressure, the displacement of the accelerator pedal, and the acceleration of the vehicle, and obtains the current driving environment information based on the visual camera and lidar;

[0007] S2. Calculating a pedal position change rate, a pedal force change rate, a throttle opening change rate, and a relative distance ratio between the current vehicle and the preceding vehicle based on the collected pressure applied to the brake pedal and the displacement of the brake pedal under the applied pressure;

[0008] S3: Generate preliminary recognition intent using random forests based on brake pedal displacement, brake pedal pressure, vehicle acceleration, throttle opening change rate, relative distance ratio between the current vehicle and the preceding vehicle, pedal force change rate, and current driving environment information;

[0009] S4. Input the preliminary recognition intent into the neural network, optimize the preliminary recognition intent using the neural network classification cross entropy function, and generate the final recognition intent;

[0010] S5. Determine whether there is jamming based on the brake pedal displacement, accelerator pedal displacement, accelerator opening change rate, and the finally recognized intention.

[0011] Preferably, the current driving environment information described in step S1 includes lane lines, traffic signs, signal lights, the distance information between the current vehicle and the vehicle ahead, the relative speed information between the current vehicle and the vehicle ahead, and the road gradient information.

[0012] Preferably, the pedal position change rate described in step S2 , the pedal force change rate , the accelerator opening change rate and the relative distance ratio between the current vehicle and the vehicle ahead are calculated as follows:

[0013] (1);

[0014] (2);

[0015] (3);

[0016] (4);

[0017] In the formula, represents the displacement of the brake pedal from the initial position to the current position during the time period ; represents the pressure applied by the driver on the brake pedal during the time period ; represents the displacement of the accelerator pedal from the initial position to the current position during the time period ; represents the minimum value; represents the distance between the current vehicle and the vehicle ahead; represents the relative speed between the current vehicle and the vehicle ahead.

[0018] Preferably, in step S3, input , which represents the vehicle acceleration, into the random forest; the output categories are 0, 1, and 2, where 0 represents no braking, 1 represents normal braking, and 2 represents emergency braking.

[0019] Preferably, the neural network classification cross-entropy function described in step S4 has the following expression:

[0020] (5);

[0021] In the formula, represents the true label of the sample; represents that the model predicts that the sample belongs to the Probability of the class;

[0022] and the final recognition intention generated by the neural network , represents the fusion weight, and respectively represent the random forest prediction probability and the neural network prediction probability.

[0023] Preferably, in step S5, if any of the following situations exist, it is determined that there is a jam:

[0024] Situation 1: If there are both brake pedal displacement and accelerator pedal displacement, it is determined as a jam;

[0025] Situation 2: If the change rate of the throttle opening is opposite to the expected movement direction, a weight of 0.4 is added;

[0026] If the growth rate of the brake master cylinder pressure is less than the set threshold, a weight of 0.4 is added;

[0027] If there is an abnormal fluctuation in the current of the brake electronic pedal motor, a weight of 0.2 is added;

[0028] If the sum of the weights added above is greater than 0.6, it is determined that there is a jam;

[0029] Situation 3: If there is , it is determined as a jam;

[0030] wherein, represents the adjusted threshold of the change rate of the brake pedal position; represents the adjusted threshold of the change rate of the pedal force; and both represent the environmental correction coefficient; represents and;

[0031] Value range: ; and for dry and smooth road surface: ; for wet and slippery road surface: ; for ice and snow road surface: ; for gravel road surface: ;

[0032] Value range: ; and for normal driving: ; for emergency braking: ; for heavy-duty vehicles: .

[0033] Preferably, in step S5, the thresholds of the change rate of the brake pedal position and the change rate of the pedal force are adjusted according to the current driving environment information:

[0034] (6);

[0035] (7);

[0036] In the formula, Adjustment variation.

[0037] Preferably, after step S5, step S6 is further included. When it is determined that the brake pedal is stuck, the stuck signal is transmitted to the vehicle controller through the CAN signal, and the vehicle controller prompts the driver that the pedal is stuck through voice or instrument, so as to realize early warning.

[0038] Therefore, the present invention adopts the above-mentioned pedal sticking monitoring method based on multi-source data fusion to identify braking intention, and has the following beneficial effects:

[0039] 1. Multi-source data fusion improves accuracy: Integrate multi-source data such as brake pedal displacement, pressure, throttle state, and environmental parameters to comprehensively capture braking-related information, avoid the limitations of single data, and improve the accuracy of braking intention recognition and sticking monitoring;

[0040] 2. Dual-model decision-making enhances reliability: Adopt the dual models of random forest and neural network for processing, combine the advantages of both, and make a decision through result fusion to improve the reliability and robustness of braking intention recognition under complex working conditions;

[0041] 3. Dynamic threshold adapts to complex working conditions: Dynamically adjust thresholds such as displacement rate and force change rate based on environmental perception to adapt to different driving scenarios (such as emergency braking, bumpy road surface), reduce misjudgment and missed judgment, and improve the adaptability of sticking detection;

[0042] 4. Multi-rule weighted judgment for comprehensive detection: Through multi-parameter weighted analysis, kinematic threshold matching and other multi-rule fusions, comprehensively identify pedal sticking from different dimensions (such as pressure change, motor state), and improve the detection coverage rate.

[0043] Next, through the drawings and embodiments, the technical solution of the present invention will be further described in detail. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flowchart of a pedal sticking monitoring method based on multi-source data fusion to identify braking intention of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] In order to make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer and more understandable, the embodiments of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not used to limit the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of this application. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout.

[0046] It should be noted that the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0047] The embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0048] As Figure 1 shown, a pedal jamming monitoring method for identifying braking intent based on multi-source data fusion includes the following steps:

[0049] S1. Synchronously collect the pressure applied to the brake pedal, the displacement of the brake pedal under the applied pressure, the displacement of the accelerator pedal, and the vehicle acceleration, and obtain the current driving environment information based on a vision camera and a lidar.

[0050] In this embodiment, the electronic service brake collects the brake pedal position information once every 5 ms, and at the same time collects the pressure information at the pedal end once every 1 ms.

[0051] The current driving environment information described in step S1 includes lane lines, traffic signs, signal lights, the distance information between the current vehicle and the vehicle in front, the relative speed information between the current vehicle and the vehicle in front, and the road gradient information.

[0052] S2. Calculate the pedal position change rate, the pedal force change rate, the throttle opening change rate, and the relative distance ratio between the current vehicle and the vehicle in front based on the collected pressure applied to the brake pedal and the displacement of the brake pedal under the applied pressure.

[0053] The pedal position change rate , the pedal force change rate , the throttle opening change rate , and the relative distance ratio between the current vehicle and the vehicle in front are calculated according to the following formulas:

[0054] (1);

[0055] (2);

[0056] (3);

[0057] (4);

[0058] In the formula, represents the displacement of the brake pedal from the initial position to the current position within the time period; within the time period represents the pressure applied by the driver on the brake pedal; within the time period represents the displacement of the accelerator pedal from the initial position to the current position within the time period; within the time period represents the minimum value; represents the distance between the current vehicle and the vehicle in front; represents the relative speed between the current vehicle and the vehicle in front.

[0059] S3. Based on the brake pedal displacement, brake pedal pressure, vehicle acceleration, accelerator opening change rate, relative distance ratio between the current vehicle and the vehicle in front, pedal force change rate, and current driving environment information, use a random forest to generate a preliminary recognition intention;

[0060] In step S3, input , [[ID={45]] represents the vehicle acceleration; the output categories are 0, 1, 2, where 0 represents no braking, 1 represents normal braking, and 2 represents emergency braking.

[0061] S4. Input the preliminary recognition intention into a neural network, and use the neural network classification cross-entropy function to optimize the preliminary recognition intention to generate a final recognition intention;

[0062] The neural network classification cross-entropy function described in step S4 has the following expression:

[0063] (5);

[0064] In the formula, represents the true label of the sample; represents the probability that the model predicts that the sample belongs to the th class;

[0065] and the final recognition intention generated by the neural network , represents the fusion weight, [[ID=7s]] and Represents the random forest prediction probability and the neural network prediction probability respectively. The fusion weight is based on the vehicle itself, with more emphasis on or .

[0066] S5. Determine whether there is a jam based on the brake pedal displacement, the accelerator pedal displacement, the accelerator opening change rate, and the final recognition intention.

[0067] In step S5, if the following conditions exist, it is determined that a jam occurs:

[0068] Case 1: If both the brake pedal and the accelerator pedal are displaced, it is considered stuck.

[0069] Case 2: If the throttle opening change rate is opposite to the expected movement direction, add a weight of 0.4;

[0070] If the brake master cylinder pressure growth rate is less than the set threshold, add a weight of 0.4;

[0071] If there is abnormal fluctuation in the brake electronic pedal motor current, add a weight of 0.2;

[0072] If the sum of the weights added above is greater than 0.6, it is determined that there is a stuck state;

[0073] Case 3: If exists , it is judged to be stuck;

[0074] in, Indicates the adjusted brake pedal position change rate threshold; represents the adjusted pedal force change rate threshold; and Both represent environmental correction factors; Indicates and; Value range: ; and dry and smooth road surface: ; Slippery road: ; Icy and snowy roads: Gravel road surface: ; Value range: ; and driving normally: ; Emergency brake: ; Heavy-load vehicles: . It can be seen that in case 3, when When , it indicates that the pedal displacement rate is abnormally low and there may be a jam; when , it indicates that the driver has stepped on the pedal hard but the displacement has not increased synchronously (force increases but displacement is slow), which may indicate a jam.

[0075] Preferably, in step S5, the threshold values of the change rate of the brake pedal position and the change rate of the pedal force are adjusted according to the current driving environment information:

[0076] (6);

[0077] (7);

[0078] Wherein, Adjustment variation. In this embodiment, the threshold value can be set to increase by +0.1 on congested roads to reduce the system alarm condition; in the case of heavy load, the threshold value is -0.1 to improve the system sensitivity.

[0079] Preferably, after step S5, there is also step S6. When it is determined that the brake pedal is stuck, the stuck signal is transmitted to the vehicle controller through the CAN signal, and the vehicle controller prompts the driver that the pedal is stuck through voice or instrument to achieve early warning.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A pedal jamming monitoring method for identifying braking intention based on multi-source data fusion, characterized in that: It includes the following steps: S1. Synchronously collect the pressure applied to the brake pedal, the displacement of the brake pedal under the applied pressure, the displacement of the accelerator pedal, and the vehicle acceleration, and obtain the current driving environment information based on a vision camera and a lidar; S2. Calculate the pedal position change rate, the pedal force change rate, the throttle opening change rate, and the relative distance ratio between the current vehicle and the vehicle ahead based on the collected pressure applied to the brake pedal and the displacement of the brake pedal under the applied pressure; S3. Generate a preliminary recognition intention using a random forest based on the brake pedal displacement, the brake pedal pressure, the vehicle acceleration, the throttle opening change rate, the relative distance ratio between the current vehicle and the vehicle ahead, the pedal force change rate, and the current driving environment information; S4. Input the preliminary recognition intention into a neural network, and optimize the preliminary recognition intention using the neural network classification cross-entropy function to generate a final recognition intention; The neural network classification cross-entropy function described in step S4 The expression is as follows: （5）； wherein, represents the true label of the sample; represents the probability that the model predicts that the sample belongs to the th class; and the final recognition intention generated by the neural network , represents the fusion weight, and respectively represent the random forest prediction probability and the neural network prediction probability; S5. Judge whether there is jamming based on the brake pedal displacement, the accelerator pedal displacement, the throttle opening change rate, and the final recognition intention; In step S5, if any of the following situations occur, it is judged that there is jamming: Situation 1: If there are both the brake pedal displacement and the accelerator pedal displacement, it is judged as jamming; Situation 2: If the throttle opening change rate is opposite to the expected movement direction, a weight of 0.4 is added; If the growth rate of the brake master cylinder pressure is less than the set threshold, a weight of 0.4 is added; If there is abnormal fluctuation in the current of the brake electronic pedal motor, a weight of 0.2 is added; If the sum of the above added weights is greater than 0.6, it is determined that there is jamming; Case 3: If there exists , it is judged as stuck; Among them, represents the adjusted threshold value of the braking pedal position change rate; represents the adjusted threshold value of the pedal force change rate; and both represent the environmental correction coefficient; represents "and"; represents the pedal position change rate; represents the pedal force change rate; Value range: ∈[0.5, 1.5]; and for dry and smooth road surface: = 1.0; for wet and slippery road surface: = 0.7 - 0.9; for icy road surface: = 0.5 - 0.6; for gravel road surface: = 1.2 - 1.5; Value range: ∈[0.8, 2.0]; and for normal driving: = 1; Emergency braking: = 1.5 - 2.0; For heavy-duty vehicles: = 0.8 - 1.

0.

2. The pedal jamming monitoring method for identifying braking intention based on multi-source data fusion according to claim 1, wherein: The current driving environment information described in step S1 includes lane lines, traffic signs, signal lights, the distance information between the current vehicle and the vehicle ahead, the relative speed information between the current vehicle and the vehicle ahead, and the road slope information.

3. A pedal jamming monitoring method for identifying braking intention based on multi-source data fusion according to claim 1, characterized in that: The pedal position change rate described in step S2 , the pedal force change rate , the throttle opening change rate and the relative distance ratio between the current vehicle and the vehicle ahead are calculated as follows: （1）； （2）； （3）； （4）； Wherein, represents the displacement of the brake pedal from the initial position to the current position within the time period ; represents the pressure applied by the driver to the brake pedal within the time period ; represents the displacement of the accelerator pedal from the initial position to the current position within the time period ; represents the minimum value; represents the distance between the current vehicle and the vehicle ahead; represents the relative speed between the current vehicle and the vehicle ahead.

4. A pedal jamming monitoring method for identifying braking intention based on multi-source data fusion according to claim 3, characterized in that: In step S3, input into the random forest , representing the vehicle acceleration; the output categories are 0, 1, and 2, where 0 represents no braking, 1 represents normal braking, and 2 represents emergency braking.

5. A pedal jamming monitoring method for identifying braking intention based on multi-source data fusion according to claim 1, characterized in that: In step S5, adjust the brake pedal position change rate threshold and the pedal force change rate threshold according to the current driving environment information; （6）； （7）； In the formula, Adjustment variation.

6. The pedal jamming monitoring method for identifying braking intention based on multi-source data fusion according to claim 1, wherein: After step S5, there is also step S6. When it is judged that the brake pedal is jammed, transmit the jamming signal to the vehicle controller through the CAN signal, and the vehicle controller prompts the driver of the pedal jamming through voice or instrument to achieve early warning.

Images