An intelligent anti-lock braking system for electric vehicles considering driver characteristics

Abstract

The application discloses a kind of electric vehicle intelligent anti-lock braking system considering driver characteristics, including comprehensive information acquisition module, driver characteristic extraction module, anti-lock braking trigger threshold adjustment module, automobile brake boost ratio adjustment module, single wheel electro-hydraulic braking torque distribution mode selection module and anti-lock braking system safety guarantee module.Comprehensive information acquisition module is used to obtain driver, vehicle and road environment information;Driver characteristic extraction module is used to extract the characteristics of driver tension level, fatigue degree and decision-making ability;According to the characteristics of the driver, the anti-lock braking trigger threshold adjustment module and the automobile brake boost ratio adjustment module calculate the adjusted trigger threshold of the anti-lock braking and the adjusted brake boost ratio of the automobile respectively;Single wheel electro-hydraulic braking torque distribution mode selection module is used to determine the source of the braking torque applied to a single wheel;Anti-lock braking system safety guarantee module is used to improve the safety and reliability of the whole system.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of automobile safety, and relates to an active safety control technology for an electric vehicle, in particular to an intelligent anti-lock braking system for an electric vehicle considering driver characteristics. BACKGROUND

[0002] With the rapid development of electric vehicle technology, automobile active safety and driving experience have become the focus of attention. The traditional anti-lock braking system can effectively prevent wheel lock in most cases and ensure the driving stability of the vehicle in emergency braking. However, the existing anti-lock braking system mainly controls braking based on the dynamics parameters of the vehicle and the road conditions, and fails to fully consider the characteristics and state of the driver.

[0003] In actual driving process, the nervousness, fatigue state and decision-making ability of the driver have an important influence on the braking effect. A nervous driver may excessively press the brake pedal, while a fatigued driver may be slow to react, and these factors may affect the braking performance and safety of the vehicle. However, the traditional anti-lock braking system relying only on vehicle and environmental information has certain limitations and cannot dynamically adjust the braking control strategy to adapt to the characteristics of different drivers. Therefore, it is urgent to develop an anti-lock braking system considering the characteristics of the driver, which can dynamically optimize the braking control strategy according to the characteristics of the driver, thereby improving the control effect of the anti-lock braking system. SUMMARY

[0004] The purpose of the present application is to provide an intelligent anti-lock braking system for an electric vehicle considering the characteristics of the driver, to solve the problems faced in the background art.

[0005] In order to achieve the above-mentioned purpose, the present application provides the following scheme:

[0006] An intelligent anti-lock braking system for an electric vehicle considering the characteristics of the driver, comprising a comprehensive information acquisition module, a driver characteristic extraction module, an anti-lock braking trigger threshold adjustment module, an automobile braking assist ratio adjustment module, a single-wheel electro-hydraulic braking torque distribution mode selection module and an anti-lock braking system safety guarantee module;

[0007] The comprehensive information acquisition module is used for acquiring driver basic information, vehicle state information and road environment information; the driver basic information includes driver heart rate, driver heart rate variability, driver skin conductance, driver electroencephalogram beta, theta and gamma band power and driver blink frequency; the vehicle state information includes vehicle type and vehicle speed; and the road environment information includes road type, road surface adhesion coefficient, visibility and illumination intensity;

[0008] The driver feature extraction module is configured to calculate a driver tension level representation factor, a driver fatigue degree representation factor, and a driver decision-making ability representation factor.

[0009] The anti-lock braking trigger threshold adjustment module is configured to calculate an anti-lock braking adjusted trigger threshold according to the driver tension level representation factor, the driver fatigue degree representation factor, the driver decision-making ability representation factor, and an anti-lock braking initial trigger threshold, wherein the anti-lock braking initial trigger threshold refers to an ideal wheel slip rate of a road currently traveled by the vehicle, and the anti-lock braking adjusted trigger threshold is calculated according to the following formula:

[0010]

[0011] In the formula, S ABS is the anti-lock braking adjusted trigger threshold; S ideal is the anti-lock braking initial trigger threshold, which is equal to the ideal wheel slip rate of the road currently traveled by the vehicle; f T is the driver tension level representation factor; f F is the driver fatigue degree representation factor; and f D is the driver decision-making ability representation factor.

[0012] The automobile brake assist ratio adjustment module is configured to determine an automobile adjusted brake assist ratio according to the driver tension level representation factor, the driver fatigue degree representation factor, the driver decision-making ability representation factor, and an automobile standard brake assist ratio by using a deep neural network algorithm, wherein the deep neural network algorithm is designed according to the following structure:

[0013] The input layer includes four nodes corresponding to the driver tension level representation factor, the driver fatigue degree representation factor, the driver decision-making ability representation factor, and the automobile standard brake assist ratio; the hidden layer includes three layers, i.e., a first hidden layer, a second hidden layer, and a third hidden layer; the first hidden layer, the second hidden layer, and the third hidden layer include 64, 32, and 16 neurons, respectively; and the output layer includes one node, which outputs the automobile adjusted brake assist ratio.

[0014] The calculation formula of the automobile adjusted brake assist ratio is as follows:

[0015]

[0016] In the formula, f T is the driver tension level representation factor; f F is the driver fatigue degree representation factor; f D is the driver decision-making ability representation factor; r standard is the automobile standard brake assist ratio, which is determined by a manufacturer of a brake booster through experiments; and r adjustThe automobile adjustment rear brake assist ratio is adjusted; h1, h2 and h3 are respectively the output of the first hidden layer, the output of the second hidden layer and the output of the third hidden layer; ReLU is an activation function, and is specifically defined as ReLU(x) = max(0, x), wherein x is the input of the activation function; W1, W2, W3 and W4 are respectively a weight matrix from the input layer to the first hidden layer, a weight matrix from the first hidden layer to the second hidden layer, a weight matrix from the second hidden layer to the third hidden layer and a weight matrix from the third hidden layer to the output layer, and W1, W2, W3 and W4 are determined through a training process; b1, b2, b3 and b4 are respectively a bias term of the first hidden layer, a bias term of the second hidden layer, a bias term of the third hidden layer and a bias term of the output layer, and b1, b2, b3 and b4 are determined through the training process; and α and β are adjustment coefficients obtained through the training process.

[0017] The single-wheel electro-hydraulic braking torque distribution mode selection module determines the source of the braking torque applied to a single wheel according to the driver torque distribution mode influence factor and the road adhesion coefficient, and in combination with a torque distribution mode selection rule.

[0018] The anti-lock braking system safety guarantee module guarantees the safety and reliability of the intelligent anti-lock braking system of the electric vehicle considering the driver characteristics according to the multi-stage anti-lock braking safety trigger threshold and the multi-stage automobile braking safety assist ratio.

[0019] The driver tension level representation factor is obtained through weighted calculation according to a driver tension level physiological feedback index, a vehicle driving state index, a vehicle internal and external operating environment index and a driving trip requirement index.

[0020] The calculation formula of the driver tension level physiological feedback index is:

[0021]

[0022] In the formula, D Tp is the driver tension level physiological feedback index; h r is the driver heart rate; is the driver resting heart rate; is the driver maximum heart rate; c s is the driver skin conductance; is the skin conductance of the driver in a resting state; is the maximum skin conductance of the driver in a resting state; P β is the driver electroencephalogram beta band power; is the maximum driver electroencephalogram beta band power; α1, α2 and α3 are respectively a driver heart rate weight coefficient, a driver skin conductance weight coefficient and a driver electroencephalogram beta band power weight coefficient, and the specific values are determined through simulation and test.

[0023] The calculation formula of the vehicle driving state index is:

[0024]

[0025] In the formula, V s is the vehicle driving state index; v is the vehicle speed; F is the automobile fault state, F = 0 indicates that the automobile is not in fault, F = 0.4 indicates that the driving motor of the automobile is in fault, F = 0.6 indicates that the hydraulic braking system of the automobile is in fault, and F = 1 indicates that both the driving motor and the hydraulic braking system of the automobile are in fault; T v is the electric automobile type representation, T v = 0.2 indicates an electric car, T v = 0.4 indicates an electric bus, T v = 0.6 indicates an electric truck, and T v = 1 indicates an electric semi-trailer; β1 and β2 are weight coefficients, and the specific values are determined through simulation and test;

[0026] The calculation formula of the vehicle internal and external operating environment index is:

[0027] V e = γ1·c r + γ2·c t + γ3·c w + γ4·b p

[0028] In the formula, V e is the vehicle internal and external operating environment index; c r is the road condition, c r = 0.4 indicates a rural road, c r = 0.7 indicates an urban road, and c r = 1 indicates a highway; c t is the traffic condition, c t = 1 indicates that the vehicle is parked due to traffic congestion, c t = 0.5 indicates that the vehicle is parked due to a red traffic signal, c t = 0 indicates that the vehicle is not parked and is in a running state; c w is the weather condition, c w = 0 indicates sunny or cloudy weather, c w = 0.5 indicates rainy weather, and c w = 1 indicates snowy weather; b p is the volume of the voice of the passengers in the vehicle, b p = 0.2 indicates that the volume is between 60-70 decibels, b p = -0.6 indicates that the volume is between 70-85 decibels, and b p=1 represents that the volume is at 85-100 decibels; γ1, γ2, γ3 and γ4 are weight coefficients, and specific values are determined through simulation and test;

[0029] The calculation formula of the driving trip requirement index is:

[0030]

[0031] In the formula, D j is the driving trip requirement index; n is the historical number of times of the driver going to the trip requirement destination, n=0 represents that the historical number of times of the driver going to the trip requirement destination is 0, n=0.2 represents that the historical number of times of the driver going to the trip requirement destination is 1-3, n=0.6 represents that the historical number of times of the driver going to the trip requirement destination is 4-6, and n=1 represents that the historical number of times of the driver going to the trip requirement destination is more than 6; t is the time limit of the driver going to the trip requirement destination, t=0.2 represents that the time limit is within 15 minutes, t=0.4 represents that the time limit is 15-30 minutes, t=0.6 represents that the time limit is 30-60 minutes, and t=1 represents that the time limit is more than 60 minutes; δ1 and δ2 are weight coefficients, and specific values are determined through simulation and test;

[0032] The calculation formula of the driver tension level representation factor is:

[0033]

[0034] In the formula, f T is the driver tension level representation factor; f T0 is the driver tension level non-normalized initial value; D Tp is the driver tension level physiological feedback index; V s is the vehicle driving state index; V e is the vehicle internal and external operating environment index; D j is the driving trip requirement index; ε1, ε2 and ε3 are weight coefficients, and specific values are determined through simulation and test.

[0035] The driver fatigue degree representation factor is obtained through weighted calculation according to the driver fatigue degree physiological feedback index, the driver operation feedback index and the driving environment influence index;

[0036] The calculation formula of the driver fatigue degree physiological feedback index is:

[0037]

[0038] In the formula, D Fp is the driver fatigue degree physiological feedback index; h rv is the driver heart rate variability; the heart rate variability of the driver at rest; the maximum heart rate variability of the driver;b f the blink frequency of the driver; the average blink frequency of the driver; the maximum blink frequency of the driver;P θ the electroencephalogram theta band power of the driver; the maximum electroencephalogram theta band power of the driver;λ1, λ2 and λ3 are respectively the heart rate variability weight coefficient of the driver, the blink frequency weight coefficient of the driver and the electroencephalogram theta band power weight coefficient of the driver, and the specific values are determined through simulation and test;

[0039] The calculation formula of the driver operation feedback index is:

[0040] D o = σ1·F w + σ2·F b + σ3·F a

[0041] In the formula, D o is the driver operation feedback index; F w is the steering wheel angle adjustment frequency; F b is the brake pedal opening change frequency; F a is the accelerator pedal opening change frequency; σ1, σ2 and σ3 are weight coefficients, and the specific values are determined through simulation and test.

[0042] The calculation formula of the driving environment influence index is:

[0043]

[0044] In the formula, E is the driving environment influence index; T d is the continuous driving time of the driver; V is the visibility of the driving environment; and L is the illumination intensity of the driving environment.

[0045] The calculation formula of the driver fatigue degree representation factor is:

[0046]

[0047] In the formula, f F is the driver fatigue degree representation factor; f F0 is the unnormalized initial value of the driver fatigue degree; D Fp is the driver fatigue degree physiological feedback index; D o is the driver operation feedback index; E is the driving environment influence index; κ1, κ2 and κ3 are weight coefficients, and the specific values are determined through simulation and test.

[0048] The driver decision-making ability characteristic factor is calculated according to driver electroencephalogram gamma band power, driver braking reaction time, the sum of the time ratio of the nearest vehicle in front of the driver's gaze to the traffic signal, and the number of lane deviations, and the specific calculation formula is:

[0049]

[0050] In the formula, f D is the driver decision-making ability characteristic factor; f D0 is the non-normalized initial value of the driver decision-making ability; P γ is the driver electroencephalogram gamma band power; t db is the driver braking reaction time; P t is the sum of the time ratio of the nearest vehicle in front of the driver's gaze to the traffic signal; n d is the number of lane deviations; and ξ1, ξ2, ξ3 and ξ4 are weight coefficients, and the specific values are determined through simulation and test.

[0051] The calculation formula of the driver torque distribution mode influence factor is:

[0052]

[0053] In the formula, I driver is the driver torque distribution mode influence factor; f T is the driver tension level characteristic factor; f D is the driver decision-making ability characteristic factor; and ω1, ω2 and ω3 are weight coefficients, and the specific values are determined through simulation and test.

[0054] The torque distribution mode selection rule is specifically:

[0055] When the calculation value of the driver torque distribution mode influence factor is less than 0.3, and the road surface adhesion coefficient is not greater than 0.2, only the motor braking torque is applied to the single wheel by the drive motor;

[0056] When the calculation value of the driver torque distribution mode influence factor is less than 0.3, and the road surface adhesion coefficient is greater than 0.2, the motor braking torque is preferentially applied to the single wheel by the drive motor, and when the motor braking torque cannot meet the braking demand, the hydraulic braking torque is applied by the hydraulic braking system to supplement the insufficient part;

[0057] When the calculation value of the driver torque distribution mode influence factor is between 0.3 and 0.6, the motor braking torque is preferentially applied to the single wheel by the drive motor, and when the motor braking torque cannot meet the braking demand, the hydraulic braking torque is applied by the hydraulic braking system to supplement the insufficient part;

[0058] When the calculation value of the driver torque distribution mode influence factor is greater than 0.6, only the hydraulic brake system is used to apply a hydraulic brake torque to the single wheel.

[0059] The multi-level anti-lock braking safety trigger threshold is specifically divided into an anti-lock braking maximum trigger threshold and an anti-lock braking minimum trigger threshold; when the calculation value of the anti-lock braking adjusted trigger threshold is greater than the anti-lock braking maximum trigger threshold, the calculation value of the anti-lock braking adjusted trigger threshold is set as the anti-lock braking maximum trigger threshold; when the calculation value of the anti-lock braking adjusted trigger threshold is less than the anti-lock braking minimum trigger threshold, the calculation value of the anti-lock braking adjusted trigger threshold is set as the anti-lock braking minimum trigger threshold.

[0060] The multi-level automobile braking safety boost ratio is specifically divided into an automobile braking maximum boost ratio and an automobile braking minimum boost ratio; when the calculation value of the automobile adjusted braking boost ratio is greater than the automobile braking maximum boost ratio, the calculation value of the automobile adjusted braking boost ratio is set as the automobile braking maximum boost ratio; when the calculation value of the automobile adjusted braking boost ratio is less than the automobile braking minimum boost ratio, the calculation value of the automobile adjusted braking boost ratio is set as the automobile braking minimum boost ratio.

[0061] The beneficial effects of the present application are:

[0062] 1. The present application can acquire and analyze the tension degree, fatigue state and decision-making ability of the driver in real time, so as to dynamically adjust the trigger threshold and braking control strategy of the anti-lock braking system according to the actual state of the driver, and realize individualized braking control.

[0063] 2. The present application can adjust the anti-lock braking trigger threshold according to the driver characteristics, so that the anti-lock braking system can be triggered at the right time, effectively avoiding the problems of too early or too late braking, and improving the accuracy and response speed of the braking system.

[0064] 3. The present application can dynamically determine the braking torque source of the single wheel according to the driver characteristics and the current vehicle driving environment, optimize the distribution of braking force, and improve the braking stability and safety of the vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0065] The present application will be further described below in conjunction with the drawings:

[0066] Figure 1 It is a framework diagram of the intelligent anti-lock braking system of the electric vehicle considering the driver characteristics. DETAILED DESCRIPTION

[0067] The present application will be further described below in conjunction with the drawings and specific embodiments.

[0068] ReferenceFigure 1 The application discloses an intelligent anti-lock braking system for an electric vehicle considering driver characteristics, which comprises a comprehensive information acquisition module, a driver characteristic extraction module, an anti-lock braking trigger threshold adjustment module, an automobile braking assist ratio adjustment module, a single-wheel electro-hydraulic braking torque distribution mode selection module and an anti-lock braking system safety guarantee module.

[0069] The comprehensive information acquisition module is used for acquiring driver basic information, vehicle state information and road environment information; the driver basic information comprises a driver heart rate, a driver heart rate variability, a driver skin conductivity, a driver electroencephalogram beta, theta and gamma band power and a driver blink frequency; the vehicle state information comprises a vehicle type and a vehicle speed; and the road environment information comprises a road type, a road surface adhesion coefficient, a visibility and an illumination intensity.

[0070] The driver characteristic extraction module is used for calculating a driver tension level representation factor, a driver fatigue degree representation factor and a driver decision-making ability representation factor.

[0071] The anti-lock braking trigger threshold adjustment module is used for calculating an anti-lock braking adjusted trigger threshold according to the driver tension level representation factor, the driver fatigue degree representation factor, the driver decision-making ability representation factor and an anti-lock braking initial trigger threshold; the anti-lock braking initial trigger threshold is an ideal wheel slip rate of a road currently traveled by the vehicle; and the anti-lock braking adjusted trigger threshold is specifically calculated according to the following formula:

[0072]

[0073] In the formula, S ABS is the anti-lock braking adjusted trigger threshold; S ideal is the anti-lock braking initial trigger threshold, which is equal to the ideal wheel slip rate of the road currently traveled by the vehicle; f T is the driver tension level representation factor; f F is the driver fatigue degree representation factor; f D is the driver decision-making ability representation factor.

[0074] The automobile braking assist ratio adjustment module is used for determining an automobile adjusted braking assist ratio according to the driver tension level representation factor, the driver fatigue degree representation factor, the driver decision-making ability representation factor and an automobile standard braking assist ratio by using a deep neural network algorithm; and the structure design of the deep neural network algorithm is specifically as follows:

[0075] The input layer includes 4 nodes corresponding to the driver tension level representation factor, the driver fatigue degree representation factor, the driver decision-making ability representation factor and the automobile standard braking assist force ratio respectively; the hidden layer adopts a 3-layer structure including a first hidden layer, a second hidden layer and a third hidden layer; the first hidden layer, the second hidden layer and the third hidden layer include 64, 32 and 16 neurons respectively; the output layer includes one node, and the output is the automobile adjusted braking assist force ratio;

[0076] The calculation formula of the automobile adjusted braking assist force ratio is as follows:

[0077]

[0078] In the formula, f T is the driver tension level representation factor; f F is the driver fatigue degree representation factor; f D is the driver decision-making ability representation factor; r standard is the automobile standard braking assist force ratio, which is determined by the manufacturer of the brake booster through experiments; r adjust is the automobile adjusted braking assist force ratio; h1, h2 and h3 are the output of the first hidden layer, the output of the second hidden layer and the output of the third hidden layer respectively; ReLU is an activation function, and the specific definition is ReLU(x) = max(0, x), wherein x is the input of the activation function; W1, W2, W3 and W4 are the weight matrix from the input layer to the first hidden layer, the weight matrix from the first hidden layer to the second hidden layer, the weight matrix from the second hidden layer to the third hidden layer and the weight matrix from the third hidden layer to the output layer respectively, and W1, W2, W3 and W4 are determined through a training process; b1, b2, b3 and b4 are the bias term of the first hidden layer, the bias term of the second hidden layer, the bias term of the third hidden layer and the bias term of the output layer respectively, and b1, b2, b3 and b4 are determined through the training process; and α and β are adjustment coefficients obtained through the training process;

[0079] The single-wheel electro-hydraulic braking torque distribution mode selection module determines the source of the braking torque applied to a single wheel according to the driver torque distribution mode influence factor and the road adhesion coefficient, and in combination with the torque distribution mode selection rule;

[0080] The anti-lock braking system safety guarantee module guarantees the safety and reliability of the anti-lock braking system of the electric vehicle according to the multi-stage anti-lock braking safety trigger threshold and the multi-stage automobile braking safety assist force ratio.

[0081] The driver tension level representation factor is obtained through weighted calculation according to the driver tension level physiological feedback index, the vehicle driving state index, the vehicle internal and external operating environment index and the driving trip requirement index.

[0082] The calculation formula of the driver tension level physiological feedback index is:

[0083]

[0084] In the formula, D Tp is the driver tension level physiological feedback index; h r is the driver heart rate; is the driver resting heart rate; is the driver maximum heart rate; c s is the driver skin conductance; is the skin conductance of the driver in a resting state; is the maximum skin conductance of the driver; P β is the driver electroencephalogram beta band power; is the maximum driver electroencephalogram beta band power; α1, α2 and α3 are respectively the driver heart rate weight coefficient, the driver skin conductance weight coefficient and the driver electroencephalogram beta band power weight coefficient, and the specific values are determined through simulation and test;

[0085] The calculation formula of the vehicle driving state index is:

[0086]

[0087] In the formula, V s is the vehicle driving state index; v is the vehicle speed; F is the automobile fault state, F = 0 indicates that the automobile has no fault, F = 0.4 indicates that the driving motor of the automobile has a fault, F = 0.6 indicates that the hydraulic braking system of the automobile has a fault, and F = 1 indicates that both the driving motor and the hydraulic braking system of the automobile have faults; T v is the electric vehicle type representation, T v = 0.2 indicates an electric car, T v = 0.4 indicates an electric bus, T v = 0.6 indicates an electric truck, and T v = 1 indicates an electric semi-trailer; β1 and β2 are weight coefficients, and the specific values are determined through simulation and test;

[0088] The calculation formula of the vehicle internal and external operating environment index is:

[0089] V e = γ1·c r + γ2·c t + γ3·c w + γ4·b p

[0090] In the formula, V e is the vehicle internal and external operating environment index; c rfor road condition, c r = 0.4 represents a rural road, c r = 0.7 represents an urban road, c r = 1 represents a highway; c t for traffic condition, c t = 1 represents that the vehicle is parked due to traffic congestion, c t = 0.5 represents that the vehicle is parked due to a red traffic light, c t = 0 represents that the vehicle is not parked and is in a traffic state; c w for weather condition, c w = 0 represents sunny or cloudy weather, c w = 0.5 represents rainy weather, c w = 1 represents snowy weather; b p for the volume of the voice of the passenger in the vehicle, b p = 0.2 represents that the volume is at 60-70 decibels, b p = 0.6 represents that the volume is at 70-85 decibels, b p = 1 represents that the volume is at 85-100 decibels; γ1, γ2, γ3 and γ4 are weight coefficients, and specific values are determined through simulation and test;

[0091] The calculation formula of the driving trip requirement index is:

[0092]

[0093] In the formula, D j is the driving trip requirement index; n is the historical number of times that the driver goes to the trip requirement destination, n = 0 represents that the historical number of times that the driver goes to the trip requirement destination is 0 times, n = 0.2 represents that the historical number of times that the driver goes to the trip requirement destination is 1-3 times, n = 0.6 represents that the historical number of times that the driver goes to the trip requirement destination is 4-6 times, and n = 1 represents that the historical number of times that the driver goes to the trip requirement destination is more than 6 times; t is the time limit for the driver to go to the trip requirement destination, t = 0.2 represents that the time limit is within 15 minutes, t = 0.4 represents that the time limit is 15-30 minutes, t = 0.6 represents that the time limit is 30-60 minutes, and t = 1 represents that the time limit is more than 60 minutes; δ1 and δ2 are weight coefficients, and specific values are determined through simulation and test;

[0094] The calculation formula of the driver tension level representation factor is:

[0095]

[0096] In the formula, f T is the driver tension level representation factor; f T0 is the non-normalized initial value of the driver tension level; is a physiological feedback index of driver tension level; V s is a vehicle driving state index; V e is an internal and external operating environment index of vehicle; D j is a driving trip requirement index; ε1, ε2 and ε3 are weight coefficients, and specific values are determined through simulation and test.

[0097] The driver fatigue degree representation factor is obtained through weighted calculation according to the physiological feedback index of driver fatigue degree, the driver operation feedback index and the driving environment influence index;

[0098] The calculation formula of the physiological feedback index of driver fatigue degree is:

[0099]

[0100] In the formula, D Fp is a physiological feedback index of driver fatigue degree; h rv is a heart rate variability of driver; is a resting heart rate variability of driver; is a maximum heart rate variability of driver; bf is a blink frequency of driver; is an average blink frequency of driver; is a maximum blink frequency of driver; P θ is a θ band power of electroencephalogram of driver; is a maximum θ band power of electroencephalogram of driver; λ1, λ2 and λ3 are respectively a weight coefficient of heart rate variability of driver, a weight coefficient of blink frequency of driver and a weight coefficient of θ band power of electroencephalogram of driver, and specific values are determined through simulation and test;

[0101] The calculation formula of the driver operation feedback index is:

[0102] D o = σ1·F w + σ2·F b + σ3·F a

[0103] In the formula, D o is a driver operation feedback index; F w is a steering wheel angle adjustment frequency; F b is a brake pedal opening change frequency; F a is an accelerator pedal opening change frequency; σ1, σ2 and σ3 are weight coefficients, and specific values are determined through simulation and test.

[0104] The calculation formula of the driving environment influence index is:

[0105]

[0106] wherein E is a driving environment influence index; T d is a continuous driving time of the driver; V is a visibility of the driving environment; and L is an illumination intensity of the driving environment;

[0107] The calculation formula of the driver fatigue degree representation factor is:

[0108]

[0109] wherein f F is the driver fatigue degree representation factor; f F0 is an unnormalized initial value of the driver fatigue degree; D Fp is a driver fatigue degree physiological feedback index; D o is a driver operation feedback index; E is a driving environment influence index; and κ1, κ2 and κ3 are weight coefficients, and the specific values are determined through simulation and test.

[0110] The driver decision-making ability representation factor is obtained by weighting the driver electroencephalogram gamma band power, the driver brake reaction time, the sum of the time proportion of the driver's gaze on the nearest vehicle in front and the traffic signal lamp, and the number of lane deviations, and the specific calculation formula is:

[0111]

[0112] wherein f D is the driver decision-making ability representation factor; f D0 is an unnormalized initial value of the driver decision-making ability; P γ is a driver electroencephalogram gamma band power; t db is a driver brake reaction time; P t is the sum of the time proportion of the driver's gaze on the nearest vehicle in front and the traffic signal lamp; and n d is the number of lane deviations; and ξ1, ξ2, ξ3 and ξ4 are weight coefficients, and the specific values are determined through simulation and test.

[0113] The calculation formula of the driver torque distribution mode influence factor is:

[0114]

[0115] wherein I driver is the driver torque distribution mode influence factor; f T is a driver tension level representation factor; f D is a driver decision-making ability representation factor; and ω1, ω2 and ω3 are weight coefficients, and the specific values are determined through simulation and test.

[0116] The torque distribution mode selection rule is specifically:

[0117] When the calculated value of the driver torque distribution mode influence factor is less than 0.3, and the road adhesion coefficient is not greater than 0.2, only the motor braking torque is applied to the single wheel by the drive motor;

[0118] When the calculated value of the driver torque distribution mode influence factor is less than 0.3, and the road adhesion coefficient is greater than 0.2, the motor braking torque is preferentially applied to the single wheel by the drive motor, and when the motor braking torque cannot meet the braking demand, the hydraulic braking torque is applied by the hydraulic braking system to supplement the insufficient part;

[0119] When the calculated value of the driver torque distribution mode influence factor is between 0.3 and 0.6, the motor braking torque is preferentially applied to the single wheel by the drive motor, and when the motor braking torque cannot meet the braking demand, the hydraulic braking torque is applied by the hydraulic braking system to supplement the insufficient part;

[0120] When the calculated value of the driver torque distribution mode influence factor is greater than 0.6, only the hydraulic braking torque is applied to the single wheel by the hydraulic braking system.

[0121] The multi-level anti-lock braking safety triggering threshold is specifically divided into an anti-lock braking maximum triggering threshold and an anti-lock braking minimum triggering threshold; when the calculated value of the anti-lock braking adjusted triggering threshold is greater than the anti-lock braking maximum triggering threshold, the calculated value of the anti-lock braking adjusted triggering threshold is set as the anti-lock braking maximum triggering threshold; when the calculated value of the anti-lock braking adjusted triggering threshold is less than the anti-lock braking minimum triggering threshold, the calculated value of the anti-lock braking adjusted triggering threshold is set as the anti-lock braking minimum triggering threshold;

[0122] The multi-level automobile braking safety assistance ratio is specifically divided into an automobile braking maximum assistance ratio and an automobile braking minimum assistance ratio; when the calculated value of the automobile adjusted braking assistance ratio is greater than the automobile braking maximum assistance ratio, the calculated value of the automobile adjusted braking assistance ratio is set as the automobile braking maximum assistance ratio; when the calculated value of the automobile adjusted braking assistance ratio is less than the automobile braking minimum assistance ratio, the calculated value of the automobile adjusted braking assistance ratio is set as the automobile braking minimum assistance ratio.

Claims

1. An intelligent anti-lock braking system for an electric vehicle taking into account driver characteristics, characterized in that, The application relates to a kind of comprehensive information acquisition module, driver feature extraction module, anti-lock braking trigger threshold adjustment module, automobile brake boost ratio adjustment module, single wheel electro-hydraulic braking torque distribution mode selection module and anti-lock braking system safety guarantee module. The comprehensive information acquisition module is used to acquire driver basic information, vehicle state information and road environment information; the driver basic information includes driver heart rate, driver heart rate variability, driver skin conductance, driver electroencephalogram beta, theta and y band power and driver blink frequency; the vehicle state information includes vehicle type and vehicle speed; the road environment information includes road type, road adhesion coefficient, visibility and illumination intensity. The driver feature extraction module is used to calculate driver tension level representation factor, driver fatigue degree representation factor and driver decision-making ability representation factor. The anti-lock braking trigger threshold adjustment module is used to calculate anti-lock braking adjusted trigger threshold according to driver tension level representation factor, driver fatigue degree representation factor, driver decision-making ability representation factor and anti-lock braking initial trigger threshold. The anti-lock braking initial trigger threshold refers to ideal wheel slip rate of the road where the vehicle is currently located; the specific calculation formula of the anti-lock braking adjusted trigger threshold is as follows: In the formula, S ABS is an anti-lock braking adjustment trigger threshold; S ideal is an anti-lock braking initial trigger threshold, equal to the ideal wheel slip rate of the road surface on which the vehicle is currently located; f T is a driver tension level representation factor; f F is a driver fatigue level representation factor; f D is a driver decision-making ability representation factor; The automobile brake boost ratio adjustment module is used to determine automobile adjusted brake boost ratio by using deep neural network algorithm according to driver tension level representation factor, driver fatigue degree representation factor, driver decision-making ability representation factor and automobile standard brake boost ratio; the structure design of the deep neural network algorithm is as follows: The input layer includes four nodes corresponding to driver tension level representation factor, driver fatigue degree representation factor, driver decision-making ability representation factor and automobile standard brake boost ratio; the hidden layer adopts three-layer structure including first hidden layer, second hidden layer and third hidden layer; the first hidden layer, the second hidden layer and the third hidden layer respectively include 64, 32 and 16 neurons; the output layer includes one node, and the output is automobile adjusted brake boost ratio; The specific calculation formula of the automobile adjusted brake boost ratio is as follows: wherein f T is a driver tension level representation factor; f F is a driver fatigue degree representation factor; f D is a driver decision-making ability representation factor; r standard is a standard braking assist ratio of the car, which is determined by the manufacturer of the brake booster through experiments; r adjust is an adjusted braking assist ratio of the car; h1, h2 and h3 are respectively an output of the first hidden layer, an output of the second hidden layer and an output of the third hidden layer; ReLU is an activation function, which is specifically defined as ReLU(x) = max(0, x), wherein x is an input of the activation function; W1, W2, W3 and W4 are respectively a weight matrix from the input layer to the first hidden layer, a weight matrix from the first hidden layer to the second hidden layer, a weight matrix from the second hidden layer to the third hidden layer and a weight matrix from the third hidden layer to the output layer, and W1, W2, W3 and W4 are all determined through a training process; b1, b2, b3 and b4 are respectively a bias term of the first hidden layer, a bias term of the second hidden layer, a bias term of the third hidden layer and a bias term of the output layer, and b1, b2, b3 and b4 are all determined through the training process; and α and β are adjustment coefficients obtained through the training process. The single wheel electro-hydraulic braking torque distribution mode selection module is used to determine the source of braking torque applied to a single wheel according to driver torque distribution mode influence factor and road adhesion coefficient in combination with torque distribution mode selection rules. The anti-lock braking system safety guarantee module is used to guarantee the safety and reliability of the intelligent anti-lock braking system of the electric vehicle considering driver characteristics according to multi-stage anti-lock braking safety trigger threshold and multi-stage automobile braking safety boost ratio.

2. An electric vehicle intelligent anti-lock braking system considering driver characteristics according to claim 1, characterized in that, The driver tension level representation factor is obtained by weighted calculation according to driver tension level physiological feedback index, vehicle driving state index, vehicle internal and external operating environment index and driving trip requirement index. The specific calculation formula of the driver tension level physiological feedback index is as follows: In the formula, D Tp is a physiological feedback index of the driver's tension level; h r is the heart rate of the driver; is the resting heart rate of the driver; is the maximum heart rate of the driver;c s is the skin conductance of the driver; is the skin conductance of the driver in a resting state; is the maximum skin conductance of the driver at rest;P β is the electroencephalogram beta band power of the driver; is the maximum electroencephalogram beta band power of the driver; α1, α2 and α3 are respectively a heart rate weight coefficient of the driver, a skin conductance weight coefficient of the driver and an electroencephalogram beta band power weight coefficient of the driver, and specific values are determined through simulation and testing; The specific calculation formula of the vehicle driving state index is as follows: In the formula, V s is a vehicle driving state index; v is a vehicle speed; F is an automobile fault state, F = 0 indicates that the automobile is not in a fault state, F = 0.4 indicates that a drive motor of the automobile is in a fault state, F = 0.6 indicates that a hydraulic braking system of the automobile is in a fault state, and F = 1 indicates that both the drive motor and the hydraulic braking system of the automobile are in fault states; T v is an electric automobile type representation quantity, T v = 0.2 indicates an electric car, T v = 0.4 indicates an electric bus, T v = 0.6 indicates an electric truck, and T v = 1 indicates an electric semi-trailer; β1 and β2 are weight coefficients, and specific values are determined through simulation and tests. The specific calculation formula of the vehicle internal and external operating environment index is as follows: V e = γ1 · c r + γ2 · c t + γ3 · c w + γ4 · b p wherein V e is an indicator of the operating environment inside and outside the vehicle; c r is a road condition, c r = 0.4 indicates a rural road, c r = 0.7 indicates an urban road, c r = 1 indicates a highway; c t is a traffic condition, c t = 1 indicates that the vehicle is stopped due to traffic congestion, c t = 0.5 indicates that the vehicle is stopped due to a red traffic light, c t = 0 indicates that the vehicle is not stopped and is in a traffic state; c w is a weather condition, c w = 0 indicates sunny or cloudy weather, c w = 0.5 indicates rainy weather, c w = 1 indicates snowy weather; b p is the volume of the voice of the passenger in the vehicle, b p = 0.2 indicates that the volume is between 60-70 decibels, b p = 0.6 indicates that the volume is between 70-85 decibels, b p = 1 indicates that the volume is between 85-100 decibels; γ1, γ2, γ3, and γ4 are weight coefficients, and the specific values are determined through simulation and testing; The specific calculation formula of the driving trip requirement index is as follows: In the formula, D j is a driving trip requirement index; n is the historical number of times that the driver goes to the trip requirement destination, n=0 indicates that the historical number of times that the driver goes to the trip requirement destination is 0, n=0.2 indicates that the historical number of times that the driver goes to the trip requirement destination is 1-3, n=0.6 indicates that the historical number of times that the driver goes to the trip requirement destination is 4-6, and n=1 indicates that the historical number of times that the driver goes to the trip requirement destination is 6 or more; t is a time limit for the driver going to the trip requirement destination, t=0.2 indicates that the time limit is 15 minutes or less, t=0.4 indicates that the time limit is 15-30 minutes, t=0.6 indicates that the time limit is 30-60 minutes, and t=1 indicates that the time limit is 60 minutes or more; and δ1 and δ2 are weight coefficients, and specific values are determined through simulation and testing. The calculation formula of the driver tension level representation factor is: wherein f T is a driver stress level representation factor; f T0 is a driver stress level non-normalized initial value; D Tp is a driver stress level physiological feedback index; V s is a vehicle driving state index; V e is a vehicle interior and exterior operating environment index; D j is a driving trip requirement index; ε1, ε2, and ε3 are weight coefficients, and specific values are determined through simulation and testing.

3. The electric vehicle intelligent anti-lock braking system considering driver characteristics according to claim 1, characterized in that, The driver fatigue degree representation factor is calculated by weighting the driver fatigue degree physiological feedback index, the driver operation feedback index and the driving environment influence index; The calculation formula of the driver fatigue degree physiological feedback index is: In the formula, D Fp is a physiological feedback index of the driver's fatigue degree; h rv is the heart rate variability of the driver; is the resting heart rate variability of the driver; is the maximum heart rate variability of the driver; b f is the blink frequency of the driver; is the average blink frequency of the driver; is the maximum blink frequency of the driver; P θ is the electroencephalogram theta band power of the driver; is the maximum electroencephalogram theta band power of the driver; λ1, λ2 and λ3 are respectively the driver heart rate variability weight coefficient, the driver blink frequency weight coefficient and the driver electroencephalogram theta band power weight coefficient, and the specific values are determined by simulation and test; The calculation formula of the driver operation feedback index is: D o = σ1 · F w + σ2 · F b + σ3 · F a In the formula, D o is the driver operation feedback index; F w is the steering wheel angle adjustment frequency; F b is the brake pedal opening change frequency; F a is the accelerator pedal opening change frequency; σ1, σ2 and σ3 are weight coefficients, and specific values are determined through simulation and test; The calculation formula of the driving environment influence index is: In the formula, E is a driving environment influence index; T d is a continuous driving time of the driver; V is the visibility of the driving environment; and L is the light intensity of the driving environment. The calculation formula of the driver fatigue degree representation factor is: In the formula, f F is a driver fatigue degree representation factor; f F0 is an unnormalized initial value of the driver fatigue degree; D Fp is a driver fatigue degree physiological feedback index; D o is a driver operation feedback index; E is a driving environment influence index; κ1, κ2, and κ3 are weight coefficients, and specific values are determined through simulation and testing.

4. The electric vehicle intelligent anti-lock braking system considering driver characteristics according to claim 1, characterized in that, The driver decision-making ability representation factor is calculated by weighting the driver electroencephalogram gamma band power, the driver brake reaction time, the sum of the time ratio of the nearest vehicle in front of the driver to the traffic signal lamp and the lane deviation frequency, and the specific calculation formula is: wherein f D is a driver decision-making ability characterization factor; f D0 is a driver decision-making ability unnormalized initial value; P γ is a driver electroencephalogram gamma band power; t db is the driver's braking reaction time; P t is the sum of the time proportion of the vehicle in front of the driver's gaze and the traffic signal light; n d is the number of lane deviation times; ξ1, ξ2, ξ3, and ξ4 are weight coefficients, and the specific values are determined through simulation and test.

5. The electric vehicle intelligent anti-lock braking system considering driver characteristics according to claim 1, characterized in that, The calculation formula of the driver torque distribution mode influence factor is: In the formula, I driver is a driver torque distribution mode influence factor; f T is a driver tension level representation factor; f D is a driver decision-making ability representation factor; ω1, ω2, and ω3 are weight coefficients, and specific values are determined through simulation and test. The torque distribution mode selection rule is specifically: When the calculation value of the driver torque distribution mode influence factor is less than 0.3, and the road adhesion coefficient is not greater than 0.2, only the motor brake torque is applied to the single wheel by the drive motor; When the calculation value of the driver torque distribution mode influence factor is less than 0.3, and the road adhesion coefficient is greater than 0.2, the motor brake torque is preferentially applied to the single wheel by the drive motor, and when the motor brake torque cannot meet the braking demand, the hydraulic brake torque is applied to the single wheel by the hydraulic brake system to supplement the insufficient part; When the calculation value of the driver torque distribution mode influence factor is between 0.3 and 0.6, the motor brake torque is preferentially applied to the single wheel by the drive motor, and when the motor brake torque cannot meet the braking demand, the hydraulic brake torque is applied to the single wheel by the hydraulic brake system to supplement the insufficient part; When the calculation value of the driver torque distribution mode influence factor is greater than 0.6, only the hydraulic brake torque is applied to the single wheel by the hydraulic brake system.

6. The electric vehicle intelligent anti-lock braking system considering driver characteristics according to claim 1, characterized in that, The multi-level anti-lock braking safety trigger threshold is specifically divided into an anti-lock braking maximum trigger threshold and an anti-lock braking minimum trigger threshold; when the calculation value of the anti-lock braking adjusted trigger threshold is greater than the anti-lock braking maximum trigger threshold, the calculation value of the anti-lock braking adjusted trigger threshold is set as the anti-lock braking maximum trigger threshold; when the calculation value of the anti-lock braking adjusted trigger threshold is less than the anti-lock braking minimum trigger threshold, the calculation value of the anti-lock braking adjusted trigger threshold is set as the anti-lock braking minimum trigger threshold; The multi-level automobile braking safety boost ratio is specifically divided into an automobile braking maximum boost ratio and an automobile braking minimum boost ratio; when the calculation value of the automobile adjusted braking boost ratio is greater than the automobile braking maximum boost ratio, the calculation value of the automobile adjusted braking boost ratio is set as the automobile braking maximum boost ratio; when the calculation value of the automobile adjusted braking boost ratio is less than the automobile braking minimum boost ratio, the calculation value of the automobile adjusted braking boost ratio is set as the automobile braking minimum boost ratio.

Images