Pedal feeling feedback mechanism and control method thereof

Abstract

The invention discloses a pedal sense feedback mechanism and a control method thereof.The pedal sense feedback mechanism comprises a brake pedal, a push rod, a double-cavity pedal feedback cavity and a connecting spring, the push rod is arranged on one side of the brake pedal in the inclination direction, and the end, away from the brake pedal, of the push rod abuts against the double-cavity pedal feedback cavity; the connecting spring is installed in the double-cavity type pedal feedback cavity, the double-cavity type pedal feedback cavity is provided with a limiting block connected with the spring for blocking, and the double-cavity type pedal feedback cavity is communicated with a liquid storage tank. The device has a continuous variable pedal sense feedback working mode, a double-cavity variable-diameter structure is adopted structurally, the principle of a hydraulic system is fully utilized, the requirement for the power of a booster pump can be reduced, energy consumption is reduced, and through the two working modes of the linear pedal sense feedback mode and the continuous variable pedal sense feedback mode, energy consumption is reduced. The requirements of users with different driving habits can be met; meanwhile, the mechanism is simple in composition, and mass production can be rapidly achieved.

Description

Technical Field

[0001] The invention relates to a pedal feel feedback mechanism and a control method thereof. Background Art

[0002] The wire control brake system has the advantages of fast response speed, high safety, and strong system applicability, and is considered to be the future of automobile braking systems. Currently, the electronic mechanical braking system is the main research direction of the wire control brake system.

[0003] At present, with the rapid development and iteration of intelligent connected vehicles, intelligent driving is gradually moving towards L3 and L4 levels, and the key technology for the realization of intelligent driving is wire control technology, especially wire control braking technology related to vehicle active safety. As the main development direction of wire control braking technology in the future, the importance of EMB is self-evident. No matter what level of autonomous driving the vehicle is in, the driver is always an important driving participant, and appropriate brake pedal feedback is a key link to ensure driving safety. However, since it does not have the pedal feedback mechanism of the traditional brake system, there is a problem of difficulty in foot feedback. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the present invention provides a pedal feel feedback mechanism and a control method thereof, which have a linear pedal feel working mode of a traditional brake, and adopt a BP neural network to realize a nonlinear control mode, so that the device has a continuously variable pedal feel feedback working mode. The structure adopts a dual-cavity variable diameter structure, and makes full use of the principle of the hydraulic system, which can reduce the demand for boost pump power and reduce energy consumption. Through the two working modes of linear pedal feel feedback mode and continuously variable pedal feel feedback, the needs of users with different driving habits can be met; at the same time, the mechanism is simple in composition and can be quickly mass-produced.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a pedal feel feedback mechanism, comprising a brake pedal, a push rod, a double-chamber pedal feedback chamber and a connecting spring, wherein the push rod is arranged on one side of the brake pedal in the tilting direction, the push rod is abutted against the double-chamber pedal feedback chamber at one end away from the brake pedal, the connecting spring is installed in the double-chamber pedal feedback chamber, and the double-chamber pedal feedback chamber is provided with a limit block blocked by the connecting spring, and the double-chamber pedal feedback chamber is connected to a liquid storage tank;

[0006] Among them, the hydraulic oil in the storage tank flows to the double-chamber pedal feedback chamber and squeezes the connecting spring. The reaction force generated by the pressure on the connecting spring will be transmitted to the brake pedal through the push rod, realizing feedback to the driver of the real foot feel.

[0007] As a preferred technical solution of the present invention, a bidirectional booster pump for pumping hydraulic oil is also provided between the liquid storage tank and the double-chamber pedal feedback chamber.

[0008] As a preferred technical solution of the present invention, a displacement sensor is arranged vertically below the movement path of the push rod.

[0009] As a preferred technical solution of the present invention, it also includes a central controller, and the signal transmission end of the displacement sensor is connected to the central controller.

[0010] As a preferred technical solution of the present invention, the signal control output end of the central controller is connected to four wheel brake motors.

[0011] As a preferred technical solution of the present invention, the power output ends of the four wheel brake motors are connected to a gear set, and a ball screw is inserted into the gear set.

[0012] As a preferred technical solution of the present invention, a clamping block is connected vertically below the ball screw, and a friction plate is connected to the outer side of the clamping block.

[0013] As a preferred technical solution of the present invention, a brake disc is arranged inside the friction plate.

[0014] A control method for a pedal feel feedback mechanism, the specific steps are as follows:

[0015] S1. The driver steps on the brake pedal, which pushes the push rod forward and compresses the hydraulic oil in the double-chamber pedal feedback chamber. At this time, the two-way booster pump is in a controlled open state, and the hydraulic oil is pumped from the reservoir into the double-chamber pedal feedback chamber with a connecting spring;

[0016] S1-1. The entry of hydraulic oil changes the pressure on the connecting spring in the double-chamber pedal feedback chamber, causing the connecting spring to deform and react on the push rod. The push rod retreats under the action of the elastic force and transmits it to the brake pedal for feedback.

[0017] S2. During the forward movement of the push rod, the displacement sensor detects the displacement signal and transmits the signal to the central controller in real time. The central controller distributes the braking force after analyzing the signal. The braking force signal will be distributed to the four wheel brake motors respectively. The wheel brake motors perform the braking action after receiving the control signal.

[0018] S2-1, the wheel brake motor rotates after receiving the command, and the rotational motion of the wheel brake motor is converted into linear motion through the cooperation of the gear set and the ball screw. The ball screw drives the connection clamping block to move toward the brake disc, so that the friction plate clamps the brake disc, thereby completing the braking process;

[0019] S3, linear pedal feel feedback working mode:

[0020] The push rod and the sealed end of the left side of the connecting spring form a closed chamber. At this time, the area of ​​the connecting end of the push rod is S1, and the area of ​​the connecting ends on both sides of the connecting spring is S2. The relationship between the force F1 applied by the brake pedal on one end of the push rod and the force F2 generated by the left end of the connecting spring is:

[0021] F1×S2=F2×S1

[0022] F2=K×L2 L2≤L1

[0023] Among them, F2 is the force generated by the left side of the connecting spring, and the force F3 on the right side is generated by the two-way booster pump. Then the force balance equation in the system is:

[0024]

[0025] Right now:

[0026]

[0027] When the brake pedal stroke reaches the maximum Lmax, the required brake pedal feedback force is Fmax, if:

[0028] Fmax≤K×Lmax

[0029] The connection spring stiffness meets the linear pedal force feedback requirement. At this time, the two-way booster pump is in a straight-through state, that is, F3 = 0. At this time:

[0030]

[0031] At this time, there is a linear relationship between F1 and L2, realizing a linear pedal feel feedback mode;

[0032] S4, continuously variable pedal feel feedback working mode:

[0033] According to the above steps, the size of F3 is adjusted by the bidirectional booster pump.

[0034] The BP neural network is used for control. The neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the data collected by the processed displacement sensor. The displacement value, displacement change rate, and displacement change difference obtained after processing the displacement data are used as inputs of each node in the input layer. The mathematical equation can be expressed as:

[0035] Input i =X(i), (i=1, 2, 3)

[0036] The hidden layer processes the parameters transmitted by the input layer. The hidden layer is mainly composed of parameter matrix, bias term and activation function. The process can be described by mathematical equation:

[0037]

[0038] The output of the hidden layer is:

[0039] Output(t)=f[Net i (t)], (i=1, 2, 3,..., q)

[0040] Where i is the number of neuron inputs, w ij is the weight coefficient, and the activation function used in the output layer is the Sigmoid function, whose image is an S-shaped curve, which maps any real number to the interval (0,1), and its expression is:

[0041] δ(x)=1 / (1+e -x )

[0042] Mapping it to the interval [-1,1], its form can be written as:

[0043] f(x)=tanh(x)=(e x -e -x ) / (e x +e -x ) The deviation between the calculated result and the expected result is taken into consideration to reduce the error between the expectation and the result. The problem is transformed into an optimization problem. The loss function for this problem can be described as:

[0044]

[0045] The minimum value of the loss function is found through iteration. During the iteration process, the weights between nodes in each layer are constantly updated.

[0046]

[0047] Where: Adjustment amount, α[W (t) -W (t-1) ] is the smoothing term

[0048] By updating the weights in the iterative process, forward propagation is performed with the training set data to test whether the results meet the requirements. If not, back propagation is performed and the iterative process is continued, repeating the cycle to achieve a continuously variable pedal feel feedback working mode.

[0049] Compared with the prior art, the beneficial effects that the present invention can achieve are: it has the linear pedal feel working mode of the traditional brake, and adopts the BP neural network to realize the nonlinear control mode, so that the device has a continuously variable pedal feel feedback working mode, and adopts a dual-cavity variable diameter structure in structure, making full use of the hydraulic system principle, which can reduce the demand for boost pump power and reduce energy consumption. Through the two working modes of linear pedal feel feedback mode and continuously variable pedal feel feedback mode, it can meet the needs of users with different driving habits; at the same time, the mechanism is simple in composition and can be quickly mass-produced. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A diagram of a brake-by-wire system of the present invention;

[0051] Figure 2 This is a diagram of the BP neural network training iteration process of the present invention.

[0052] Among them: 1. Brake pedal; 2. Push rod; 3. Fluid storage tank; 4. Bidirectional boost pump; 5. Double-chamber pedal feedback chamber; 6. Limit block; 7. Connecting spring; 8. Displacement sensor; 9. Central controller; 10. Brake motor; 11. Gear set; 12. Ball screw; 13. Clamping block; 14. Friction plate; 15. Brake disc. DETAILED DESCRIPTION

[0053] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further described below in conjunction with specific embodiments, but the following embodiments are only preferred embodiments of the present invention, not all. Based on the embodiments in the implementation mode, other embodiments obtained by those skilled in the art without creative work are all within the scope of protection of the present invention. The experimental methods in the following embodiments, unless otherwise specified, are conventional methods, and the materials, reagents, etc. used in the following embodiments, unless otherwise specified, can be obtained from commercial channels.

[0054] Example:

[0055] like Figure 1 - Figure 2As shown, this embodiment proposes a pedal feel feedback mechanism, including a brake pedal 1, a push rod 2, a double-chamber pedal feedback chamber 5 and a connecting spring 7, wherein the push rod 2 is arranged on one side of the brake pedal 1 in the tilting direction, and the end of the push rod 2 away from the brake pedal 1 abuts against the double-chamber pedal feedback chamber 5, and the connecting spring 7 is installed in the double-chamber pedal feedback chamber 5, and the double-chamber pedal feedback chamber 5 is provided with a limit block 6 blocked by the connecting spring 7, and the double-chamber pedal feedback chamber 5 is connected with a liquid storage tank 3; wherein, the hydraulic oil in the liquid storage tank 3 flows into the double-chamber pedal feedback chamber 5 and squeezes the connecting spring 7, and the reaction force generated by the compression of the connecting spring 7 will be transmitted to the brake pedal 1 through the push rod 2, so as to realize the feedback of the real foot feel to the driver;

[0056] By having the linear pedal feel working mode of the traditional brake and using the BP neural network to realize the nonlinear control mode, the device has a continuously variable pedal feel feedback working mode. The structure adopts a double-cavity variable diameter structure, fully utilizing the principle of the hydraulic system, which can reduce the demand for the power of the two-way boost pump 4 and reduce energy consumption. Through the two working modes of linear pedal feel feedback mode and continuously variable pedal feel feedback mode, it can meet the needs of users with different driving habits; at the same time, the structure is simple and can be quickly mass-produced.

[0057] The liquid storage tank 3 can be connected to the outside world, so as to ensure that the pedal feel simulation mechanism is not affected by atmospheric pressure. The core components of the system are the two-way booster pump and the double-chamber pedal feel generation chamber 5. The two-way booster pump 4 changes its working state under the control of the central controller 9, which is the key to realizing the anthropomorphic pedal feel feedback. The central controller 9 is the MCU.

[0058] Some parameters in the figure are as follows: the stiffness of the connecting spring 7 is K, the area of ​​the connecting ends on both sides of the connecting spring 7 is S2, the area of ​​the connecting end of the push rod 2 is S1, the free stroke of the connecting spring 7 is L1, the compressed stroke of the connecting spring 7 is L2, the free stroke of the brake pedal 1 is L3, and the maximum stroke of the brake pedal 1 is Lmax.

[0059] The working modes of the brake pedal feel feedback device are divided into two types: the first type: linear pedal feel feedback working mode; the second type: continuously variable pedal feel feedback working mode. The control of the two working modes is described in detail below.

[0060] 1. Linear pedal feel feedback working mode

[0061] In this mode, pedal feel feedback is mainly generated by the elasticity of the connecting spring 7 and the boosting capacity of the two-way booster pump 4. In this mode, the calculation method of the brake pedal 1 feedback force is as follows: the push rod 2 and the left sealed end of the connecting spring 7 travel to close the chamber. At this time, since the area of ​​the connecting end of the push rod 2 is S1, and the area of ​​the connecting ends on both sides of the connecting spring 7 is S2, according to Pascal's law, the relationship between the force F1 applied by the brake pedal 1 to one end of the push rod 2 and the force F2 generated by the left end of the connecting spring 7 is:

[0062] F1×S2=F2×S1 (1)

[0063] F2=K×L2 L2≤L1 (2)

[0064] Where F2 is the force generated on the left side of the connecting spring 7, and the force F3 on the right side is generated by the two-way booster pump 4. Then the force balance equation in the system is:

[0065]

[0066] Right now:

[0067] The stiffness K of the connecting spring 7 is a key parameter in the linear pedal feel feedback working mode. When the brake pedal 1 stroke reaches the maximum Lmax, the required brake pedal 1 feedback force is Fmax, if:

[0068] Fmax≤K×Lmax (5)

[0069] This means that the stiffness of the connecting spring 7 can meet the linear pedal force feedback requirements. At this time, the two-way booster pump 4 is in a straight-through state, that is, F3 = 0. At this time, equation 4 can be rewritten as:

[0070]

[0071] At this point it is clear from Equation 6 that there is a linear relationship between F1 and L2, so the system is able to achieve a linear pedal feel feedback mode.

[0072] Continuously variable pedal feel feedback working modes:

[0073] In this working mode, the control of the two-way boost pump 4 is the core of the function realization. According to equation 4, the size of F3 in the equation is adjusted by the two-way boost pump 4. Therefore, efficient control of the two-way boost pump 4 is the key to realizing the continuous variable pedal feel feedback working mode. Therefore, this patent proposes to use the BP (Back Propagation) neural network method for control.

[0074] BP neural network is a multi-layer feedforward neural network trained according to the error back propagation algorithm. It is a typical nonlinear algorithm and also the most widely used neural network. The neural network is structurally composed of an input layer, a hidden layer, and an output layer. The input layer is mainly responsible for receiving the data collected by the pedal displacement sensor after processing. The displacement value, displacement change rate, and displacement change difference obtained after processing the displacement data are used as inputs for each node of the input layer. The mathematical equation can be expressed as:

[0075] Input i =X(i), (i=1, 2, 3)

[0076] The hidden layer processes the parameters transmitted by the input layer. The hidden layer is mainly composed of parameter matrix, bias term and activation function. The process can be described by mathematical equation:

[0077]

[0078] The output of the hidden layer is:

[0079] Output(t)=f[Net i (t)], (i=1, 2, 3,..., q)

[0080] Where i is the number of neuron inputs, w ij is the weight coefficient, and the activation function used in the output layer is the Sigmoid function, which is a commonly used mathematical function. Its image is an S-shaped curve. It maps any real number to the interval (0,1). Its expression is:

[0081] δ(x)=1 / (1+e -x )

[0082] Mapping it to the interval [-1,1], its form can be written as:

[0083] f(x)=tanh(x)=(e x -e -x ) / (e x +e -x )

[0084] The deviation between the network calculation result and the expected result needs to be taken into consideration. In order to reduce the error between the expectation and the result, the problem is transformed into an optimization problem. Therefore, the loss function for this problem can be described as:

[0085]

[0086] The minimum value of the loss function is found through iteration. During the iteration process, the weights between nodes in each layer are constantly updated.

[0087]

[0088] Where: Adjustment amount, α[W (t) -W (t-1) ] is the smoothing term

[0089] By updating the weights in the iterative process, forward propagation is performed with the training set data to test whether the results meet the requirements. If not, back propagation is performed and the iterative process is continued, repeating the cycle until a satisfactory result is obtained.

[0090] The training process of BP neural network is as follows: Figure 2 As shown, the trained network is used to control the working process of the boost pump. Through the control of the BP neural network, nonlinear control of the pedal feel simulator can be achieved, nonlinear foot feel feedback can be achieved, and thus a continuously variable pedal feel feedback working mode can be achieved.

[0091] A control method for a pedal feel feedback mechanism, the specific steps are as follows:

[0092] S1, the driver steps on the brake pedal 1, the brake pedal 1 pushes the push rod 2, so that the push rod 2 is pushed forward and the hydraulic oil in the double-chamber pedal feedback chamber 5 is compressed. At this time, the two-way booster pump 4 is in a controlled open state, and the hydraulic oil is pumped from the liquid storage tank 3 into the double-chamber pedal feedback chamber 5 with the connecting spring 7;

[0093] S1-1, the entry of hydraulic oil changes the pressure on the connecting spring 7 in the double-chamber pedal feedback chamber 5, causing the connecting spring 7 to deform and react on the push rod 2. The push rod 2 retreats under the action of the elastic force and transmits it to the brake pedal 1 for feedback;

[0094] S2. During the forward movement of the push rod 2, the displacement sensor 8 detects the displacement signal and transmits the signal to the central controller in real time. The central controller 9 distributes the braking force after analyzing the signal. The braking force signal will be distributed to the four wheel brake motors 10 respectively. The wheel brake motors 10 perform the braking action after receiving the control signal.

[0095] S2-1, the wheel brake motor 10 rotates after receiving the command, and the rotational motion of the wheel brake motor 10 is converted into linear motion through the cooperation of the gear set 11 and the ball screw 12. The ball screw 12 drives the connection clamping block 13 to move toward the brake disc 15, so that the friction plate 14 clamps the brake disc 15, thereby completing the braking process;

[0096] S3, linear pedal feel feedback working mode:

[0097] The push rod 2 and the sealed end of the left side of the connecting spring 7 form a closed chamber. At this time, the area of ​​the connecting end of the push rod 2 is S1, and the area of ​​the connecting ends on both sides of the connecting spring 7 is S2. The relationship between the force F1 applied by the brake pedal 1 to one end of the push rod and the force F2 generated by the left end of the connecting spring 7 is: F1×S2=F2×S1 (1)

[0098] F2=K×L2 L2≤L1 (2)

[0099] Where F2 is the force generated on the left side of the connecting spring 7, and the force F3 on the right side is generated by the two-way booster pump 4. Then the force balance equation in the system is:

[0100]

[0101] Right now:

[0102]

[0103] When the brake pedal 1 stroke reaches the maximum Lmax, the required brake pedal 1 feedback force is Fmax, if:

[0104] Fmax≤K×Lmax (5)

[0105] The stiffness of the connecting spring 7 meets the linear pedal force feedback requirement. At this time, the bidirectional booster pump 4 is in a straight-through state, that is, F3 = 0. At this time, equation 4 is:

[0106]

[0107] At this time, there is a linear relationship between F1 and L2, realizing a linear pedal feel feedback mode;

[0108] S4, continuously variable pedal feel feedback working mode:

[0109] According to equation 4 in the above steps, the size of F3 is adjusted by the bidirectional booster pump 4.

[0110] The BP neural network is used for control. The neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the data collected by the processed displacement sensor. The displacement value, displacement change rate, and displacement change difference obtained after processing the displacement data are used as inputs of each node in the input layer. The mathematical equation can be expressed as:

[0111] Input i =X(i), (i=1, 2, 3)

[0112] The hidden layer processes the parameters transmitted by the input layer. The hidden layer is mainly composed of parameter matrix, bias term and activation function. The process can be described by mathematical equation:

[0113]

[0114] The output of the hidden layer is:

[0115] Output(t)=f[Net i (t)], (i=1, 2, 3,..., q)

[0116] Where i is the number of neuron inputs, w ij is the weight coefficient, and the activation function used in the output layer is the Sigmoid function, whose image is an S-shaped curve, which maps any real number to the interval (0,1), and its expression is:

[0117] δ(x)=1 / (1+e -x ) is mapped to the interval [-1,1], then its form can be written as:

[0118] f(x)=tanh(x)=(e x -e -x ) / (e x +e -x )

[0119] The deviation between the calculated result and the expected result is taken into consideration to reduce the error between the expectation and the result. The problem is transformed into an optimization problem. The loss function for this problem can be described as:

[0120]

[0121] The minimum value of the loss function is found through iteration. During the iteration process, the weights between nodes in each layer are constantly updated.

[0122]

[0123] Where: Adjustment amount, α[W (t) -W (t-1) ] is the smoothing term

[0124] By updating the weights in the iterative process, forward propagation is performed with the training set data to test whether the results meet the requirements. If not, back propagation is performed and the iterative process is continued, repeating the cycle to achieve a continuously variable pedal feel feedback working mode.

[0125] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.

[0126] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above implementation regulations. The above implementation regulations and descriptions are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A pedal feel feedback mechanism, characterized in that: The invention comprises a brake pedal (1), a push rod (2), a double-chamber pedal feedback chamber (5) and a connecting spring (7), wherein the push rod (2) is arranged on one side of the brake pedal (1) in an inclined direction, and the end of the push rod (2) away from the brake pedal (1) is in contact with the double-chamber pedal feedback chamber (5), the connecting spring (7) is installed in the double-chamber pedal feedback chamber (5), and the double-chamber pedal feedback chamber (5) is provided with a limit block (6) blocked by the connecting spring (7), and the double-chamber pedal feedback chamber (5) is connected to a liquid storage tank (3); The hydraulic oil in the liquid storage tank (3) flows into the double-chamber pedal feedback chamber (5) and compresses the connecting spring (7). The reaction force generated by the compression of the connecting spring (7) is transmitted to the brake pedal (1) through the push rod (2), thereby providing the driver with a real foot feeling.

2. The pedal feel feedback mechanism according to claim 1, characterized in that: A bidirectional booster pump (4) for pumping hydraulic oil is also provided between the liquid storage tank (3) and the double-chamber pedal feedback chamber (5).

3. The pedal feel feedback mechanism according to claim 1, characterized in that: A displacement sensor (8) is arranged vertically below the movement path of the push rod (2).

4. The pedal feel feedback mechanism according to claim 3, characterized in that: It also includes a central controller (9), and the signal transmission end of the displacement sensor (8) is connected to the central controller (9).

5. The pedal feel feedback mechanism according to claim 4, characterized in that: The signal control output end of the central controller (9) is connected to four wheel brake motors (10).

6. The pedal feel feedback mechanism according to claim 5, characterized in that: The power output ends of the four wheel brake motors (10) are connected to a gear set (11), and a ball screw (12) is inserted into the gear set (11).

7. The pedal feel feedback mechanism according to claim 6, characterized in that: A clamping block (13) is vertically connected below the ball screw (12), and a friction plate (14) is connected to the outside of the clamping block (13).

8. The pedal feel feedback mechanism according to claim 7, characterized in that: A brake disc (15) is arranged inside the friction plate (14).

9. A control method for a pedal feel feedback mechanism according to any one of claims 1 to 8, characterized in that: S1. The driver steps on the brake pedal (1), and the brake pedal (1) pushes the push rod (2), so that the push rod (2) is pushed forward and the hydraulic oil in the double-chamber pedal feedback chamber (5) is compressed. At this time, the bidirectional booster pump (4) is in a controlled open state, and the hydraulic oil is pumped from the liquid storage tank (3) into the double-chamber pedal feedback chamber (5) having a connecting spring (7); S1-1, the entry of hydraulic oil changes the pressure on the connecting spring (7) in the double-chamber pedal feedback chamber (5), causing the connecting spring (7) to deform and react on the push rod (2). The push rod (2) retreats under the action of the elastic force and transmits the pressure to the brake pedal (1) for feedback; S2. During the forward movement of the push rod (2), the displacement sensor (8) detects the displacement signal and transmits the signal to the central controller (9) in real time. The central controller (9) distributes the braking force after analyzing the signal. The braking force signal is distributed to the four wheel brake motors (10) respectively. The wheel brake motors (10) perform the braking action after receiving the control signal. S2-1, the wheel brake motor (10) rotates after receiving the command, and the rotational motion of the wheel brake motor (10) is converted into linear motion through the cooperation of the gear set (11) and the ball screw (12), and the ball screw (12) drives the connecting clamping block (13) to move toward the brake disc (15), so that the friction plate (14) clamps the brake disc (15), thereby completing the braking process; S3, linear pedal feedback working mode: The push rod (2) and the left sealed end of the connecting spring (7) form a closed chamber. At this time, the area of ​​the connecting end of the push rod (2) is S1, and the area of ​​the connecting ends on both sides of the connecting spring (7) is S2. The relationship between the force F1 applied by the brake pedal (1) to one end of the push rod (2) and the force F2 generated by the left end of the connecting spring (7) is: F1×S2=F2×S1 F2=K×L2 L2≤L1 Where F2 is the force generated on the left side of the connecting spring (7), and the force F3 on the right side is generated by the bidirectional booster pump (4). Then the force balance equation in the system is: Right now: When the brake pedal (1) travel reaches the maximum Lmax, the required brake pedal (1) feedback force is Fmax, if: Fmax≤K×Lmax The stiffness of the connecting spring (7) meets the linear pedal force feedback requirement. At this time, the bidirectional booster pump (4) is in a straight-through state, that is, F3 = 0. At this time, equation 4 is: At this time, there is a linear relationship between F1 and L2, realizing a linear pedal feel feedback mode; S4, continuously variable pedal feel feedback working mode: According to the above steps, the size of F3 is adjusted by the bidirectional booster pump (4). The BP neural network is used for control. The neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the data collected by the processed displacement sensor. The displacement value, displacement change rate, and displacement change difference obtained after processing the displacement data are used as inputs of each node in the input layer. The mathematical equation can be expressed as: Input i =X(i),(i=1,2,3) The hidden layer processes the parameters transmitted by the input layer. The hidden layer is mainly composed of parameter matrix, bias term and activation function. The process can be described by mathematical equation: The output of the hidden layer is: Output(t)=f[Net i (t)],(i=1,2,3,…,q) Where i is the number of neuron inputs, w ij is the weight coefficient, and the activation function used in the output layer is the Sigmoid function, whose image is an S-shaped curve, which maps any real number to the interval (0,1), and its expression is: δ(x)=1 / 1+e -x ) Mapping it to the interval [-1,1], its form can be written as: f(x)=tanh(x)=(e x -And -x ) / (And x +e -x ) The deviation between the calculated result and the expected result is taken into consideration to reduce the error between the expectation and the result. The problem is transformed into an optimization problem. The loss function for this problem can be described as: The minimum value of the loss function is found through iteration. During the iteration process, the weights between nodes in each layer are constantly updated. Where: Adjustment amount, α[W (t) -W (t-1) ] is the smoothing term By updating the weights in the iterative process, forward propagation is performed with the training set data to test whether the results meet the requirements. If not, back propagation is performed and the iterative process is continued, repeating the cycle to achieve a continuously variable pedal feel feedback working mode.

Images