CN112987574B - A control method of cloud-controlled intelligent chassis system based on multi-agent - Google Patents

Abstract

The invention discloses a multi-agent-based cloud-controlled intelligent chassis system and a control method, comprising: a wire-controlled hydraulic steering subsystem, a wire-controlled hydraulic braking subsystem, a wire-controlled suspension subsystem, a perception module, and a reinforcement learning module , communication module, cloud processing center and multi-agent coordination controller; the invention is based on multi-agent system coordination control theory, and realizes various aspects of wire-controlled hydraulic steering subsystem, wire-controlled hydraulic braking subsystem and wire-controlled suspension subsystem. Coupling to form a multi-agent wire-controlled chassis system, which greatly improves the stability and reliability of the wire-controlled chassis system, and is of great significance to the improvement of the comfort, safety and smoothness of the commercial vehicle.

Description

A control method of cloud-controlled intelligent chassis system based on multi-agent

technical field

The invention belongs to the technical field of automobile chassis systems, and in particular relates to a cloud-controlled intelligent chassis system and a control method based on multi-agents.

Background technique

The drive-by-wire chassis system is a hot spot in the current automotive chassis research. The drive-by-wire chassis system of commercial vehicles mainly includes a steering-by-wire subsystem, a brake-by-wire subsystem and a suspension-by-wire subsystem. Each subsystem cancels the mechanical connection of the traditional chassis. , can be controlled separately. At present, although the functions of each subsystem are relatively complete, the coupling between the subsystems has not been fully considered in all aspects. The most commonly used methods are to improve a certain parameter to design a control strategy, or to mechanically superimpose multiple performance indicators. . However, these measures cannot reflect the performance indicators of the vehicle chassis. Therefore, on the road to intelligence, finding new methods to reflect the performance indicators of the vehicle chassis has become a rigid need.

Today, multi-agent technology has been widely used in intelligent robots, traffic control, distributed intelligent decision-making and other fields. Its super coordination and learning, coupled with high robustness and reliability in solving practical problems It is particularly suitable for the control of commercial vehicle by wire chassis. Applying the multi-agent system theory to the drive-by-wire chassis system of commercial vehicles, fully considering the dynamic relationship between the vertical, longitudinal and lateral directions of commercial vehicles under different working conditions, will greatly improve the stability and performance of the drive-by-wire chassis system. reliability, which is of great significance to the improvement of the comfort, safety and smoothness of the commercial vehicle.

SUMMARY OF THE INVENTION

In view of the above-mentioned deficiencies of the prior art, the purpose of the present invention is to provide a multi-agent-based cloud-controlled intelligent chassis system and a control method, so as to solve the problem that the prior art cannot perform vertical, vertical and The problem of longitudinal and lateral coordinated control; the invention regards the key subsystems of the wire-controlled chassis as a separate intelligent body, establishes the connection between the various subsystems, forms a multi-intelligent wire-controlled chassis structure, and realizes the coordinated control of the multi-agent system , to improve the overall performance of the vehicle.

For achieving the above object, the technical scheme adopted in the present invention is as follows:

A multi-agent-based cloud-controlled intelligent chassis system of the present invention includes: a wire-controlled hydraulic steering subsystem, a wire-controlled hydraulic braking subsystem, a wire-controlled suspension subsystem, a perception module, a reinforcement learning module, and a communication module, Cloud processing center and multi-agent coordination controller;

The wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem are all distributed on the multi-agent wire-controlled chassis for realizing the control of the vehicle chassis;

The perception module is respectively connected to each vehicle-mounted sensor and the multi-agent coordination controller for collecting information;

The reinforcement learning module evaluates the motion state of the vehicle itself according to the output results of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem, and obtains the complete vehicle according to the current motion state of the vehicle itself. Action behavior with the best stability and continuous reinforcement learning; data connection with the multi-agent coordination controller, and data connection with the cloud processing center through the communication module;

The cloud processing center is used to regularly check whether the reinforcement learning results of the vehicle are reasonable, and adjust the unreasonable reinforcement learning results;

The multi-agent coordination controller is used to adjust the information input of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem.

Further, the wire-controlled hydraulic steering subsystem includes: a steering wheel unit, a wire-controlled steering control unit, a hydraulic steering unit, an electric steering unit, and a steering road sensing unit; the rotation signal of the steering wheel unit is input to the wire-controlled steering control unit, The steering-by-wire control unit coordinates and controls the hydraulic steering unit and the electric steering unit to work together to complete the vehicle steering. The steering road sensing unit extracts vehicle information and road surface state information, and generates a positive torque to act on the steering wheel unit to simulate the real road feeling for the driver;

The wire-controlled hydraulic brake subsystem includes: a wire-controlled brake control unit, a hydraulic brake unit, a regenerative brake unit and a pedal simulation unit; the pedal simulation unit converts the information of the strength of the driver's pedal depression into a corresponding electrical signal And transmit it to the brake-by-wire unit. The brake-by-wire unit controls the hydraulic braking unit to complete the braking operation according to the received electrical signal. The regenerative braking unit stores kinetic energy when the vehicle is braking, and the energy storage motor of the regenerative braking unit is connected to the vehicle. on the axle;

The wire-controlled suspension subsystem includes: an active suspension unit, a suspension control unit; the active suspension unit adjusts the damping and stiffness of the suspension, and the suspension control unit adjusts the damping and stiffness of the suspension according to the vehicle height, vehicle speed, steering angle and rate, and braking The signal is used to control the active suspension unit to adjust suspension damping and stiffness.

Further, the information collected by the sensing module includes: the operation command input by the driver, the smoothness of the road surface, the curvature of the road, the lateral speed and acceleration of the vehicle, the longitudinal speed and acceleration of the vehicle, the vertical acceleration of the vehicle, the lateral Swing angular velocity, body slip angle, suspension damping force, suspension dynamic disturbance, wheel dynamic load, wheel slip rate.

Further, the communication module adopts the 5G communication mode to realize data communication between the vehicle and the cloud processing center.

Further, the cloud processing center also optimizes the riding comfort and safety of the target vehicle according to the learning results of other vehicles.

Further, the multi-agent coordination controller is used for, according to the current state of the vehicle, the operating conditions of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem, the driver's input information and the road surface. Flatness and road curvature, with vehicle safety and comfort as the ultimate goal, control three subsystems, so as to achieve coordinated control of the entire vehicle chassis.

Further, the dynamics model of the wire-controlled hydraulic steering subsystem includes: lateral motion and yaw motion models, respectively:

where m is the vehicle mass, v is the vehicle speed, γ is the yaw rate, β is the body slip angle, F _yfl is the left front wheel cornering force of the vehicle, F _yfr is the vehicle right front wheel cornering force, F _yrl is the cornering force of the left rear wheel of the vehicle, F _yrr is the cornering force of the right rear wheel of the vehicle, I _s is the yaw moment of inertia of the vehicle, the distance from the center of mass of L _f to the front axle, and the distance from the center of mass of L _r to the rear axle distance.

Further, the wheel rotation dynamics model of the control-by-wire hydraulic brake subsystem is:

In the formula, I ₁ is the moment of inertia of the left front wheel of the vehicle, I ₂ is the moment of inertia of the right front wheel of the vehicle, I ₃ is the moment of inertia of the left rear wheel of the vehicle, I ₄ is the moment of inertia of the right rear wheel of the vehicle, ω ₁ is The rotational angular velocity of the left front wheel of the vehicle, ω ₂ is the rotational angular velocity of the right front wheel of the vehicle, ω ₃ is the rotational angular velocity of the left rear wheel of the vehicle, ω ₄ is the rotational angular velocity of the right rear wheel of the vehicle, and F _xfl is the longitudinal direction of the left front wheel of the vehicle Adhesion, F _xfr is the longitudinal adhesion of the right front wheel of the vehicle, F _xrl is the longitudinal adhesion of the left rear wheel of the vehicle, F _xrr is the longitudinal adhesion of the right rear wheel of the vehicle, R ₁ is the radius of the left front wheel of the vehicle, R ₂ is the radius of the front wheel on the right side of the vehicle, R ₃ is the radius of the rear wheel on the left side of the vehicle, R ₄ is the radius of the rear wheel on the right side of the vehicle, T _b1 is the braking torque of the front wheel on the left side of the vehicle, and T _b2 is the braking torque of the front wheel on the right side of the vehicle. dynamic torque, T _b3 is the braking torque of the left rear wheel of the vehicle, and T _b4 is the braking torque of the right rear wheel of the vehicle;

The wheel slip formula is:

where S is the slip rate of the vehicle, v ₀ is the longitudinal speed of the vehicle, R is the wheel radius, and ω is the rotational angular velocity of the wheel;

The calculation formula of braking torque is:

T _b =(PP ₀ )A _w ηB _F R _g

In the formula, T _b is the braking torque of the brake, P is the pressure of the brake line, P ₀ is the initial pressure of the brake line, A _w is the effective working area of the brake cylinder, η is the efficiency of the brake cylinder, B _F For the braking efficiency factor, R _g is the radius of the brake disc.

Further, the vertical dynamic model at the center of mass of the vehicle body of the suspension-by-wire subsystem is:

The body pitch dynamics model is:

The body roll dynamics model is:

The vertical motion model of the four unsprung mass parts is:

The vertical motion model at the four endpoints of the body is:

为车辆质心处垂直方向的加速度，I_p为簧载质量部分的俯仰转动惯量，θ为簧载质量部分的俯仰角，

为簧载质量部分的俯仰角角加速度，I_r为簧载质量部分的侧倾转动惯量，

为簧载质量部分的侧倾角，

为簧载质量部分的侧倾角角加速度，L_f为车身质心到前轴的距离，L_r为车身质心到后轴的距离，B_f为前轴轮距，B_r为后轴轮距；K_si为第i轮处的悬架刚度，K_ti为第i轮处的轮胎刚度，C_si为第i轮处的悬架阻尼系数，z_wi为第i处出车轮垂向位移，

为第i处出车轮垂向速度，

为第i处出车轮垂向加速度，z_bi为第i轮处车身端点垂向位移，f_i为第i轮处的悬架阻尼力， m_wi为车辆的第i轮处的非簧载质量，z_gi为第i轮处路面输入的垂向位移，i＝1,2,3,4.，1代表左侧前轮，2代表右侧前轮，3代表左侧后轮，4代表右侧后轮。where m _b is the sprung mass of the vehicle, z _b is the vertical displacement at the center of mass of the vehicle body,

is the vertical acceleration at the center of mass of the vehicle, I _p is the pitch moment of inertia of the sprung mass part, θ is the pitch angle of the sprung mass part,

is the pitch angle acceleration of the sprung mass part, I _r is the roll moment of inertia of the sprung mass part,

is the roll angle of the sprung mass part,

is the roll angle acceleration of the sprung mass part, L _f is the distance from the center of mass of the body to the front axle, L _r is the distance from the center of mass of the body to the rear axle, B _f is the wheel track of the front wheel, and B _r is the wheel track of the rear wheel; K _si is the suspension stiffness at the i-th wheel, K _ti is the tire stiffness at the i-th wheel, _Csi is the suspension damping coefficient at the i-th wheel, _zwi is the vertical displacement of the wheel at the i-th wheel,

is the vertical speed of the wheel exiting at the i-th place,

is the vertical acceleration of the wheel at the i-th wheel, _zbi is the vertical displacement of the body end point at the i-th wheel, f _i is the suspension damping force at the i-th wheel, _mwi is the unsprung mass at the i-th wheel of the vehicle , z _gi is the vertical displacement of the road input at the i-th wheel, i=1, 2, 3, 4., 1 represents the left front wheel, 2 represents the right front wheel, 3 represents the left rear wheel, and 4 represents the right side rear wheel.

Further, the specific control method of the multi-agent coordination controller is: when the road surface is uneven and there is interference, the vehicle will tilt in the vertical direction and the pitch direction, which will lead to different forces acting on the suspension subsystem. Stability; when cornering, the friction of the tires causes yaw motion; when braking, the resulting acceleration will cause the vehicle to produce violent yaw motion and pitch motion, resulting in balance due to changes in load distribution The breaking of the state will reduce the operating stability of the car; the driving state of the vehicle will be affected by the steering angle, pedal braking, accelerator input and ground conditions, etc. At the same time, the current working conditions of each subsystem will also affect The work of other subsystems; in order to restore the vehicle to a good balance state, it is necessary to adjust the various subsystems at this time according to the size of the role, and each subsystem cooperates with each other to complete the control of the vehicle chassis, so that the vehicle maintains good safety. with comfort.

A control method of a cloud-controlled intelligent chassis system based on multi-agents of the present invention is based on the above system, and the steps are as follows:

1) The perception module collects the information on the smoothness and curvature of the road surface, the information input by the driver and the motion information of the vehicle, and transmits the collected information to the multi-agent coordination controller;

2) The multi-agent coordination controller controls each subsystem according to the above-mentioned information received, and each subsystem controls the vehicle chassis cooperatively;

3) The reinforcement learning module evaluates the current output results of each subsystem according to the set final goal, and dynamically adjusts the next output according to the evaluation results to complete the optimization of the multi-agent collaborative controller;

4) The communication module regularly uploads the vehicle information and the current control algorithm to the cloud processing center; the cloud processing center checks the reliability of the received algorithm and feeds back the test results;

5) The reinforcement learning module further optimizes the control algorithm according to the feedback results of the cloud processing center.

各车轮动载荷|z_wi-z_bi|，各悬架动扰度 |z_wi-z_gi|，车辆当前的横摆角速度γ，质心侧偏角

车身俯仰角θ，车轮滑移率S，车辆的纵向速度v。Further, the signal transmitted to the multi-agent coordination controller in the step 1) is based on the damping force f _i of each suspension in the dynamic model established by each subsystem, the vertical acceleration

The dynamic load of each wheel |z _wi -z _bi |, the dynamic disturbance of each suspension |z _wi -z _gi |, the current yaw rate γ of the vehicle, the side slip angle of the center of mass

Body pitch angle θ, wheel slip rate S, vehicle longitudinal speed v.

Further, the control process of the multi-agent collaborative controller in the step 2) specifically includes:

21) The multi-agent collaborative controller receives the signal input by the perception module and the signal fed back by the reinforcement learning module, and makes preliminary adjustments to the input signals of the wire-controlled hydraulic steering subsystem, wire-controlled hydraulic braking subsystem, and wire-controlled suspension subsystem. ;

22) The wire-controlled hydraulic steering subsystem and wire-controlled hydraulic brake subsystem receive the driver input information and vehicle dynamic load signal adjusted by the multi-agent collaborative controller, and the wire-controlled suspension subsystem receives the multi-agent collaborative controller adjustment The input information of the rear road surface, the vehicle yaw motion signal and the dynamic load signal;

23) The wire-controlled hydraulic steering subsystem controls the lateral movement and yaw movement of the vehicle according to the received information; the wire-controlled hydraulic brake subsystem controls the longitudinal movement of the vehicle according to the received information; the wire-controlled suspension subsystem controls the vehicle's longitudinal movement according to the received information The received information controls the vertical motion of the vehicle; the lateral motion of the vehicle determines the dynamic load, and the longitudinal motion of the vehicle determines the road to be driven, providing new road signals;

24) The multi-agent cooperative controller receives the dynamic load signal of the vehicle, the yaw motion signal and the new road signal, and adjusts the input signal of each subsystem again; the output of each subsystem is used as a feedback input to control itself and other subsystems, each The subsystems coordinate with each other to restore or maintain the vehicle to a balanced state, realizing the coordinated control of the multi-agent-by-wire chassis system.

体现车辆的安全性的车轮的相对动载荷；体现整车舒适性的悬架动扰度。Further, the index of evaluation in described step 3) comprises: the acceleration of the vehicle vertical direction that embodies the smoothness of the whole vehicle

The relative dynamic load of the wheel, which reflects the safety of the vehicle; the suspension disturbance, which reflects the comfort of the whole vehicle.

Further, the step 3) specifically includes: designing an evaluation function of each performance index to evaluate the system output result;

The performance of the designed suspension-by-wire subsystem refers to the evaluation function:

为车辆垂直方向的加速度，|z_wi-z_bi|为各车轮动载荷；In the formula, α ₁ is the evaluation index of suspension performance,

is the acceleration in the vertical direction of the vehicle, |z _wi -z _bi | is the dynamic load of each wheel;

Design the performance index function of the hydraulic steering-by-wire subsystem:

α ₂ =(γ'-γ) ²

时车辆的横摆角速度，γ为车辆当前的横摆角速度；In the formula, α ₂ is the performance index of the steering system, and γ' is the ideal steering yaw rate, namely

is the yaw angular velocity of the vehicle, γ is the current yaw angular velocity of the vehicle;

Design the performance index function of the control-by-wire hydraulic brake subsystem:

α ₃ =(S'-S) ²

In the formula, α ₃ is the performance index of the braking system, S' is the ideal slip rate, S' ≈ 20%, and S is the current slip rate of the wheel;

The comprehensive performance index function designed to evaluate the comprehensive performance of the multi-agent wire-controlled chassis system:

In the formula, e is the comprehensive performance evaluation index of the multi-agent wire-controlled chassis system, and the value of its modulo |e| is used to evaluate the pros and cons of this output result. The effect is good; Q _e is the contribution value matrix, k ₁ is the contribution value of the suspension-by-wire subsystem to the comprehensive performance index, indicating the proportion of the contribution of the suspension-by-wire subsystem to the performance optimization of the whole vehicle in the comprehensive evaluation; k ₂ is the contribution value of the hydraulic steering-by-wire subsystem to the comprehensive performance index, indicating the proportion of the contribution of the hydraulic steering-by-wire subsystem to the overall vehicle performance optimization in the comprehensive evaluation; k ₃ is the hydraulic brake-by-wire subsystem The contribution value to the comprehensive performance index represents the proportion of the contribution of the hydraulic brake-by-wire subsystem to the overall vehicle performance optimization in the comprehensive evaluation.

Further, in the step 3), the dynamic adjustment needs to refer to the comprehensive performance evaluation index e. When the |e| of this action is smaller than the |e| of the previous action, record the |e| value and the output result of this time. On this basis, the output result is adjusted; when the |e| of the current action is greater than the |e| of the previous action, the output result recorded last time is output. This method can ensure that the input and output of the vehicle are always in an ideal state, and constantly optimize itself in new attempts, thereby ensuring the stability, comfort and safety of the vehicle.

Further, the reliability test of the algorithm in the described step 4) is realized by the method of simulation test and expert review, first by computer to carry out simulation test to the algorithm and draw the preliminary analysis result, then by the expert to review the analysis result and for poor reliability. The algorithm gives the corresponding modification opinion as the test result.

Further, the optimization of the algorithm in the step 5) is represented by the vehicle modifying the parameters of the algorithm according to the inspection result fed back by the cloud processing center.

Beneficial effects of the present invention:

1. The present invention is based on the multi-agent system coordinated control theory, realizes the multi-faceted coupling of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem to form a multi-agent wire-controlled chassis system, which greatly improves the performance of the multi-agent wire-controlled chassis system. The improvement of the stability and reliability of the wire-controlled chassis system is of great significance to the improvement of the comfort, safety and smoothness of the commercial vehicle.

2. The present invention adopts reinforcement learning and multi-agent theory, fully considers the dynamic relationship between vertical, longitudinal and lateral directions of commercial vehicles under different working conditions, and adopts the method of cloud processing center to test the reinforcement learning results, The reliability of the system is further ensured, the performance of the whole vehicle is improved, and the working reliability of the multi-agent body-by-wire chassis system is ensured, and the safety of the whole vehicle is improved.

Description of drawings

1 is a schematic diagram of the system structure of the present invention;

Figure 2 is a schematic diagram of a multi-agent cooperative controller control algorithm;

Fig. 3 is the working flow chart of the method of the present invention.

Detailed ways

In order to facilitate the understanding of those skilled in the art, the present invention will be further described below with reference to the embodiments and the accompanying drawings, and the contents mentioned in the embodiments are not intended to limit the present invention.

Referring to FIG. 1 , a multi-agent-based cloud-controlled intelligent chassis system of the present invention includes: a wire-controlled hydraulic steering subsystem, a wire-controlled hydraulic braking subsystem, a wire-controlled suspension subsystem, a perception module, a Learning module, communication module, cloud processing center and multi-agent coordination controller;

The wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem are all distributed on the multi-agent wire-controlled chassis for realizing the control of the vehicle chassis;

The perception module is respectively connected to each vehicle-mounted sensor and the multi-agent coordination controller, and is used to collect information, including: the operation instructions input by the driver, the smoothness of the road surface, the curvature of the road, the lateral speed and acceleration of the vehicle, and the longitudinal speed of the vehicle. And acceleration, vertical acceleration of vehicle, yaw rate of vehicle, body slip angle, suspension damping force, suspension disturbance, wheel dynamic load, wheel slip rate.

The reinforcement learning module evaluates the motion state of the vehicle itself according to the output results of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem, and obtains the complete vehicle according to the current motion state of the vehicle itself. Action behavior with the best stability and continuous reinforcement learning; data connection with the multi-agent coordination controller, and data connection with the cloud processing center through the communication module; the communication module adopts the 5G communication method to realize the vehicle and the cloud processing center data communications;

The cloud processing center is used to regularly check whether the reinforcement learning results of the vehicle are reasonable, and adjust the unreasonable reinforcement learning results;

The multi-agent coordination controller is used to adjust the information input of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem.

The hydraulic steering-by-wire subsystem includes: a steering wheel unit, a steering-by-wire control unit, a hydraulic steering unit, an electric steering unit, and a steering road sensing unit; the rotation signal of the steering wheel unit is input to the steering-by-wire control unit, and the wire The steering control unit coordinates and controls the hydraulic steering unit and the electric steering unit to work together to complete the vehicle steering. The steering road sensing unit extracts vehicle information and road surface state information, and generates a positive torque to act on the steering wheel unit to simulate the real road feeling for the driver;

The wire-controlled hydraulic brake subsystem includes: a wire-controlled brake control unit, a hydraulic brake unit, a regenerative brake unit and a pedal simulation unit; the pedal simulation unit converts the information of the strength of the driver's pedal depression into a corresponding electrical signal And transmit it to the brake-by-wire unit. The brake-by-wire unit controls the hydraulic braking unit to complete the braking operation according to the received electrical signal. The regenerative braking unit stores kinetic energy when the vehicle is braking, and the energy storage motor of the regenerative braking unit is connected to the vehicle. on the axle;

The wire-controlled suspension subsystem includes: an active suspension unit, a suspension control unit; the active suspension unit adjusts the damping and stiffness of the suspension, and the suspension control unit adjusts the damping and stiffness of the suspension according to the vehicle height, vehicle speed, steering angle and rate, and braking The signal is used to control the active suspension unit to adjust suspension damping and stiffness.

The cloud processing center also optimizes the ride comfort and safety of the target vehicle according to the learning results of other vehicles.

The multi-agent coordination controller is used for, according to the current state of the vehicle, the working conditions of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem, the driver's input information, and the smoothness of the road surface. Road curvature, with vehicle safety and comfort as the ultimate goal, controls three subsystems, so as to achieve coordinated control of the entire vehicle chassis.

The dynamic model of the wire-controlled hydraulic steering subsystem includes: lateral motion and yaw motion models, respectively:

where m is the vehicle mass, v is the vehicle speed, γ is the yaw rate, β is the body slip angle, F _yfl is the left front wheel cornering force of the vehicle, F _yfr is the vehicle right front wheel cornering force, F _yrl is the cornering force of the left rear wheel of the vehicle, F _yrr is the cornering force of the right rear wheel of the vehicle, I _s is the yaw moment of inertia of the vehicle, the distance from the center of mass of L _f to the front axle, and the distance from the center of mass of L _r to the rear axle distance.

The wheel rotation dynamics model of the wire-controlled hydraulic brake subsystem is:

In the formula, I ₁ is the moment of inertia of the left front wheel of the vehicle, I ₂ is the moment of inertia of the right front wheel of the vehicle, I ₃ is the moment of inertia of the left rear wheel of the vehicle, I ₄ is the moment of inertia of the right rear wheel of the vehicle, ω ₁ is The rotational angular velocity of the left front wheel of the vehicle, ω ₂ is the rotational angular velocity of the right front wheel of the vehicle, ω ₃ is the rotational angular velocity of the left rear wheel of the vehicle, ω ₄ is the rotational angular velocity of the right rear wheel of the vehicle, and F _xfl is the longitudinal direction of the left front wheel of the vehicle Adhesion, F _xfr is the longitudinal adhesion of the right front wheel of the vehicle, F _xrl is the longitudinal adhesion of the left rear wheel of the vehicle, F _xrr is the longitudinal adhesion of the right rear wheel of the vehicle, R ₁ is the radius of the left front wheel of the vehicle, R ₂ is the radius of the front wheel on the right side of the vehicle, R ₃ is the radius of the rear wheel on the left side of the vehicle, R ₄ is the radius of the rear wheel on the right side of the vehicle, T _b1 is the braking torque of the front wheel on the left side of the vehicle, and T _b2 is the braking torque of the front wheel on the right side of the vehicle. dynamic torque, T _b3 is the braking torque of the left rear wheel of the vehicle, and T _b4 is the braking torque of the right rear wheel of the vehicle;

The wheel slip formula is:

where S is the slip rate of the vehicle, v ₀ is the longitudinal speed of the vehicle, R is the wheel radius, and ω is the rotational angular velocity of the wheel;

The calculation formula of braking torque is:

T _b =(PP ₀ )A _w ηB _F R _g

In the formula, T _b is the braking torque of the brake, P is the pressure of the brake line, P ₀ is the initial pressure of the brake line, A _w is the effective working area of the brake cylinder, η is the efficiency of the brake cylinder, B _F For the braking efficiency factor, R _g is the radius of the brake disc.

The vertical dynamic model at the center of mass of the vehicle body of the suspension-by-wire subsystem is:

The body pitch dynamics model is:

The body roll dynamics model is:

The vertical motion model of the four unsprung mass parts is:

The vertical motion model at the four endpoints of the body is:

为车辆质心处垂直方向的加速度，I_p为簧载质量部分的俯仰转动惯量，θ为簧载质量部分的俯仰角，

为簧载质量部分的俯仰角角加速度，I_r为簧载质量部分的侧倾转动惯量，

为簧载质量部分的侧倾角，

为簧载质量部分的侧倾角角加速度，L_f为车身质心到前轴的距离，L_r为车身质心到后轴的距离， B_f为前轴轮距，B_r为后轴轮距；K_si为第i轮处的悬架刚度，K_ti为第i轮处的轮胎刚度，C_si为第i轮处的悬架阻尼系数，z_wi为第i处出车轮垂向位移，

为第i处出车轮垂向速度，

为第i处出车轮垂向加速度，z_bi为第i轮处车身端点垂向位移，f_i为第i轮处的悬架阻尼力，m_wi为车辆的第i轮处的非簧载质量，z_gi为第i轮处路面输入的垂向位移，i＝1,2,3,4.，1代表左侧前轮，2代表右侧前轮，3代表左侧后轮，4代表右侧后轮。where m _b is the sprung mass of the vehicle, z _b is the vertical displacement at the center of mass of the vehicle body,

is the vertical acceleration at the center of mass of the vehicle, I _p is the pitch moment of inertia of the sprung mass part, θ is the pitch angle of the sprung mass part,

is the pitch angle acceleration of the sprung mass part, I _r is the roll moment of inertia of the sprung mass part,

is the roll angle of the sprung mass part,

is the roll angle acceleration of the sprung mass part, L _f is the distance from the center of mass of the body to the front axle, L _r is the distance from the center of mass of the body to the rear axle, B _f is the wheel track of the front wheel, and B _r is the wheel track of the rear wheel; K _si is the suspension stiffness at the i-th wheel, K _ti is the tire stiffness at the i-th wheel, _Csi is the suspension damping coefficient at the i-th wheel, _zwi is the vertical displacement of the wheel at the i-th wheel,

is the vertical speed of the wheel exiting at the i-th place,

is the vertical acceleration of the wheel at the i-th wheel, z _bi is the vertical displacement of the body end point at the i-th wheel, f _i is the suspension damping force at the i-th wheel, and _mwi is the unsprung mass at the i-th wheel of the vehicle , z _gi is the vertical displacement of the road input at the i-th wheel, i=1, 2, 3, 4., 1 represents the left front wheel, 2 represents the right front wheel, 3 represents the left rear wheel, and 4 represents the right side rear wheel.

The specific control method of the multi-agent coordination controller is as follows: when the road surface is uneven and there is interference, the vehicle will tilt in the vertical direction and the pitch direction, resulting in different forces acting on the suspension subsystem and affecting the stability; When turning, the frictional force of the tire makes it produce yaw motion; when braking, the resulting acceleration will cause the vehicle to produce violent yaw motion and pitch motion, and the balance state will be broken due to the change of load distribution. , reducing the operational stability of the vehicle; the driving state of the vehicle will be affected by the steering angle, pedal braking, accelerator input and ground conditions, and at the same time, the current operating conditions of each subsystem will also affect other subsystems In order to restore the vehicle to a good balance state, it is necessary to adjust the various subsystems at the same time according to the size of the role, and each subsystem cooperates with each other to complete the control of the vehicle chassis, so that the vehicle maintains good safety and comfort. .

Referring to Fig. 3, a control method of a cloud-controlled intelligent chassis system based on multi-agents of the present invention, based on the above system, the steps are as follows:

1) The perception module collects the information on the smoothness and curvature of the road surface, the information input by the driver and the motion information of the vehicle, and transmits the collected information to the multi-agent coordination controller;

各车轮动载荷|z_wi-z_bi|，各悬架动扰度|z_wi-z_gi|，车辆当前的横摆角速度γ，质心侧偏角

车身俯仰角θ，车轮滑移率S，车辆的纵向速度v。The signal transmitted to the multi-agent coordinated controller in the step 1) is based on the damping force f _i of each suspension in the dynamic model established by each subsystem, the vertical acceleration

The dynamic load of each wheel |z _wi -z _bi |, the dynamic disturbance of each suspension |z _wi -z _gi |, the current yaw rate γ of the vehicle, the side slip angle of the center of mass

Body pitch angle θ, wheel slip rate S, vehicle longitudinal speed v.

2) The multi-agent coordinated controller controls each subsystem according to the received above-mentioned information, and each subsystem controls the vehicle chassis cooperatively; as shown in FIG. 2 , the control process of the multi-agent coordinated controller specifically includes:

21) The multi-agent collaborative controller receives the signal input by the perception module and the signal fed back by the reinforcement learning module, and makes preliminary adjustments to the input signals of the wire-controlled hydraulic steering subsystem, wire-controlled hydraulic braking subsystem, and wire-controlled suspension subsystem. ;

22) The wire-controlled hydraulic steering subsystem and wire-controlled hydraulic brake subsystem receive the driver input information and vehicle dynamic load signal adjusted by the multi-agent collaborative controller, and the wire-controlled suspension subsystem receives the multi-agent collaborative controller adjustment The input information of the rear road surface, the vehicle yaw motion signal and the dynamic load signal;

23) The wire-controlled hydraulic steering subsystem controls the lateral movement and yaw movement of the vehicle according to the received information; the wire-controlled hydraulic brake subsystem controls the longitudinal movement of the vehicle according to the received information; the wire-controlled suspension subsystem controls the vehicle's longitudinal movement according to the received information The received information controls the vertical motion of the vehicle; the lateral motion of the vehicle determines the dynamic load, and the longitudinal motion of the vehicle determines the road to be driven, providing new road signals;

24) The multi-agent cooperative controller receives the dynamic load signal of the vehicle, the yaw motion signal and the new road signal, and adjusts the input signal of each subsystem again; the output of each subsystem is used as a feedback input to control itself and other subsystems, each The subsystems coordinate with each other to restore or maintain the vehicle to a balanced state, realizing the coordinated control of the multi-agent-by-wire chassis system.

3) The reinforcement learning module evaluates the current output results of each subsystem according to the set final goal, and dynamically adjusts the next output according to the evaluation results to complete the optimization of the multi-agent collaborative controller;

体现车辆的安全性的车轮的相对动载荷；体现整车舒适性的悬架动扰度。The index evaluated in the described step 3) includes: the acceleration in the vertical direction of the vehicle that embodies the smoothness of the entire vehicle

The relative dynamic load of the wheel, which reflects the safety of the vehicle; the suspension disturbance, which reflects the comfort of the whole vehicle.

The step 3) specifically includes: designing an evaluation function of each performance index to evaluate the system output result;

The performance of the designed suspension-by-wire subsystem refers to the evaluation function:

为车辆垂直方向的加速度，|z_wi-z_bi|为各车轮动载荷；In the formula, α ₁ is the evaluation index of suspension performance,

is the acceleration in the vertical direction of the vehicle, |z _wi -z _bi | is the dynamic load of each wheel;

Design the performance index function of the hydraulic steering-by-wire subsystem:

α ₂ =(γ'-γ) ²

时车辆的横摆角速度，γ为车辆当前的横摆角速度；In the formula, α ₂ is the performance index of the steering system, and γ' is the ideal steering yaw rate, namely

is the yaw angular velocity of the vehicle, γ is the current yaw angular velocity of the vehicle;

Design the performance index function of the control-by-wire hydraulic brake subsystem:

α ₃ =(S'-S) ²

In the formula, α ₃ is the performance index of the braking system, S' is the ideal slip rate, S' ≈ 20%, and S is the current slip rate of the wheel;

The comprehensive performance index function designed to evaluate the comprehensive performance of the multi-agent wire-controlled chassis system:

In the formula, e is the comprehensive performance evaluation index of the multi-agent wire-controlled chassis system, and the value of its modulo |e| is used to evaluate the pros and cons of this output result. The effect is good; Q _e is the contribution value matrix, k ₁ is the contribution value of the suspension-by-wire subsystem to the comprehensive performance index, indicating the proportion of the contribution of the suspension-by-wire subsystem to the performance optimization of the whole vehicle in the comprehensive evaluation; k ₂ is the contribution value of the hydraulic steering-by-wire subsystem to the comprehensive performance index, indicating the proportion of the contribution of the hydraulic steering-by-wire subsystem to the overall vehicle performance optimization in the comprehensive evaluation; k ₃ is the hydraulic brake-by-wire subsystem The contribution value to the comprehensive performance index represents the proportion of the contribution of the hydraulic brake-by-wire subsystem to the overall vehicle performance optimization in the comprehensive evaluation.

In the step 3), the dynamic adjustment needs to refer to the comprehensive performance evaluation index e. When the |e| of this action is smaller than the |e| of the previous action, the |e| value and the output result of this time are recorded, and on this basis Adjust the output result; when the |e| of the current action is greater than the |e| of the previous action, it will be output with the output result recorded last time. This method can ensure that the input and output of the vehicle are always in an ideal state, and constantly optimize itself in new attempts, thereby ensuring the stability, comfort and safety of the vehicle. 4) The communication module regularly uploads the vehicle information and the current control algorithm to the cloud processing center; the cloud processing center checks the reliability of the received algorithm and feeds back the test results;

The reliability test of the algorithm in the described step 4) is realized by the method of simulation test and expert review. First, the algorithm is simulated and tested by the computer and the preliminary analysis result is obtained, and then the analysis result is reviewed by the expert and given for the algorithm with poor reliability. Appropriate amendments shall be made as the test results.

5) The reinforcement learning module further optimizes the control algorithm according to the feedback results of the cloud processing center;

The optimization of the algorithm in the step 5) shows that the vehicle modifies the parameters of the algorithm according to the inspection result fed back by the cloud processing center.

There are many specific application ways of the present invention, and the above are only the preferred embodiments of the present invention. It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, several improvements can be made. These Improvements should also be considered as the protection scope of the present invention.

Claims (4)

1. A control method for a multi-agent-based cloud-controlled intelligent chassis system, the multi-agent-based cloud-controlled intelligent chassis system comprising: a wire-controlled hydraulic steering subsystem, a wire-controlled hydraulic braking subsystem, a wire-controlled Suspension subsystem, perception module, reinforcement learning module, communication module, cloud processing center and multi-agent coordination controller; The wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem are all distributed on the multi-agent wire-controlled chassis for realizing the control of the vehicle chassis; The perception module is respectively connected to each vehicle-mounted sensor and the multi-agent coordination controller for collecting information; The reinforcement learning module evaluates the motion state of the vehicle itself according to the output results of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem, and obtains the complete vehicle according to the current motion state of the vehicle itself. Action behavior with the best stability and continuous reinforcement learning; data connection with the multi-agent coordination controller, and data connection with the cloud processing center through the communication module; The cloud processing center is used to regularly check whether the reinforcement learning results of the vehicle are reasonable, and adjust the unreasonable reinforcement learning results; The multi-agent coordination controller is used to adjust the information input of the wire-controlled hydraulic steering subsystem, the wire-controlled hydraulic brake subsystem and the wire-controlled suspension subsystem; It is characterized in that, the steps are as follows: 1) The perception module collects the information on the smoothness and curvature of the road surface, the information input by the driver and the motion information of the vehicle, and transmits the collected information to the multi-agent coordination controller; 2) The multi-agent coordination controller controls each subsystem according to the above-mentioned information received, and each subsystem controls the vehicle chassis cooperatively; 3) The reinforcement learning module evaluates the current output results of each subsystem according to the set final goal, and dynamically adjusts the next output according to the evaluation results to complete the optimization of the multi-agent collaborative controller; 4) The communication module regularly uploads the vehicle information and the current control algorithm to the cloud processing center; the cloud processing center checks the reliability of the received algorithm and feeds back the test results; 5) The reinforcement learning module further optimizes the control algorithm according to the feedback results of the cloud processing center; The step 3) specifically includes: designing an evaluation function of each performance index to evaluate the system output result; The performance of the designed suspension-by-wire subsystem refers to the evaluation function:

In the formula, α ₁ is the evaluation index of suspension performance,

is the acceleration in the vertical direction of the vehicle, |z _wi -z _bi | is the dynamic load of each wheel; Design the performance index function of the hydraulic steering-by-wire subsystem: α ₂ =(γ'-γ) ² In the formula, α ₂ is the performance index of the steering system, and γ' is the ideal steering yaw rate, namely

is the yaw angular velocity of the vehicle, γ is the current yaw angular velocity of the vehicle; Design the performance index function of the control-by-wire hydraulic brake subsystem: α ₃ =(S'-S) ² In the formula, α ₃ is the performance index of the braking system, S' is the ideal slip rate, S' ≈ 20%, and S is the current slip rate of the wheel; The comprehensive performance index function designed to evaluate the comprehensive performance of the multi-agent wire-controlled chassis system:

In the formula, e is the comprehensive performance evaluation index of the multi-agent wire-controlled chassis system, and the value of its modulo |e| is used to evaluate the pros and cons of this output result. The effect is good; Q _e is the contribution value matrix, k ₁ is the contribution value of the suspension-by-wire subsystem to the comprehensive performance index, indicating the proportion of the contribution of the suspension-by-wire subsystem to the performance optimization of the whole vehicle in the comprehensive evaluation; k ₂ is the contribution value of the hydraulic steering-by-wire subsystem to the comprehensive performance index, indicating the proportion of the contribution of the hydraulic steering-by-wire subsystem to the overall vehicle performance optimization in the comprehensive evaluation; k ₃ is the hydraulic brake-by-wire subsystem The contribution value to the comprehensive performance index represents the proportion of the contribution of the hydraulic brake-by-wire subsystem to the overall vehicle performance optimization in the comprehensive evaluation. 2. the control method of the cloud-controlled intelligent chassis system based on multi-agent according to claim 1, is characterized in that, the algorithm that multi-agent cooperative controller controls in described step 2) specifically comprises: 21) The multi-agent collaborative controller receives the signal input by the perception module and the signal fed back by the reinforcement learning module, and makes preliminary adjustments to the input signals of the wire-controlled hydraulic steering subsystem, wire-controlled hydraulic braking subsystem, and wire-controlled suspension subsystem. ; 22) The wire-controlled hydraulic steering subsystem and wire-controlled hydraulic brake subsystem receive the driver input information and vehicle dynamic load signal adjusted by the multi-agent collaborative controller, and the wire-controlled suspension subsystem receives the multi-agent collaborative controller adjustment The input information of the rear road surface, the vehicle yaw motion signal and the dynamic load signal; 23) The wire-controlled hydraulic steering subsystem controls the lateral movement and yaw movement of the vehicle according to the received information; the wire-controlled hydraulic brake subsystem controls the longitudinal movement of the vehicle according to the received information; the wire-controlled suspension subsystem controls the vehicle's longitudinal movement according to the received information The received information controls the vertical motion of the vehicle; the lateral motion of the vehicle determines the dynamic load, and the longitudinal motion of the vehicle determines the road to be driven, providing new road signals; 24) The multi-agent cooperative controller receives the dynamic load signal of the vehicle, the yaw motion signal and the new road signal, and adjusts the input signal of each subsystem again; the output of each subsystem is used as a feedback input to control itself and other subsystems, each The subsystems coordinate with each other to restore or maintain the vehicle to a balanced state, realizing the coordinated control of the multi-agent-by-wire chassis system. 3. The control method of a multi-agent-based cloud-controlled intelligent chassis system according to claim 1, wherein the signal transmitted to the multi-agent coordination controller in the step 1) is based on the Damping force f _i of each suspension in the dynamic model, vertical acceleration

The dynamic load of each wheel |z _wi -z _bi |, the dynamic disturbance of each suspension |z _wi -z _gi |, the current yaw rate γ of the vehicle, the side slip angle of the center of mass

Body pitch angle θ, wheel slip rate S, vehicle longitudinal speed v. 4. the control method of the cloud-controlled intelligent chassis system based on multi-agent according to claim 1, is characterized in that, the index that evaluates in described step 3) comprises: the acceleration of the vehicle vertical direction that embodies the smoothness of the whole vehicle

The relative dynamic load of the wheel, which reflects the safety of the vehicle; the suspension disturbance, which reflects the comfort of the whole vehicle.

Images