CN108725453A - Control system and its switch mode are driven altogether based on pilot model and manipulation the man-machine of inverse dynamics - Google Patents

Abstract

The invention discloses a kind of based on pilot model and manipulates the man-machine of inverse dynamics and drives control system and its switch mode altogether, including sensing system, driving intention identification model, man-machine drive model, handover control system with sovereign right altogether and execute system.Driving intention identification model is passed to by sensing system monitoring running environment information, car status information and driver status information, wherein driver status signal first；Secondly it establishes and drives model altogether based on manipulation inverse dynamics and the man-machine of pilot model, calculate desired handle input and the handle input of driver；It is established based on this and man-machine drives switching law altogether, system carries out necessary amendment on the basis of understanding driver's driving intention to it, when driver is temporarily lost with control ability or auxiliary system is temporarily lost with back work ability, controlled by another party's full powers, avoid it is man-machine between generate unnecessary interference.The present invention reduces driver's operating burden, avoids man-machine alternating uncoordinated.

Description

Human-machine co-driving control system and its cut-off based on driver model and handling inverse dynamics change mode

technical field

The invention belongs to the technical field of automobile safety, and in particular relates to a man-machine co-driving control system based on a driver model and manipulation inverse dynamics and a switching mode thereof.

Background technique

The popularization of unmanned vehicles can greatly reduce accidents caused by human factors such as driver fatigue and improper operation. According to SAE's grading standards, the development of automobile intelligence can be divided into five levels: fully human driving, driving assistance, partial automatic driving, conditional automatic driving, highly automatic driving and fully automatic driving. Although unmanned driving technology has received long-term attention, and its research has made great progress, but from the perspective of practical promotion and mass application, unmanned vehicles will face legal constraints if they want to become a means of transportation for human beings. , insurance, technical difficulty, accident liability, driving pleasure and other issues, the complex road traffic environment also determines that unmanned driving is impossible to achieve in the short term. The situation of joint control of man-machine.

With the improvement of automobile intelligence and the continuous development of driving assistance systems, the relationship between automobiles and drivers has become very complicated. Various active vehicle control systems based on environmental information perception and drivers with individual differences constitute the For the parallel binary control of smart cars, there is a dynamic interaction between man and machine. The improvement of automobile intelligence and the expansion of autonomous authority lead to the inevitable coupling and restriction between the intention of the car and the intention of people. Automobiles are products with strong individual needs, and users' acceptance of autonomous decision-making and control of automobiles is an important indicator to measure the value of automobiles. Therefore, the establishment of a human-machine co-driving control system and its switching mode is a key problem to be solved in the development of smart cars.

A good human-machine co-driving control system should adaptively switch the driving rights between man and machine according to the driver's behavior and vehicle safety situation on the premise of ensuring the safety of the car movement, so as to realize the man-machine cooperative driving. The human-machine co-driving control system and its switching mode designed in the present invention is a dual-driving single-control system in which the driver takes control. The man-machine co-driving control system recognizes the driving intention and estimates the driver's control amount. Carry out auxiliary control that conforms to the driver's intention and desired control objectives. Shared control with the driver is achieved with as little intervention as possible, giving the driver more freedom to control the car alone.

Contents of the invention

Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides a human-machine co-driving control system and its switching mode based on the driver model and manipulation inverse dynamics, so as to realize the interaction with the driver with as little intervention as possible. Shared control reduces the driver's operating burden and avoids uncoordinated man-machine alternation.

The invention is a man-machine co-driving control system and its switching mode based on driver model and manipulation inverse dynamics. The human-machine co-driving control system includes a sensor system, a driving intention recognition model, a human-machine co-driving model, a sovereign switching control system, and an execution system. First, the sensor system monitors the driving environment information, vehicle status information, and driver status information, in which the driver status signal is transmitted to the driving intention recognition model; secondly, a human-machine co-driving model based on the inverse dynamics of manipulation and the driver model is established to calculate The expected manipulation input and the driver's manipulation input are determined; and based on this, the human-machine co-driving switching rule is established. The system makes necessary corrections on the basis of understanding the driver's driving intention. When the driver temporarily loses control ability or the auxiliary system When the auxiliary working ability is temporarily lost, the other party has full control to avoid unnecessary interference between man and machine. The human-machine co-driving control system and its switching mode designed in the present invention can not only be applicable to the current legal conditions and environmental perception technology level, but also can further reduce the driver's operation burden and avoid the incoordination of human-machine alternation. The invention is easy to realize and has good application prospect.

Technical scheme: in order to achieve the above object, the technical scheme adopted in the present invention is:

A human-machine co-driving control system based on a driver model and manipulation inverse dynamics, including a sequentially connected sensor system, a driving intention recognition model, a human-machine co-driving model, a sovereign switching control system, and an execution system, wherein,

The sensor system includes:

The driver state perception module is used to transmit the driver state signal to the driving intention recognition model,

The vehicle driving state perception module is used to transmit the vehicle driving state signal to the human-machine co-driving model,

The driving environment perception module is used to transmit the driving environment signal to the human-machine co-driving model;

The driving intention recognition model is used to identify the driving intention according to the driver's state signal, and transmit it to the human-machine co-driving model;

The human-machine co-driving model includes a vehicle-road model, a driver model, and a manipulation inverse dynamics model. The vehicle-road model receives vehicle driving state signals, driving environment signals, and driving intentions, and establishes a system based on the driver model and driving intention on this basis. Manipulate the man-machine co-driving model of inverse dynamics, calculate the expected manipulation input and the driver's manipulation input, and transmit it to the sovereign switching control system;

The sovereign switching system includes a mode switching decision-making layer, which is used to judge and decide the driving mode, and transmit the decision signal to the execution system;

The execution system includes a steering motor and a controller. The controller commands the steering motor to perform an action by controlling the current output rotation angle of the steering motor.

Further, the driver state perception module includes a steering wheel angle/torque sensor and a brake pedal sensor for monitoring driver action state signals;

The vehicle driving state perception module includes a wheel sensor, a sideslip sensor, a lateral acceleration sensor, and a center-of-mass side slip angle sensor for monitoring vehicle driving state signals;

The driving environment perception module includes a visual recognition sensor, a radar sensor, an ultrasonic sensor, and an infrared sensor for monitoring driving environment signals.

Further, the driving intention recognition model includes an HMM model and an SVM model.

Further, the expected manipulation input and the driver's manipulation input are respectively an ideal steering wheel angle signal and an actual driver's steering wheel angle signal.

Further, the driving modes include driver-led mode, machine-assisted mode and automatic driving mode.

At the same time, the present invention also provides the switching mode of the above-mentioned human-machine co-driving control system based on the driver model and manipulation inverse dynamics, including the following steps:

Step 1), the sensor system perceives and monitors the driving environment signal, the vehicle state signal and the driver state signal, transmits the driver state signal to the driving intention recognition model, and transmits the vehicle driving state signal and the driving environment signal to the human-machine co-driving model ;

Step 2), after the driving intention recognition model receives the driver's state signal, it makes a preliminary judgment on the driver's intention through the HMM model, and outputs the maximum likelihood estimation value; the SVM model performs the second recognition according to the output maximum likelihood estimation value to obtain More accurate driver intentions and pass them on to the vehicle-road model;

Step 3), after the vehicle-road model receives the driver's intention signal, combined with the vehicle driving state signal and driving environment signal obtained by the sensor system, it is processed and calculated and transferred to the driver model and the handling inverse dynamics model to calculate the driver's The output driving control and control inverse dynamics model output the optimal control, respectively refer to the driver output steering wheel angle and expected steering wheel angle, and transmit the signal to the sovereign switching decision system;

Step 4), after the sovereign switching decision-making system receives the signal, the mode-switching decision-making layer takes the received driving manipulation signal and optimal manipulation signal as the decision-making object, takes the driving mode as the decision-making result, formulates the sovereign switching principle, and makes a judgment , to transmit the decision result signal to the execution system;

Step 5), the controller in the execution system enters the corresponding driving state after receiving the mode switching decision instruction, and takes the steering wheel angle as the control target and the steering motor angle as the control object, calculates the corresponding steering motor angle, and controls The command is passed to the steering motor actuator;

Step 6), the steering motor in the execution system executes an action after receiving the command to realize steering control.

Further, in step 3), the step of manipulating the inverse dynamics model includes:

Step 3.1), establishing a three-degree-of-freedom vehicle model;

In the formula, v is the lateral velocity of the vehicle; u is the longitudinal velocity of the vehicle; ω _r is the yaw rate of the vehicle; m is the total mass of the vehicle; I _z is the moment of inertia of the vehicle around the vertical axis; is the distance from the center of mass of the vehicle to the front and rear axles; δ is the front wheel rotation angle; F _yf is the cornering force of the front wheels; F _yr is the cornering force of the rear wheels; F _xf is the driving force/braking force of the front wheels, F _xf ≥ 0 is the driving force, F _xf <0 is the braking force; F _xr is the rear wheel driving force/braking force; F _f is the rolling resistance, and F _f = mgf, f is the rolling resistance coefficient; F _w is the air resistance, F _w ＝C _D A(3.6u) ² /21.15, _{C D} is the air resistance coefficient, A is the windward area;

Step 3.2), establishing an optimal control model;

State variable x(t)＝[v(t) ω(t) u(t) x(t) y(t) θ(t)] ^T , x is lateral displacement; y is longitudinal displacement; θ is heading angle;

The control variable Z(t) is steering wheel angle δ _sw (t) and front wheel driving/braking force F _xf (t);

The control task is to pass the given path in the shortest time;

Step 3.3), solving the optimal control problem;

Step 3.4), obtain the optimal manipulation input, including the expected steering wheel angle δ ^* .

Further, in step 3.3), the method for solving the optimal control problem is: converting the solution of the optimal control problem transformed from the manipulative inverse dynamics problem into the solution of the nonlinear programming problem:

C[X(τ _k ), Z(τ _k ),τ _k ; t ₀ ,t _e ]≤0

Among them, the initial point τ ₀ =-1, τ _k is the LG point, t ₀ is the initial point before the time interval conversion; t _e is the end point before the time interval conversion; τ _e is the end point after the time interval conversion.

Further, in the mode switching decision-making layer, the sovereign switching principle of the sovereign switching module is as follows:

Assuming that the driver turns the steering wheel to the right is positive, set the thresholds δ _max and δ _min as the upper limit and lower limit of the steering wheel angle respectively. When the vehicle is in the steering state, the monitoring driver model outputs the steering wheel angle The value of the ideal steering wheel angle δ ^* output;

1) If |δ|>δ _max , it means that the driver is in a highly concentrated state at this time and enters the driver-dominant mode, and the motor does not perform any operation;

2) If δ _min <|δ|<δ _max , according to the ideal steering wheel angle δ ^* output by the vehicle handling inverse dynamics model and the direction of the steering wheel angle δ output by the driver model, taking the car currently turning right as an example, Divided into the following states:

State 1: If δ _min <δ<δ ^* , it means that the driver and the controller cooperate to control the car to turn right at this moment. Since the steering wheel angle provided by the driver is not enough to make the car complete the steering operation, it enters the machine-assisted mode at this time, and the motor always provide assistance;

State 2: If δ ^* >0, δ<-δ _min , it is judged that the driver made an error at the moment, and the controller takes control and enters the automatic driving mode;

3) If |δ|<δ _min , it is considered that the driver is not paying attention, at this time, the controller takes control and enters the automatic driving mode.

Beneficial effects: Compared with the prior art, the man-machine co-driving control system and its switching mode based on the driver model and manipulation inverse dynamics provided by the present invention have the following advantages:

1. According to the driver's behavior intention and vehicle safety situation, the driving right between man and machine can be adaptively switched to realize man-machine cooperative driving. And share control with the driver with as little intervention as possible, so as to give the driver more freedom to control the car alone

2. The human-machine co-driving model can better simulate the man-machine co-driving, and simulate and verify the designed man-machine co-driving control system.

Description of drawings

Fig. 1 is the overall block diagram of man-machine co-driving control system and switching mode thereof in the present invention;

Fig. 2 is the principle diagram of automobile handling inverse dynamics method in the present invention;

Fig. 3 is the flow chart of automobile handling inverse dynamics model in the present invention;

FIG. 4 is a flow chart of the method for switching modes in the present invention.

Detailed ways

The invention is a man-machine co-driving control system and its switching mode based on driver model and manipulation inverse dynamics. First, the sensor system detects the driver's action status signal, the vehicle driving status signal, and the driving environment signal. Among them, the driver's state signal is transmitted to the driving intention recognition model, which is used to identify the driving intention and transmitted to the vehicle road model. At the same time, the vehicle driving state signal and driving environment signal are transmitted to the vehicle road model. Through the vehicle road model, driver model and handling inverse dynamics model, the actual driver steering wheel angle signal and the ideal steering wheel angle signal are obtained and transmitted to the sovereign switching system. The sovereign switching system includes a switching decision-making module, which is used to judge and make decisions about the driving mode, and transmit the decision-making signal to the executive system. The execution system includes a steering motor and a controller. The controller commands the steering motor to perform actions by controlling the current output rotation angle of the steering motor.

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

As shown in Figure 1, the working steps of the present invention are:

Step 1), the sensor system perceives and monitors the driving environment signal, the vehicle state signal and the driver state signal, wherein the driver state signal is transmitted to the driving intention recognition model, and the vehicle driving state signal and the driving environment signal are transmitted to the human-machine co-driving Model.

Step 2), after the driving intention recognition model receives the driver's state signal, it makes a preliminary judgment on the driver's intention through the HMM model, and outputs the maximum likelihood estimation value; the SVM model performs the second recognition according to the output maximum likelihood estimation value to obtain More accurate driver intentions and pass them on to the vehicle-road model;

Step 3), after the vehicle-road model receives the driver's intention signal, combined with the vehicle driving state signal and driving environment signal obtained by the sensor system, it is processed and calculated and transferred to the driver model and the handling inverse dynamics model, thereby calculating The driving control output by the driver and the optimal control output by the control inverse dynamics model refer to the steering wheel angle output by the driver and the expected steering wheel angle respectively, and the signals are transmitted to the sovereign switching decision system;

Step 4), after the sovereign switching decision-making system receives the signal, the switching decision-making module takes the received driving manipulation signal (the driver's output steering wheel angle), the optimal manipulation signal (the expected steering wheel angle) as the decision object, and the driving mode ( Driver-led mode, machine-assisted mode, automatic driving mode) as the decision result, formulate the principle of sovereign switching, make judgments, and pass the decision result signal to the execution system;

Step 5), the controller in the execution system enters the corresponding driving state after receiving the mode switching decision command, and controls the target with the steering wheel angle, takes the steering motor angle as the control object, calculates the corresponding steering motor angle, and sends the control command Passed to the steering motor actuator;

Step 6), the steering motor in the execution system executes an action after receiving the command to realize steering control.

As shown in Figures 2 and 3, the specific steps for manipulating the inverse dynamics model include:

Step 3.1), establishing a three-degree-of-freedom vehicle model;

In the formula, v is the lateral velocity of the vehicle; u is the longitudinal velocity of the vehicle; ω _r is the yaw rate of the vehicle; m is the total mass of the vehicle; I _z is the moment of inertia of the vehicle around the vertical axis; is the distance from the center of mass of the vehicle to the front and rear axles; δ is the front wheel rotation angle; F _yf is the cornering force of the front wheels; F _yr is the cornering force of the rear wheels; F _xf is the driving force/braking force of the front wheels (F _xf ≥ 0 is the driving force, F _xf <0 is the braking force); F _xr is the rear wheel driving force/braking force; F _f is the rolling resistance (F _f = mgf, f is the rolling resistance coefficient); F _w is the air resistance (F _w ＝C _D A(3.6u) ² /21.15, _{C D} is the air resistance coefficient, A is the windward area).

If the influence of driving force/braking force on cornering force is considered, then:

In the formula, F _zf is the vertical force of the front wheel; F _zr is the vertical force of the rear wheel; k ₁ and k ₂ are the comprehensive cornering stiffness of the front and rear wheels respectively.

Considering the longitudinal load transfer, there are:

In the formula, h _g is the height of the center of mass of the car.

Step 3.2), establishing an optimal control model

State variable x(t)=[v(t) ω(t) u(t) x(t) y(t) θ(t)] ^T , x is the lateral displacement; y is the longitudinal displacement; θ is the heading angle. The control variable Z(t) is the steering wheel angle δ _sw (t) and the front wheel driving force/braking force F _xf (t), and the control task is to pass the given path in the shortest time.

Restrictions:

(1) Boundary value constraints

The initial values are as follows:

x(0)=[0,0,u ₀ ,0,0,0] ^T

(2) Process constraints

Considering preventing the car from rolling over while avoiding obstacles, the following process constraints are established:

In the formula, L is the wheelbase, and K is the stability factor.

When the car is driven and is front wheel drive, there are:

When the car is subjected to braking force, and assuming that the front and rear wheels are locked, there are:

(3) Control variable and state variable constraints

Affected by factors such as vehicle performance and road conditions, the vehicle must satisfy certain state variables and control variable constraints to ensure the smooth completion of the steering process. Therefore, the following constraints are established:

u _min ≤ u ≤ u _max

δ _min ≤ δ ≤ δ _max

Step 3.3), solving the optimal control problem

Step 3.2.1), using the Gauss pseudospectral method to convert the optimal control problem into a nonlinear programming problem

(1) Summarize the above inverse dynamics problem into an optimal control problem with Mayer type as the optimization objective:

min J＝ψ(x(t _e ),t _e )

C[x(t),z(t),t]≤0

(2) Interval transformation

Convert the time interval t∈[t ₀ ,t _e ] of the optimal control problem into τ∈[-1,1], and transform the time variable t:

τ=2t/(t _e -t ₀ )-(t _e +t ₀ )/(t _e -t ₀ ),

Available

min J＝ψ(x(τ _e ),t _e )

C[x(τ),z(τ),τ;t ₀ ,t _e ]≤0

(3) Global interpolation polynomial approximation of state variables and control variables

The Gauss pseudospectral method selects N LG points and an initial point τ ₀ =-1 as nodes, constructs N+1 Lagrange interpolation polynomials L _i (τ)(i=0,…,N), and uses this as the basis function approximate state variable

Among them, the Lagrange interpolation polynomial function

Make the approximate state on the node equal to the actual state, that is, x(τ _i )=X(τ _i ), (i=0,...,N).

Use the Lagrange interpolation polynomial L ^* _i (τ), (i=1,...,N) as the basis function to approximate the control variable, namely:

In the formula, τ _i (i=1,...,N) is an LG point.

(3) Conversion of kinematic differential equation constraints into algebraic constraints

Transform the kinetic differential equation constraints into algebraic constraints, namely:

Where the differential matrix D _ki ∈ R ^N×(N+1) is a constant value when the number of interpolation nodes is given, the expression is:

Among them, τ _k (k=1,...,N) is a point in the set κ, and τ _i (i=0,...,N) belongs to the set κ ₀ ={τ ₀ ,τ ₁ ,...,τ _N }. Discrete at the interpolation node τ _k (1≤k≤N). In this way, the dynamic differential equation constraints of the optimal control problem can be converted into algebraic constraints, for k=1,...,N:

(4) Terminal state constraints under discrete conditions

The terminal state should also satisfy the kinetic equation constraints:

Discretizing the terminal constraints and approximating them with Gauss integrals gives:

in is the Gauss weight, and τ _k is the LG point.

The following formula can be obtained by discretizing the boundary value constraint and path constraint at the interpolation point:

C[X(τ _k ), Z(τ _k ),τ _k ; t ₀ ,t _e ]≤0

After the above transformation, the solution of the optimal control problem transformed from the vehicle handling inverse dynamics problem is transformed into the solution of the nonlinear programming problem.

Step 3.2.1), using SQP to solve the nonlinear programming problem

The general nonlinear constraint optimal control problem is as follows

In the formula, f(x), _ci (x) are real-valued continuous functions and at least one of them is nonlinear, I={m _e +1,...,m}, constructor problem

in g _k is the gradient of the function f(x) at point x _k , and B _k is the approximation of the Lagrange function's sea color matrix. Note that the solution of the above sub-problem is d _k , the Wilson-Han-Powell method used in this patent is to use d _k as the search direction of the kth iteration, which is the descending direction of many penalty functions.

Stepwise quadratic programming algorithm steps:

(1) Given x ₁ ∈ R ⁿ , σ＞0, δ＞0, B ₁ ∈ R ^n×n , ε≥0, k:=1

(2) Solve the above sub-problems to obtain d _k ; if ||d _k ||≤ε, stop; find α _k ∈ [0, δ] such that

(3) x _k+1 = x _k +α _k d _k ; calculate B _k+1 ; k:=k+1; go to step 2.

The penalty function P(x,σ) is an exact penalty function of L ₁ , ε _k is a non-negative sequence and satisfies

Use the quasi-Newton formula to iterate step by step to calculate B _k+1 , take

s _k =x _k+1 -x _k ,

Calculate B _k+1 using BFGS correction formula

Step 3.4), obtain the optimal manipulation input, including the expected steering wheel angle δ ^* .

As shown in Figure 4, the sovereign switching principle of the sovereign switching module is as follows:

Assuming that the driver turns the steering wheel to the right is positive, set the thresholds δ _max , δ _min , which are the upper limit and lower limit of the steering wheel angle respectively. When the vehicle is in the steering state, monitor the driver model to output the steering wheel angle Output the value of ideal steering wheel angle δ ^* .

1) If |δ|>δ _max , it means that the driver is in a highly concentrated state at this time and enters the driver-dominant mode, and the motor does not perform any operation;

2) If δ _min <|δ|<δ _max , according to the ideal steering wheel angle δ ^* output by the vehicle handling inverse dynamics model and the direction of the steering wheel angle δ output by the driver model, taking the car currently turning right as an example, Divided into the following states:

State 1: If δ _min <δ<δ ^* , it means that the driver and the controller cooperate to control the car to turn right at this moment. Since the steering wheel angle provided by the driver is not enough to make the car complete the steering operation, it enters the machine-assisted mode at this time, and the motor always Provide assistance.

State 2: If δ ^* >0, δ<-δ _min , it is determined that the driver made an error at the moment, and the controller takes control and enters the automatic driving mode.

When the car turns left, the switching process of the sovereign switching module is similar to this.

3) If |δ|<δ _min , it is considered that the driver is not paying attention, at this time, the controller takes control and enters the automatic driving mode.

The above is only a preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications are also possible. It should be regarded as the protection scope of the present invention.

Claims (9)

1. A human-machine co-driving control system based on a driver model and manipulation inverse dynamics, characterized in that it includes a sequentially connected sensor system, a driving intention recognition model, a human-machine co-driving model, a sovereign switching control system, and an execution system ,in, The sensor system includes: The driver state perception module is used to transmit the driver state signal to the driving intention recognition model, The vehicle driving state perception module is used to transmit the vehicle driving state signal to the human-machine co-driving model, The driving environment perception module is used to transmit the driving environment signal to the human-machine co-driving model; The driving intention recognition model is used to identify the driving intention according to the driver's state signal, and transmit it to the human-machine co-driving model; The human-machine co-driving model includes a vehicle-road model, a driver model, and a manipulation inverse dynamics model. The vehicle-road model receives vehicle driving state signals, driving environment signals, and driving intentions, and establishes a system based on the driver model and driving intention on this basis. Manipulate the man-machine co-driving model of inverse dynamics, calculate the expected manipulation input and the driver's manipulation input, and transmit it to the sovereign switching control system; The sovereign switching system includes a mode switching decision-making layer, which is used to judge and decide the driving mode, and transmit the decision signal to the execution system; The execution system includes a steering motor and a controller. The controller commands the steering motor to perform an action by controlling the current output rotation angle of the steering motor. 2. The human-machine co-driving control system based on the driver model and manipulation inverse dynamics according to claim 1, wherein the driver state perception module includes a steering wheel angle/torque sensor, a brake pedal sensor, Used to monitor the driver's action status signal; The vehicle driving state perception module includes a wheel sensor, a sideslip sensor, a lateral acceleration sensor, and a center-of-mass side slip angle sensor for monitoring vehicle driving state signals; The driving environment perception module includes a visual recognition sensor, a radar sensor, an ultrasonic sensor, and an infrared sensor for monitoring driving environment signals. 3. The human-machine co-driving control system based on driver model and manipulation inverse dynamics according to claim 1, characterized in that: the driving intention recognition model includes an HMM model and an SVM model. 4. The man-machine co-driving control system based on driver model and manipulation inverse dynamics according to claim 1, characterized in that: the expected manipulation input and the driver's manipulation input are respectively the ideal steering wheel angle and the actual steering wheel angle. The driver's steering wheel angle signal. 5. The human-machine co-driving control system based on driver model and manipulation inverse dynamics according to claim 1, wherein the driving modes include driver-led mode, machine-assisted mode and automatic driving mode. 6. According to any one of claims 1 to 5, the switching mode of the human-machine co-driving control system based on the driver model and manipulation inverse dynamics is characterized in that: it comprises the following steps: Step 1), the sensor system perceives and monitors the driving environment signal, the vehicle state signal and the driver state signal, transmits the driver state signal to the driving intention recognition model, and transmits the vehicle driving state signal and the driving environment signal to the human-machine co-driving model ; Step 2), after the driving intention recognition model receives the driver's state signal, it makes a preliminary judgment on the driver's intention through the HMM model, and outputs the maximum likelihood estimation value; the SVM model performs the second recognition according to the output maximum likelihood estimation value to obtain More accurate driver intentions and pass them on to the vehicle-road model; Step 3), after the vehicle-road model receives the driver's intention signal, combined with the vehicle driving state signal and driving environment signal obtained by the sensor system, it is processed and calculated and transferred to the driver model and the handling inverse dynamics model to calculate the driver's The output driving control and control inverse dynamics model output the optimal control, respectively refer to the driver output steering wheel angle and expected steering wheel angle, and transmit the signal to the sovereign switching decision system; Step 4), after the sovereign switching decision-making system receives the signal, the mode-switching decision-making layer takes the received driving manipulation signal and optimal manipulation signal as the decision-making object, takes the driving mode as the decision-making result, formulates the sovereign switching principle, and makes a judgment , to transmit the decision result signal to the execution system; Step 5), the controller in the execution system enters the corresponding driving state after receiving the mode switching decision instruction, and takes the steering wheel angle as the control target and the steering motor angle as the control object, calculates the corresponding steering motor angle, and controls The command is passed to the steering motor actuator; Step 6), the steering motor in the execution system executes an action after receiving the command to realize steering control. 7. The switching mode of the man-machine co-driving control system based on the driver model and manipulation inverse dynamics according to claim 6, characterized in that: in step 3), the step of the manipulation inverse dynamics model comprises: Step 3.1), establishing a three-degree-of-freedom vehicle model; In the formula, v is the lateral velocity of the vehicle; u is the longitudinal velocity of the vehicle; ω _r is the yaw rate of the vehicle; m is the total mass of the vehicle; I _z is the moment of inertia of the vehicle around the vertical axis; is the distance from the center of mass of the vehicle to the front and rear axles; δ is the front wheel rotation angle; F _yf is the cornering force of the front wheels; F _yr is the cornering force of the rear wheels; F _xf is the driving force/braking force of the front wheels, F _xf ≥ 0 is the driving force, F _xf < 0 is the braking force; F _xr is the rear wheel driving force/braking force; F _f is the rolling resistance, and F _f = mgf, f is the rolling resistance coefficient; F _w is the air resistance, F _w ＝C _D A(3.6u) ² /21.15, _{C D} is the air resistance coefficient, A is the windward area; Step 3.2), establishing an optimal control model; State variable x(t)＝[v(t) ω(t) u(t) x(t) y(t) θ(t)] ^T , x is lateral displacement; y is longitudinal displacement; θ is heading angle; The control variable Z(t) is steering wheel angle δ _sw (t) and front wheel driving/braking force F _xf (t); The control task is to pass the given path in the shortest time; Step 3.3), solving the optimal control problem; Step 3.4), obtain the optimal manipulation input, including the expected steering wheel angle δ ^* . 8. The switching mode of the man-machine co-driving control system based on the driver model and manipulation inverse dynamics according to claim 7, characterized in that: step 3.3), the method for solving the optimal control problem is: the manipulation inverse dynamics The solution of the optimal control problem transformed from the scientific problem into the solution of the nonlinear programming problem: C[X(τ _k ), Z(τ _k ),τ _k ; t ₀ ,t _e ]≤0 Among them, the initial point τ ₀ =-1, τ _k is the LG point, t ₀ is the initial point before the time interval conversion; t _e is the end point before the time interval conversion; τ _e is the end point after the time interval conversion. 9. The switching mode of the human-machine co-driving control system based on the driver model and manipulation inverse dynamics according to claim 6, characterized in that: in the mode switching decision-making layer, the sovereign switching principle of the sovereign switching module is as follows: Assuming that the driver turns the steering wheel to the right is positive, set the thresholds δ _max and δ _min as the upper limit and lower limit of the steering wheel angle respectively. When the vehicle is in the steering state, the monitoring driver model outputs the steering wheel angle The value of the ideal steering wheel angle δ ^* output; 1) If |δ|>δ _max , it means that the driver is in a highly concentrated state at this time and enters the driver-dominant mode, and the motor does not perform any operation; 2) If δ _min <|δ|<δ _max , according to the ideal steering wheel angle δ ^* output by the vehicle handling inverse dynamics model and the direction of the steering wheel angle δ output by the driver model, taking the car currently turning right as an example, Divided into the following states: State 1: If δ _min <δ<δ ^* , it means that the driver and the controller cooperate to control the car to turn right at this moment. Since the steering wheel angle provided by the driver is not enough to make the car complete the steering operation, it enters the machine-assisted mode at this time, and the motor always provide assistance; State 2: If δ ^* >0, δ<-δ _min , it is judged that the driver made an error at the moment, and the controller takes control and enters the automatic driving mode; 3) If |δ|<δ _min , it is considered that the driver is not paying attention, at this time, the controller takes control and enters the automatic driving mode.