Finite-time stable control method for autonomous and safe motion speed of unmanned vehicle
By designing autonomous safe motion speed learning methods and building motion speed tracking error model and path tracking controller, the problem of autonomous driving stable motion in a limited time is solved, safe speed planning and path tracking are realized, and unmanned driving is improved The stability and safety of the car.
Patent Information
- Application Number
- CN202410572459.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-10
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-05-10
AI Technical Summary
It is difficult for driverless cars to achieve stable motion in a limited time, resulting in excessive tracking errors, which may cause collisions, and lack of autonomous safe speed planning methods to improve travel efficiency and safety.
By designing an autonomous safe motion speed learning method, using the reward and punishment value function and the longitudinal motion state equation and lateral dynamic model of the driverless car, a motion speed tracking error model and path tracking error system are constructed, and a longitudinal speed tracking controller and lateral path are designed. Tracking controller to enable driverless cars to track expected paths and safe movements stably for a limited time.
It improves the movement stability and safety of driverless cars, ensures accurate tracking of expected paths and independently plan safety speeds within a limited time, reduces collision risks and improves travel efficiency.
Smart Images

Figure CN118689095B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of driverless vehicle control, and in particular to a limited time stable control method for the autonomous safety motion speed of a driverless vehicle. Background Art
[0002] With the rapid development of the vehicle industry, cars have gradually become an important means of transportation for people to travel, but they have also brought huge challenges to people's life safety and road congestion. Therefore, society urgently needs a more efficient automobile driving technology to improve people's safe travel and reduce urban congestion. As a highly intelligent product of the integration of high-tech, driverless cars are not only the future development direction of the vehicle industry, but also an important way to solve traffic congestion and improve travel efficiency. In practical applications, vehicle motion control is one of the core technologies of driverless cars. It is directly related to whether the vehicle can travel according to the expected path and speed. It is the most basic and important part of driverless driving. Therefore, studying the stability control method of driverless cars is of great significance to ensure the accuracy, safety and stability of vehicle driving.
[0003] At present, some research results have been achieved in the tracking control method of unmanned vehicles, but the stability problem of vehicle movement in a limited time has not been solved. If the stability time adjustment during the movement of the unmanned vehicle is too long, a large tracking error will be generated in the transient movement stage of tracking the desired path, causing the unmanned vehicle to be in danger of collision; at the same time, in order to solve road congestion and improve travel efficiency, the unmanned vehicle should have the ability to autonomously plan the safe movement speed. So far, there is no planning method for the autonomous safe speed of unmanned vehicles and a finite time stability control method. Summary of the invention
[0004] In order to solve the above problems, the present invention provides a finite time stable control method for the autonomous safe movement speed of an unmanned vehicle, aiming to improve the safety and stability of the movement of the unmanned vehicle.
[0005] The present invention is achieved through the following technical solutions:
[0006] The method for finite time stable control of autonomous safety motion speed of an unmanned vehicle is characterized by comprising the following steps:
[0007] Step 1: Using the distance status information between the driverless car and the vehicle in front, an autonomous safety movement speed learning method is designed. Through the reward and punishment value function, the driverless car realizes safety movement speed planning;
[0008] Step 2: Apply the longitudinal motion state equation of the driverless vehicle and construct a motion speed tracking error model through the planned safe motion speed;
[0009] Step 3: Based on the motion speed tracking error model, design the longitudinal speed tracking controller to complete the speed tracking and motion speed constraint performance of the driverless vehicle;
[0010] Step 4: Calculate the longitudinal motion speed based on the unmanned vehicle speed tracking, and apply the lateral dynamics model of the unmanned vehicle to build a lateral path tracking error system;
[0011] Step 5: Based on the lateral path tracking error system, a finite-time lateral path tracking controller is designed to enable the unmanned vehicle to stably track the desired motion path within a finite time.
[0012] In step 1, the distance status information between the driverless car and the front car is used , , describing the speed adjustment action of the driverless car as follows: ,in, represents the distance between the driverless car and the car in front, Indicates the safe driving distance specified between the driverless car and the vehicle in front. and They represent the speed of the driverless car at the current moment and the next moment respectively. Indicates the change in the speed adjustment of the driverless car.
[0013] make , the reward and penalty value function for adjusting the speed of the driverless car is designed as: ,The learning method of autonomous and safe motion speed of the unmanned vehicle is as follows:
[0014] Step a. Initialize the driverless car behavior pair ,in Indicates the current distance status information. Indicates the current speed adjustment action; sets the learning rate of the driverless car's update state , attenuation coefficient , the selection probability of the speed adjustment action ;
[0015] Step b. The driverless car senses the distance to the vehicle in front and with Compare the size and determine the distance status information between the driverless car and the car in front , the driverless car adjusts the probability of action selection based on speed Select The speed adjustment action in the process determines the behavior of the driverless car at the current moment. ;
[0016] Step c. The driverless car Get the reward and penalty value and enter the next distance status information , and then use the speed to adjust the probability of action selection Select a new speed adjustment action , so that the driverless car can obtain the next moment behavior , at this time The value is updated, the process is: in, is the attenuation coefficient, Indicates the current behavior The value of Indicates the next moment behavior The value of; the driverless car completes a movement speed adjustment;
[0017] Step d. As the new behavior pair at the current moment, repeat steps b-c, and the driverless car continuously adjusts the movement speed to complete the autonomous and safe movement speed. planning.
[0018] In step 2, the longitudinal motion state equation of the driverless vehicle is: ,in, represents the control input acceleration, Represents the time constant. The motion speed tracking error model is constructed as: in, represents the speed tracking error of the driverless car, Represents the acceleration tracking error of the driverless car. and They represent the actual longitudinal velocity change rate and acceleration change rate of the driverless vehicle respectively; and They represent the derivatives of the speed tracking error and acceleration tracking error of the driverless vehicle respectively; The second derivative of the safe motion speed; Indicates the actual longitudinal motion acceleration of the driverless car; Indicates the actual longitudinal speed of the driverless car.
[0019] In step 3, the longitudinal velocity tracking controller is designed as follows: ,in, and Indicates the adjustable controller parameters of the velocity tracking error and the adjustable controller parameters of the acceleration tracking error.
[0020] In this way, under the action of the longitudinal speed tracking controller, the actual movement speed of the driverless car Tracking planned safe movement speed , and achieve motion speed constraint performance , Indicates the pre-specified safe speed limit value.
[0021] In step 4, the lateral dynamics model of the driverless vehicle is: ,in, , and denote the tracking errors of the lateral path and yaw angle, respectively, represents the front wheel steering angle, Represents the path curvature. , , , and Denote the cornering stiffness of the front and rear wheels respectively, and Represent the distance from the center of mass to the front axle and the rear axle respectively. represents the quality of the driverless car, Indicates that the vehicle is around The moment of inertia of the shaft, express The derivative of and denote the rates of change of lateral path and yaw angle tracking errors, respectively;
[0022] make , , converted into the following form: ,in, represents the control input for the lateral path tracking of the driverless car, and ,Furthermore, in order to enable the driverless car to achieve stable path tracking within a limited time, we define Denotes the auxiliary error variable, and the lateral path tracking error system is established as follows: ,in, , , , ; represents the vector consisting of the lateral path tracking error and the yaw angle tracking error, represents the vector consisting of the lateral path tracking error change rate and the yaw angle tracking error change rate, and Respectively and The derivative of .
[0023] In step 5, the lateral path tracking controller is designed as follows: ,in, express The pseudo-inverse matrix of , , , , and denote adjustable parameters, and . and Respectively and The coefficient matrix of represents the control matrix, The coefficient matrix representing the path curvature; , Respectively represent matrices The diagonal elements of Represents auxiliary variables Adjustable parameters of
[0024] , Respectively represent matrices The diagonal elements of Representation Controller Adjustable parameters.
[0025] In this way, under the action of the lateral path tracking controller, the driverless vehicle can stably track the desired motion path within a limited time.
[0026] Compared with the prior art, the present invention has the following advantages and effects:
[0027] The present invention provides a finite-time stable control method for the autonomous and safe motion speed of an unmanned vehicle. By measuring the distance state information from the vehicle in front, an autonomous and safe motion speed learning method for the unmanned vehicle is designed. Using the reward and punishment value function, the unmanned vehicle continuously adjusts the speed and autonomously plans the safe motion speed; using the longitudinal motion state equation of the unmanned vehicle, a longitudinal speed tracking control method is proposed and the motion speed constraint performance is realized, so that the unmanned vehicle can stably track the planned safe speed; using the lateral dynamics model of the unmanned vehicle and applying the longitudinal speed, a finite-time lateral path tracking controller is constructed, so that the unmanned vehicle can achieve stable tracking of the desired path within a finite time. From a new perspective of planning autonomous and safe motion speed, the present invention proposes a new method of longitudinal speed tracking control with constraint performance and finite-time lateral path tracking control, which ensures the stability and safety of the unmanned vehicle's tracking motion. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1 This is a flow chart of the control method of the present invention.
[0029] Figure 2 This is a working block diagram of the longitudinal velocity tracking controller and the lateral path tracking controller of the present invention.
[0030] Figure 3 This is a circuit schematic diagram of the driverless car main control module of the present invention.
[0031] Figure 4 The figure is a schematic diagram of the communication circuit of the driverless car of the present invention. DETAILED DESCRIPTION
[0032] See also Figure 1 and Figure 2 The present invention proposes a finite time stable control method for the autonomous and safe motion speed of an unmanned vehicle. Figure 2 It can be seen that the driverless car autonomously plans the safe movement speed v d , in the longitudinal velocity tracking controller u L The actual speed of the driverless car under the action v Tracking planned safe speed v d , and achieve motion speed constraint performance; at the same time, the driverless car uses the designed lateral path tracking controller u H , achieving the desired motion path y d and The finite time stable tracking includes the following steps:
[0033] S1: Using the distance status information between the driverless car and the vehicle in front, an autonomous safe motion speed learning method is designed. Through the reward and punishment value function, the driverless car realizes safe motion speed planning;
[0034] Specifically, the distance status information between the driverless car and the vehicle in front is used , , describing the speed adjustment action of the driverless car as follows: ,in, represents the distance between the driverless car and the car in front, Indicates the safe driving distance specified between the driverless car and the vehicle in front. and They represent the speed of the driverless car at the current moment and the next moment respectively. Indicates the change in the speed adjustment of the driverless car.
[0035] make , the reward and penalty value function for adjusting the speed of the driverless car is designed as: ,The learning method of autonomous and safe motion speed of the unmanned vehicle is as follows:
[0036] Step a. Initialize the driverless car behavior pair ,in Indicates the current distance status information. Indicates the current speed adjustment action; sets the learning rate of the driverless car's update state , attenuation coefficient , the selection probability of the speed adjustment action ;
[0037] Step b. The driverless car senses the distance to the vehicle in front and with Compare the size and determine the distance status information between the driverless car and the car in front , the driverless car adjusts the probability of action selection based on speed Select The speed adjustment action in the process determines the behavior of the driverless car at the current moment. ;
[0038] Step c. The driverless car Get the reward and penalty value and enter the next distance status information , and then use the speed to adjust the probability of action selection Select a new speed adjustment action , so that the driverless car can obtain the next moment behavior , at this time The value is updated, the process is: in, is the attenuation coefficient, Indicates the current behavior The value of Indicates the next moment behavior The value of; the driverless car completes a movement speed adjustment;
[0039] Step d. As the new behavior pair at the current moment, repeat steps b-c, and the driverless car continuously adjusts the movement speed to complete the autonomous and safe movement speed. planning.
[0040] S2: Apply the longitudinal motion state equation of the unmanned vehicle to construct a motion speed tracking error model;
[0041] Specifically, the longitudinal motion state equation of the driverless car is: ,in, represents the control input acceleration, Represents the time constant. The motion speed tracking error model is constructed as: ,in, represents the speed tracking error of the driverless car, Represents the acceleration tracking error of the driverless car. and They represent the actual longitudinal velocity change rate and acceleration change rate of the driverless vehicle respectively; and They represent the derivatives of the speed tracking error and acceleration tracking error of the driverless vehicle respectively; The second derivative of the safe motion speed; Indicates the actual longitudinal motion acceleration of the driverless car; Indicates the actual longitudinal speed of the driverless car.
[0042] S3: Using the motion speed tracking error model, design the longitudinal speed tracking controller to achieve the speed tracking and motion speed constraint performance of the driverless vehicle;
[0043] Specifically, the longitudinal velocity tracking controller is designed as follows: ,in, and Indicates the adjustable controller parameters of the velocity tracking error and the adjustable controller parameters of the acceleration tracking error.
[0044] Substituting the longitudinal velocity tracking controller into the motion velocity tracking error model, we can obtain: , in order to illustrate the tracking error and Asymptotically stable, and the actual speed of the driverless car meets the constraints , if there exists a positive definite symmetric matrix and ,matrix and , so that the following matrix inequality holds: , ,in represents the initial motion speed of the driverless car, then the longitudinal speed tracking controller can make the actual motion speed Tracking planned safe speed , while the actual movement speed Satisfy constraint performance , thereby ensuring the safety of the actual movement speed of the driverless car.
[0045] The Lyapunov function is designed as follows: , we can get , multiply the left side of equation (9) by , right multiply , can be obtained ,in , , further from formula (11), we can get , so from formula (14), we can see that the speed tracking error system is asymptotically stable, and the actual speed of the driverless car is Stable tracking of planned safe motion speed .
[0046] Integrate both sides of equation (14) to obtain , and then we have: , multiply the left side of equation (10) by , right multiply , can be obtained , combining equation (15) and equation (16), we can get , it can be seen from formula (17) that under the action of the longitudinal speed tracking controller, the unmanned vehicle achieves the motion speed constraint performance in the process of tracking the planned safe motion speed S4: Calculate the longitudinal velocity of the unmanned vehicle and apply the lateral dynamics model of the unmanned vehicle to construct a lateral path tracking error system;
[0047] Specifically, the lateral dynamics model of the driverless car is: ,in, , denote the tracking errors of the lateral path and yaw angle, respectively, y and y d They represent the actual lateral motion path and expected motion path of the driverless car respectively. and They represent the actual yaw angle path and the expected yaw angle path of the lateral motion of the driverless vehicle, represents the front wheel steering angle, Represents the path curvature. , , , and Denote the cornering stiffness of the front and rear wheels respectively, Represent the distance from the center of mass to the front axle and the rear axle respectively. represents the quality of the driverless car, Indicates that the vehicle is around The moment of inertia of the shaft.
[0048] make , , transform formula (8) into the following form: ,in, represents the control input for the lateral path tracking of the driverless car, and , where the parameter matrix and In represents the actual longitudinal motion speed of the unmanned vehicle, which is calculated by formula (5) under the action of the longitudinal speed tracking controller.
[0049] Furthermore, in order to enable the driverless car to achieve stable path tracking within a limited time, we define Denotes the auxiliary error variable, and the lateral path tracking error system is established as follows: ,in, , , , , , Respectively represent matrices The diagonal elements of Represents auxiliary variables Adjustable parameters.
[0050] S5: Based on the lateral path tracking error system, a finite-time lateral path tracking controller is designed to enable the driverless vehicle to stably track the desired motion path within a finite time;
[0051] Specifically, the lateral path tracking controller is designed as follows: ,in, express The pseudo-inverse matrix of , , , , , Respectively represent matrices The diagonal elements of Representation Controller The adjustable parameters of In order to ensure the finite-time stability of the lateral path tracking error system, the Lyapunov function is designed as follows: , derive formula (22) along the lateral path tracking error system, and substitute the lateral path tracking controller into it, and we can get: According to Young's inequality, there exists a positive number , so that the following inequality holds: ,and ,in, , , . ,in, , , Substituting equations (24)-(26) into equation (23), we can obtain: ,in, , .
[0052] Therefore, it can be seen from formula (27) that under the action of the lateral path tracking controller, the unmanned vehicle achieves stable tracking of the desired motion path within a limited time.
[0053] See also Figure 3 and Figure 4 Finally, MC9S12XEP is used as the control unit of the main control module to run the limited time stable control method of the autonomous and safe motion speed of the unmanned vehicle proposed in the present invention. The control unit is composed of a data acquisition module, a main control module, a motor drive and execution module, and an information feedback module. The main control module input is connected to the data acquisition module of the rotary encoder and the distance sensor, and the output is connected to the motor drive module. The motor drive module is connected to the motor, and the power supply system supplies power to each electrical device. The control method of the main control module is to read the feedback signal of the motor encoder in the information feedback module, and compare it with the command signal given by the main control module to calculate the error signal. According to the error signal, the main control module calculates the control amount of the motor according to the predetermined control algorithm, and sends it to the motor drive module, so that the motor drives the wheel to rotate and moves according to the designed control method.
Claims
1. A finite time stable control method for autonomous and safe motion speed of an unmanned vehicle, characterized in that: The following steps are involved: Step 1: Using the distance status information between the driverless car and the vehicle in front, design an autonomous safety motion speed learning method. Through the reward and punishment value function, the driverless car completes the safety motion speed planning; Step 2: Apply the longitudinal motion state equation of the driverless vehicle and construct a motion speed tracking error model through the planned safe motion speed; Step 3: Based on the motion speed tracking error model, design the longitudinal speed tracking controller to complete the speed tracking and motion speed constraint performance of the driverless vehicle; Step 4: Calculate the longitudinal motion speed based on the unmanned vehicle speed tracking, and apply the lateral dynamics model of the unmanned vehicle to build a lateral path tracking error system; Step 5: Based on the lateral path tracking error system, a finite-time lateral path tracking controller is designed to enable the driverless vehicle to stably track the desired motion path within a finite time; In step 2, the longitudinal motion state equation of the driverless vehicle is: Where u1 represents the control input acceleration, τ d represents the time constant; the motion speed tracking error model is constructed as: Among them, e v =vv d represents the speed tracking error of the driverless car, represents the acceleration tracking error of the driverless car; and a represent the actual longitudinal velocity change rate and acceleration change rate of the driverless vehicle, respectively; and They represent the derivatives of the speed tracking error and acceleration tracking error of the driverless vehicle respectively; represents the second-order derivative of the safe motion speed; a represents the actual longitudinal motion acceleration of the driverless car; v represents the actual longitudinal motion speed of the driverless car; In step 4, the lateral dynamics model of the driverless vehicle is: in, e y and denote the tracking errors of the lateral path and yaw angle, δ f represents the front wheel steering angle, C g Represents the path curvature; a1=C f +G r , a2=-(l f C f -l r C r ), C f Indicates the cornering stiffness of the front wheel, C γ represents the cornering stiffness of the rear wheel, l f and l γ Respectively represent the distance from the center of mass to the front axle and the rear axle; m represents the mass of the driverless car, I z represents the moment of inertia of the vehicle around the z-axis; represents the derivative of x, and denote the rates of change of lateral path and yaw angle tracking errors, respectively; make The lateral dynamics model of the driverless car is transformed into the following form: Among them, u H =δ f represents the control input for the lateral path tracking of the driverless car, and In order to achieve stable path tracking for the driverless car within a limited time, define Denotes the auxiliary error variable, and the lateral path tracking error system is established as follows: Where, ξ(e1)=-η1Sig(e1) α ,η1=diag{η 11 , η 12 },Sig(e1) α =[|e 11 | α sgn(e 11 ),|e 12 | α sgn(e 12 )] T , 0<∝<1; e1 represents the vector formed by the lateral path tracking error and the yaw angle tracking error, e2 represents the vector formed by the lateral path tracking error change rate and the yaw angle tracking error change rate, and denote the derivatives of e1 and e2 respectively; A1 and A2 represent the coefficient matrices of e1 and e2 respectively, B represents the control matrix, and B0 represents the coefficient matrix of path curvature; η 11 , η 12 They represent the diagonal elements of the matrix η1, and β represents the auxiliary variable Adjustable parameters.
2. The finite time stable control method for autonomous safety motion speed of an unmanned vehicle according to claim 1 is characterized in that: In step 1, the distance state information s between the driverless vehicle and the preceding vehicle is used. i , i = 1, 2, 3, describing the speed adjustment action a of the driverless car i As follows: s1:S AF >S0,s2:S AF =S0,s3:S AF <S0(1), a1:v t+1 =v t +Δv d , a2:v t+1 =v t , a3:v t+1 =v t -Δv d (2), where S AF represents the distance between the driverless car and the front car, S0 represents the safe driving distance specified by the driverless car and the front car, v t and v t+1 They represent the speed of the driverless car at the current moment and the next moment, Δv d Indicates the change value of the speed adjustment of the driverless car; Let ΔS = S AF -S0, the reward and penalty value function for adjusting the speed of the driverless car is designed as: f is the reward and penalty value function for adjusting the speed of the unmanned vehicle. The autonomous safety speed learning method is as follows: Step a. Initialize the driverless car behavior pair Where s represents the current distance status information, Indicates the current speed adjustment action; sets the learning rate α∈[0, 1], the decay coefficient γ∈[0, 1], and the selection probability ε∈[0, 1] of the driverless car to update the state; Step b. The driverless car senses the distance S to the vehicle in front AF And compare the size with S0 to determine the distance status information s between the driverless car and the front car i , the driverless car selects a with the speed adjustment action selection probability ε i The speed adjustment action in the process determines the behavior of the driverless car at the current moment. Step c. The driverless car obtains the reward and penalty value according to f, and enters the next distance state information s′, and then uses the selection probability ε of the speed adjustment action to select a new speed adjustment action Enable the driverless car to obtain the next moment behavior At this time The value is updated, the process is: Where γ is the attenuation coefficient, Indicates the current behavior The value of Indicates the next moment behavior The value of; the driverless car completes a movement speed adjustment; Step d. As the new behavior pair at the current moment, repeat steps b-c, and the driverless car continuously adjusts the movement speed to complete the autonomous and safe movement speed v d planning.
3. The finite time stable control method for autonomous safety motion speed of an unmanned vehicle according to claim 1 is characterized in that: In step 3, the longitudinal velocity tracking controller is designed as follows: Among them, k p and k d represents the speed tracking error adjustable controller parameters and the acceleration tracking error adjustable controller parameters; Under the action of the longitudinal speed tracking controller, the actual longitudinal motion speed v of the driverless car tracks the planned autonomous safety motion speed v d , motion speed constraint performance |v|≤υ s , s Indicates the pre-specified safe speed limit value.
4. The finite time stable control method for autonomous safety motion speed of an unmanned vehicle according to claim 1 is characterized in that: In step 5, the finite-time lateral path tracking controller is designed as follows: in, represents the pseudo-inverse matrix of B, η1=diag{η 11 , η 22 }, and Under the action of the lateral path tracking controller, the driverless car stably tracks the desired motion path within a limited time; η 21 , η 22 They represent the diagonal elements of the matrix η2, and λ1 represents the controller u H Adjustable parameters.
Citation Information
Patent Citations
Multi-intelligent agent based unmanned electric car automatic overtaking system and method
CN106671982A
Rule and learning model-based highway leaving method for driverless cars
CN109598934A
