An optimal stability control method and system for an autonomous driving fleet

Through the LQR optimal control method and Lyapunov stability analysis, combined with the Aizermann method linearization, the optimal stability control of the autonomous driving fleet is achieved, and the problems of complex stability analysis and high calculation cost in the existing technology are solved. Improve the effectiveness of control strategies and system stability.

CN118113045BActive Publication Date: 2024-11-12HUAZHONG UNIV OF SCI & TECH
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410236791.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-03-01
Publication Date
2024-11-12
Estimated Expiration
2044-03-01

AI Technical Summary

Technical Problem

In the stability analysis and optimal control, the prior art has problems such as complex spatial domain conversion, high computational cost, and difficulty in achieving fleet stability and optimal control at the same time.

Method used

Using the optimal control method of LQR, linearize the vehicle follow-up model through the Aizerman method, define the system energy functional, and use the principle of minimum value to solve the functional extreme value. The solution to the Riccati equation is taken as the Lyapunov function to achieve the best Optimal stability control.

Benefits of technology

Implementing optimal stable control of the fleet in the time domain reduces computational costs, improves the effectiveness of control strategies and system stability, and is suitable for nonlinear follow-up systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118113045B_ABST
    Figure CN118113045B_ABST
Patent Text Reader

Abstract

The present invention belongs to the technical field of automatic driving under intelligent network environment, and discloses an optimal stability control method and system for an automatic driving fleet. The present invention first uses the Aitelman method to linearize the nonlinear terms in the vehicle following model through the origin within a reasonable interval, and then substitutes it into the original following model and converts it into the form of system state equation, and then uses the Lyapunov second method to construct the Lyapunov function to analyze the stability of the fleet, and obtains the value range of the nonlinear terms of the following model that meets the stability requirements. Secondly, using the LQR control method, the energy functional of the system state vector and the control vector is used as an indicator, and the control vector is determined in the allowable control domain using the minimum principle, and it is proved that the solution of the Riccati equation can be taken as the quadratic form of the Lyapunov function, and the optimal stability control of the fleet is achieved according to the linear feedback solution of the system state.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of autonomous driving in an intelligent network environment, and in particular relates to an optimal stability control method for an autonomous driving fleet. Background Art

[0002] In an intelligent network environment, vehicles can move in coordination with a smaller headway to form an autonomous driving mode based on a convoy. If the vehicle following behavior is unstable, it is easy to cause traffic slowdown, oscillation, and even congestion. The second Lyapunov method does not require the solution of the characteristic equation. It only judges the stability of the system by finding a special Lyapunov function. It is a clever and practical stability analysis method. If the performance indicators of the system are to be optimal in a certain sense, an optimal control method must be used to find a control law. Linear quadratic control (LQR) simply uses the state linear feedback control law to form a closed-loop optimal control system, and the results obtained by the optimal control of the linear system are also applicable to nonlinear systems operating under small signals, which can effectively reduce the computational cost.

[0003] Through the above analysis, the problems and defects of the prior art are as follows:

[0004] (1) In stability analysis, the commonly used method based on Laplace transform has the problem of complex spatial domain conversion and difficulty in solving the characteristic equations of other complex systems;

[0005] (2) In optimal control, both model predictive control (MPC) and deep learning control (DRL) methods have the characteristics of high computing power requirements and high computational costs;

[0006] (3) Existing methods fail to conveniently achieve both fleet stability and optimal control objectives simultaneously in the time domain. Summary of the invention

[0007] In view of the problems existing in the prior art, the present invention provides a technology suitable for an autonomous driving fleet, which can simultaneously achieve LQR optimal control and Lyapunov stability.

[0008] The present invention is implemented as follows: an optimal stability control method for an automatic driving fleet, the method comprising the following steps:

[0009] Step S100, obtaining a convoy system state equation and a system equilibrium state according to a specific vehicle following model;

[0010] Step S200, linearizing the system using the Aitelman method;

[0011] Step S300, define the system energy functional, constrain the control vector, use the minimum principle to solve the functional extreme value, and obtain the optimal control of the system;

[0012] Step S400, taking the solution of the Riccati equation as the quadratic matrix in the Lyapunov function, the system can achieve optimal stable control;

[0013] Step S500: designing a vehicle following simulation program according to the optimal stability control method.

[0014] The optimal stability control method for an autonomous driving fleet as claimed in claim 1, wherein obtaining the state equation in step S100 comprises the following steps:

[0015] Step S110: Select a set of state variables that are independent of each other and can fully represent the vehicle motion, including but not limited to the vehicle speed v n , front and rear vehicle spacing Δx n ;

[0016] Step S120: Convert the vehicle following model into a differential form of the system state equation And let the speed difference between the front and rear vehicles Δv n is zero, and the system equilibrium state expression is obtained

[0017] Further, the system linearization in step S200 includes the following steps:

[0018] Step S210: Use the Aitelman method to perform cross-origin linearization (V(Δx n )=λΔx n ), and obtain the linearization coefficient λ;

[0019] Step S220: Substitute the linearized result into the original system state equation to define a new state equation Coefficient matrix A and control matrix B.

[0020] Further, the step S300 includes the following steps:

[0021] Step S310: Define general system energy functional And make certain constraints on the control vector <u<u max , the functional should include the total transient error of the current state from the equilibrium state and the sum of the consumed control energy, where Q and R are both positive definite symmetric matrices;

[0022] Step S320: Use the minimum principle to solve the functional extreme value and define the Hamiltonian function

[0023] Step S330: Calculate partial derivatives of different variables to obtain the state equation Coupling equation Ru+BT =0, adjoint equation The solution of the coupling equation is the optimal control solution;

[0024] Step S340: To determine the closed-loop control solution, assume that there is a reversible matrix transformation κ=Ps between the Hamiltonian coefficients and the state variables, and derive the closed-loop control solution.

[0025] Step S350: Combine the adjoint equation in step S330 and the closed-loop control solution derived in step S340 to obtain the matrix equation

[0026] Step S360: Substitute the state equation in step S330 and the closed-loop control solution in step S340 into the matrix equation in step S350 to obtain the Riccati equation

[0027] Further, the step S400 includes the following steps:

[0028] Step S410: Use Lyapunov's second method to analyze the stability of the system and select a general quadratic function As the Lyapunov function of the system;

[0029] Step S420: Deriving the Lyapunov function And let the quadratic matrix in the derivative result be the general negative definite quadratic matrix A T P+PA=-E=Q;

[0030] Step S430, substituting the linearization coefficient λ in step S210 into the quadratic matrix obtained by derivation in step S420, to obtain the specific form of the Lyapunov function in step S410;

[0031] Step S440, substituting the specific form of the Lyapunov function in step S430 into the general negative definite quadratic form matrix in step S420 to obtain a new general negative definite quadratic form matrix Q' containing the linearization coefficient λ;

[0032] Step S450: Using the negative definite property of the matrix, solve the inequality Q for the matrix elements of the new general negative definite quadratic form matrix in step S440. 11 >0,Q 11 Q 22 -Q 12 Q 21 >0, the value range of the linearization coefficient λ is obtained;

[0033] Further, the step S500 includes the following steps:

[0034] S510 vehicle following module: calculates the state equation and system equilibrium state of the convoy system according to the given vehicle following model;

[0035] S520 system linearization module: Use the Aitelman method to perform cross-origin linearization on the nonlinear part of the system state equation containing state variables, obtain linearization coefficients, and define new state equations, coefficient matrices, and control matrices;

[0036] S530 Optimal Control Module: defines the system energy functional, constrains the control vector, uses the minimum value principle to solve the functional extreme value, and obtains the optimal control of the system;

[0037] S540 stability analysis module: Use Lyapunov's second method to analyze the stability of the system, select a general quadratic function as the system's Lyapunov function, substitute the linearization coefficient into the quadratic matrix, and obtain the specific form of the Lyapunov function. Use the negative definite property of the matrix to solve the range of values ​​of the linearization coefficient, and prove that the system can achieve optimal stability control according to the optimal control method.

[0038] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0039] First, when the present invention applies LQR control to a nonlinear car-following system, a method of linearization directly in the time domain is adopted to achieve optimal stable control. The control strategy is redesigned under the premise of ensuring the stability of the system, which can avoid the blindness and ineffectiveness of the control strategy. First, the nonlinear terms in the vehicle-following model are linearized through the origin within a reasonable interval using the Aitelman method, and then substituted into the original car-following model and converted into the form of the system state equation. Then, the Lyapunov second method is used to construct the Lyapunov function to analyze the stability of the fleet, and the range of values ​​of the nonlinear terms of the car-following model that meets the stability requirements is obtained. Secondly, the LQR control method is used, the energy functionals of the system's state vector and control vector are used as indicators, and the control vector is determined in the allowable control domain using the minimum principle. It is proved that the solution of the Riccati equation can be taken as the quadratic form of the Lyapunov function, and the optimal stable control of the fleet is achieved based on the linear feedback solution of the system state.

[0040] Second, the Lyapunov second method used in the present invention does not require the solution of the characteristic equation, but only determines the system stability by finding a special Lyapunov function. It is a clever and practical stability analysis method. The linear quadratic control (LQR) simply uses the state linear feedback control law to form a closed-loop optimal control system, and the results obtained by the optimal control of the linear system are also applicable to nonlinear systems running under small signals, which can effectively reduce the calculation cost. The present invention simultaneously achieves fleet stability and optimal control, and has obvious advantages in indicators such as safety, energy consumption, and stability recovery time. It is a simple and optimal stability control method for autonomous driving fleets.

[0041] Third, as auxiliary evidence of the inventiveness of the claims of the present invention, it is also reflected in the following important aspects:

[0042] (1) The expected benefits and commercial value of the technical solution of the present invention after transformation are as follows: it can be used in the design of control software for pure autonomous driving fleets and mixed fleets of manual driving and autonomous driving, and can be used for both vehicle-mounted software and roadside equipment software.

[0043] (2) The technical solution of the present invention fills the technical gap in the industry at home and abroad: Currently, the stability analysis of the fleet is generally performed in the frequency domain and the optimal control is ignored, while the present invention directly performs the stability analysis in the time domain and combines it with the optimal control.

[0044] (3) The technical solution of the present invention solves a technical problem that people have long been eager to solve but have never succeeded in solving: it is very difficult to construct a Lyapunov function for stability. The present invention uses the solution of LQR optimal control to construct a Lyapunov function, thereby solving the above problem.

[0045] Fourth, the significant technical progress of the present invention in the optimal stability control method of the autonomous driving fleet is mainly reflected in the following aspects:

[0046] 1. Improved model accuracy: With the development of sensor technology, car-following models can more accurately describe the dynamic interaction between vehicles. These models are now able to consider more variables, such as vehicle acceleration, deceleration, vehicle size, road conditions, etc., thereby improving the accuracy and reliability of the control method.

[0047] 2. Development of optimization algorithms: In step S300, the process of solving the optimal control of the system using the minimum principle becomes more efficient and accurate. With the continuous development of optimization algorithms, such as the use of advanced numerical methods, heuristic algorithms or machine learning techniques, the optimal solution can be found in a shorter time, and these solutions are often closer to the actual optimal control strategy.

[0048] 3. Intelligent control strategies: By introducing advanced control strategies such as adaptive control, predictive control or reinforcement learning, autonomous driving fleets can respond more intelligently to different traffic environments and road conditions. These control strategies can dynamically adjust control inputs based on real-time traffic information, vehicle status and other relevant factors to achieve optimal stability and safety.

[0049] 4. Improvement of system integration: The optimal stability control method of the autonomous driving fleet needs to be closely integrated with other vehicle systems (such as navigation system, perception system, decision-making system, etc.). With the continuous advancement and intelligence of vehicle systems, the information exchange and collaborative work between these systems have become more efficient and accurate, thereby improving the stability and performance of the entire fleet.

[0050] 5. Advances in simulation and testing technology: In step S500, the design and implementation of the vehicle following simulation program has also made significant technological progress. By using high-performance computers, advanced simulation software and virtual reality technology, more realistic and complex traffic scenarios can be simulated, thereby more accurately evaluating the effectiveness and robustness of the control method. In addition, by conducting rigorous testing and verification on actual roads, it can be ensured that the control method can meet the expected performance and safety requirements in actual applications.

[0051] In summary, significant technological advances have made the optimal stability control method of autonomous driving fleets more accurate, efficient and intelligent, laying a solid foundation for the development of future intelligent transportation and autonomous driving technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of the optimal stability control method for an autonomous driving fleet provided by an embodiment of the present invention.

[0053] Figure 2 It is a structural block diagram of an optimal stability control system for an autonomous driving fleet provided by an embodiment of the present invention.

[0054] Figure 3 This is an optimal stability control structure diagram of an autonomous driving fleet provided by an embodiment of the present invention.

[0055] Figure 4 It is a stability range diagram of optimal stability control of an autonomous driving fleet provided by an embodiment of the present invention.

[0056] Figure 5 It is a diagram of undisturbed vehicle data before and after the optimal stability control is compared, provided by an embodiment of the present invention.

[0057] Figure 6 It is an energy consumption diagram comparing before and after the optimal stability control is implemented, provided in an embodiment of the present invention.

[0058] Figure 7 It is a traffic conflict diagram comparing before and after the optimal stability control is implemented, provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0060] The following are two specific embodiments of the optimal stability control method for the autonomous driving fleet:

[0061] Embodiment 1:

[0062] Step S100:

[0063] Step S110, select state variables, for example, the speed of the vehicle (v1), the speed of the preceding vehicle (v2), the distance between the preceding and following vehicles (d), and the time distance between the preceding and following vehicles (τ) as state variables that are independent of each other and can fully represent the movement of the vehicle.

[0064] Step S120: According to the vehicle following model, such as the Intelligent Driver Model (IDM), convert it into a differential form of the system state equation. Assume that the system state equation is:

[0065] \[\begin{bmatrix}\dot{v_1}\\\dot{d}\\\dot{\tau}\end{bmatrix}=f(v_1,v_2,d,\tau,u)\]

[0066] Where u is the control input, such as acceleration or deceleration. Let the speed difference between the front and rear vehicles be zero (v1 v2 = 0), and the system equilibrium state expression is obtained.

[0067] Step S200: Use the Ayres method to linearize the nonlinear system state equation to obtain a linearized system state equation.

[0068] Step S300: define the system energy functional, consider the constraints of the control input, such as the range of acceleration or deceleration, use the minimum principle to solve the functional extreme value, and obtain the optimal control input of the system.

[0069] Step S400: taking the solution of the Riccati equation as the quadratic matrix in the Lyapunov function, and adjusting the system parameters so that the system can achieve optimal stable control.

[0070] Step S500: Based on the above optimal stability control method, a vehicle following simulation program is designed to simulate the driving conditions of the fleet under different road and traffic conditions to verify the effectiveness of the control method.

[0071] Embodiment 2:

[0072] Step S100:

[0073] Step S110, select the vehicle speed (v), the distance between the front and rear vehicles (d) and the time distance between vehicles (τ) as state variables.

[0074] Step S120: Use the full velocity difference model as the vehicle following model and convert it into the differential form of the system state equation. Suppose the system state equation is:

[0075] \[\begin{bmatrix}\dot{v}\\\dot{d}\\\dot{\tau}\end{bmatrix}=f(v,v_{\text{lead}},d,\tau,u)\]

[0076] Where v_{\text{lead}} is the speed of the leading vehicle. Let the speed difference term (v v_{\text{lead}}) be zero, and the equilibrium state expression of the system is obtained.

[0077] Step S200: Linearize the system using the Aitelman method as well.

[0078] Step S300: define the system energy functional, consider the constraints of the control input, use the minimum principle to solve the functional extreme value, and obtain the optimal control input.

[0079] Step S400: constructing a Lyapunov function through the solution of the Riccati equation, adjusting system parameters, and achieving optimal stable control of the system.

[0080] Step S500: Based on the obtained optimal stability control method, a vehicle following simulation program is designed to simulate the driving conditions of a convoy under different road and traffic conditions to verify the effectiveness and robustness of the control method.

[0081] The above two embodiments respectively demonstrate the specific application process of the optimal stability control method for an autonomous driving fleet based on an intelligent driver model and a full speed difference model. In actual applications, a suitable vehicle following model and control method can be selected according to specific needs and conditions.

[0082] like Figure 1 As shown, the optimal stability control method for an automatic driving vehicle fleet provided by the present invention comprises the following steps:

[0083] Step S100, obtaining a convoy system state equation and a system equilibrium state according to a specific vehicle following model;

[0084] Step S200, linearizing the system using the Aitelman method;

[0085] Step S300, define the system energy functional, constrain the control vector, use the minimum principle to solve the functional extreme value, and obtain the optimal control of the system;

[0086] Step S400, taking the solution of the Riccati equation as the quadratic matrix in the Lyapunov function, the system can achieve optimal stable control;

[0087] Step S500: designing a vehicle following simulation program according to the optimal stability control method.

[0088] The step S100 provided by the present invention obtains the state equation including the following steps:

[0089] Step S110: Select a set of state variables that are independent of each other and can fully represent the vehicle motion, including but not limited to the vehicle speed v n , front and rear vehicle spacing Δx n ;

[0090] Step S120: Convert the vehicle following model into a differential form of the system state equation And let the speed difference between the front and rear vehicles Δv n is zero, and the system equilibrium state expression is obtained

[0091] The optimal stability control method for an autonomous driving fleet provided by the present invention aims to achieve the optimal driving state of each vehicle in the fleet through mathematics and control theory to ensure the overall stability and safety of the fleet. The following is the detailed working principle of the method:

[0092] Step S100: Obtaining the fleet system state equation and system equilibrium state

[0093] 1. Select state variables (step S110):

[0094] Select state variables that can fully describe the vehicle's motion characteristics, such as the vehicle's speed, the distance between the front and rear vehicles, etc.

[0095] These variables are the basis for system analysis and are used to build a dynamic model of the fleet.

[0096] 2. Establish the state equation (step S120):

[0097] Based on the car-following model, the dynamic behavior of the vehicle is converted into the differential form of the system state equations. These equations describe the change of state variables over time.

[0098] By setting the speed difference of the front and rear vehicles to zero, we can get the equilibrium state expression of the system. This represents the state of the convoy under ideal conditions (no acceleration, constant speed).

[0099] Step S200: System linearization

[0100] Ayterman Method:

[0101] It is used to linearize the nonlinear system state equation and simplify the design of control strategy. This is because the analysis and control of linear system are much simpler than nonlinear system.

[0102] Linearization is performed near the equilibrium point of the system, approximating the nonlinear equations into linear equations to facilitate the application of linear control theory.

[0103] Step S300: Define the system energy functional and solve the optimal control

[0104] System energy functional:

[0105] A performance metric is defined, usually related to the energy, deviation or error of the system, and is used to evaluate the effectiveness of the control strategy.

[0106] Constraint control vector:

[0107] The control vector is an external input that affects the system state, such as acceleration command. By imposing constraints on it, the practicality and safety of the control strategy can be ensured.

[0108] Minimum principle:

[0109] It is used to solve the minimum value of the energy functional and thus determine the optimal control strategy. This means finding a control method that minimizes the performance index and thus achieves optimal performance.

[0110] Step S400: Using the Riccati equation to achieve optimal stability control

[0111] Solution of Riccati equation:

[0112] In optimal control theory, the solution of the Riccati equation provides a method to determine the control law that minimizes a given performance indicator.

[0113] The solution of the Riccati equation is used as the quadratic matrix in the Lyapunov function to prove the stability of the system and achieve optimal stable control.

[0114] Step S500: Designing a vehicle following simulation program

[0115] Simulation program:

[0116] Based on the above theories and methods, a simulation program is designed to simulate the car-following behavior of the fleet and verify the effectiveness and stability of the proposed control strategy.

[0117] Through simulation, the performance of the control method can be tested under different driving scenarios, including vehicle speed changes, vehicle distance adjustment, etc., to ensure reliability and safety in actual applications.

[0118] In general, the present invention provides an efficient and reliable optimal stability control method for an autonomous driving fleet through system modeling, linearization processing, application of optimal control theory, and stability verification achieved through the Riccati equation, thereby improving the safety and efficiency of the fleet.

[0119] The system linearization in step S200 provided by the present invention includes the following steps:

[0120] Step S210: Use the Aitelman method to perform cross-origin linearization (V(Δx n )=λΔx n ), and obtain the linearization coefficient λ;

[0121] Step S220: Substitute the linearized result into the original system state equation to define a new state equation Coefficient matrix A and control matrix B.

[0122] The step S300 provided by the present invention includes the following steps:

[0123] Step S310: Define the general system energy functional And make certain constraints on the control vector <u<u max , the functional should include the total transient error of the current state from the equilibrium state and the sum of the consumed control energy, where Q and R are both positive definite symmetric matrices;

[0124] Step S320: Use the minimum principle to solve the functional extreme value and define the Hamiltonian function

[0125] Step S330: Calculate partial derivatives of different variables to obtain the state equation Coupling equation Ru+B T =0, adjoint equation The solution of the coupling equation is the optimal control solution;

[0126] Step S340: To determine the closed-loop control solution, assume that there is a reversible matrix transformation κ=Ps between the Hamiltonian coefficients and the state variables, and derive the closed-loop control solution.

[0127] Step S350: Combine the adjoint equation in step S330 and the closed-loop control solution derived in step S340 to obtain the matrix equation

[0128] Step S360: Substitute the state equation in step S330 and the closed-loop control solution in step S340 into the matrix equation in step S350 to obtain the Riccati equation

[0129] The step S400 provided by the present invention includes the following steps:

[0130] Step S410: Use Lyapunov's second method to analyze the stability of the system and select a general quadratic function As the Lyapunov function of the system;

[0131] Step S420: Deriving the Lyapunov function And let the quadratic matrix in the derivative result be the general negative definite quadratic matrix A T P+PA=-E=Q;

[0132] Step S430, substituting the linearization coefficient λ in step S210 into the quadratic matrix obtained by derivation in step S420, to obtain the specific form of the Lyapunov function in step S410;

[0133] Step S440, substituting the specific form of the Lyapunov function in step S430 into the general negative definite quadratic form matrix in step S420 to obtain a new general negative definite quadratic form matrix Q' containing the linearization coefficient λ;

[0134] Step S450: Using the negative definite property of the matrix, solve the inequality Q for the matrix elements of the new general negative definite quadratic form matrix in step S440. 11 >0,Q 11 Q 22 -Q 12 Q 21 >0, the value range of the linearization coefficient λ is obtained;

[0135] The step S500 provided by the present invention includes the following steps:

[0136] S510 vehicle following module: calculates the state equation and system equilibrium state of the convoy system according to the given vehicle following model;

[0137] S520 system linearization module: Use the Aitelman method to perform cross-origin linearization on the nonlinear part of the system state equation containing state variables, obtain linearization coefficients, and define new state equations, coefficient matrices, and control matrices;

[0138] S530 Optimal Control Module: defines the system energy functional, constrains the control vector, uses the minimum value principle to solve the functional extreme value, and obtains the optimal control of the system;

[0139] S540 stability analysis module: Use Lyapunov's second method to analyze the stability of the system, select a general quadratic function as the system's Lyapunov function, substitute the linearization coefficient into the quadratic matrix, and obtain the specific form of the Lyapunov function. Use the negative definite property of the matrix to solve the range of values ​​of the linearization coefficient, and prove that the system can achieve optimal stability control according to the optimal control method.

[0140] like Figure 2 As shown, an optimal stability control system for an autonomous driving fleet includes: a fleet state acquisition module, a linearization module, a system optimal control module, a stability control module, and a simulation module;

[0141] The convoy state acquisition module is used to obtain the convoy system state equation and system equilibrium state according to a specific vehicle following model;

[0142] Linearization module, used to linearize the system using the Aitelman method;

[0143] The system optimal control module is used to define the system energy functional, constrain the control vector, and use the minimum principle to solve the functional extreme value to obtain the system optimal control;

[0144] Stable control module, used to take the solution of Riccati equation as the quadratic matrix in Lyapunov function, so that the system can achieve optimal stable control;

[0145] The simulation module is used to design a vehicle following simulation program based on the optimal stability control method.

[0146] like Figure 3-7 As shown, an optimal stability control method for an autonomous driving fleet is specifically as follows:

[0147] The full velocity difference following model (FVD) is adopted; the simulation scenario is set, under periodic boundary conditions, the circumference of the circular road is L = 1500m, the number of vehicles is N = 100, and the vehicle numbers are arranged clockwise from small to large; in the initial stable state, the vehicle spacing is L h =15m (including vehicle length Lc=5m), vehicle speed v0=5.6m / s, initial disturbance is x 100 (0) = -10m, x n(0)=1500(n-1) / N, (n=2,3...N).

[0148] The consistency rate between the theoretical stable region and the simulation results in the optimal stability control is 68%. In addition, when appropriate parameter values ​​are selected, the platoon has advantages in terms of energy consumption, traffic conflicts, and stability recovery time.

[0149] The optimized speed function following model (OV) is adopted; the simulation scenario is set, under periodic boundary conditions, the circumference of the circular road is L = 1500m, the number of vehicles is N = 100, and the vehicle numbers are arranged from small to large in a clockwise direction; in the initial stable state, the vehicle spacing is L h =15m (including vehicle length Lc=5m), vehicle speed v0=5.6m / s, initial disturbance is x 100 (0) = -10m, x n (0)=1500(n-1) / N, (n=2,3...N).

[0150] When the present invention applies LQR control to a nonlinear car-following system, a method of linearization directly in the time domain is adopted to achieve optimal stable control. The control strategy is redesigned under the premise of ensuring the stability of the system, which can avoid the blindness and ineffectiveness of the control strategy. First, the nonlinear terms in the vehicle-following model are linearized through the origin within a reasonable interval using the Aitelman method, and then substituted into the original car-following model and converted into the form of the system state equation. Then, the Lyapunov second method is used to construct the Lyapunov function to analyze the stability of the fleet, and the range of values ​​of the nonlinear terms of the car-following model that meets the stability requirements is obtained. Secondly, the LQR control method is used, the energy functionals of the system state vector and the control vector are used as indicators, and the control vector is determined in the allowable control domain using the minimum principle. It is proved that the solution of the Riccati equation can be taken as the quadratic form of the Lyapunov function, and the optimal stable control of the fleet is achieved according to the linear feedback solution of the system state.

[0151] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. An optimal stability control method for an autonomous driving fleet, characterized in that: The following steps are involved: Step S100, obtaining a convoy system state equation and a system equilibrium state according to a specific vehicle following model; Step S200, linearizing the system using the Aitelman method; Step S300, define the system energy functional, constrain the control vector, use the minimum principle to solve the functional extreme value, and obtain the optimal control of the system; Step S400, taking the solution of the Riccati equation as the quadratic matrix in the Lyapunov function, the system can achieve optimal stable control; Step S500, designing a vehicle following simulation program according to an optimal stability control method; The system linearization in step S200 includes the following steps: Step S210, using the Aitelman method to perform origin linearization on the nonlinear part containing state variables in the system state equation within a reasonable interval to obtain a linearization coefficient λ; Step S220: Substitute the linearized result into the original system state equation to define a new state equation Coefficient matrix A and control matrix B; The step S300 includes the following steps: Step S310: Define the general system energy functional And make certain constraints on the control vector <u<u max , the functional should include the total transient error of the current state from the equilibrium state and the sum of the consumed control energy, where Q and R are both positive definite symmetric matrices; Step S320: Use the minimum principle to solve the functional extreme value and define the Hamiltonian function Step S330: Calculate partial derivatives of different variables to obtain the state equation Coupling equation Ru+B T =0, The solution of the coupling equation is the optimal control solution; Step S340: To determine the closed-loop control solution, assume that there is a reversible matrix transformation κ=Ps between the Hamiltonian coefficients and the state variables, and derive the closed-loop control solution. Step S350: Combine the adjoint equation in step S330 and the closed-loop control solution derived in step S340 to obtain the matrix equation Step S360: Substitute the state equation in step S330 and the closed-loop control solution in step S340 into the matrix equation in step S350 to obtain the Riccati equation 2. The optimal stability control method for an autonomous driving vehicle fleet as claimed in claim 1, characterized in that: The state equation obtained in step S100 comprises the following steps: Step S110: Select a set of state variables that are independent of each other and can fully represent the vehicle motion, including but not limited to the vehicle speed v n , front and rear vehicle spacing Δx n ; Step S120: Convert the vehicle following model into a differential form of the system state equation And let the speed difference between the front and rear vehicles Δv n is zero, and the system equilibrium state expression is obtained 3. The optimal stability control method for an autonomous driving vehicle fleet according to claim 1, characterized in that: The step S400 includes the following steps: Step S410: Use Lyapunov's second method to analyze the stability of the system and select a general quadratic function As the Lyapunov function of the system; Step S420: Deriving the Lyapunov function And let the quadratic matrix in the derivative result be the general negative definite quadratic matrix A T P+PA=-E=Q; Step S430, substituting the linearization coefficient λ in step S210 into the quadratic matrix obtained by derivation in step S420, to obtain the specific form of the Lyapunov function in step S410; Step S440, substituting the specific form of the Lyapunov function in step S430 into the general negative definite quadratic form matrix in step S420 to obtain a new general negative definite quadratic form matrix Q' containing the linearization coefficient λ; Step S450: Using the negative definite property of the matrix, solve the inequality Q for the matrix elements of the new general negative definite quadratic form matrix in step S440. 11 >0,Q 11 Q 22 -Q 12 Q 21 >0, and obtain the value range of the linearization coefficient λ.

4. The optimal stability control method for an autonomous driving vehicle fleet as claimed in claim 1, characterized in that: The step S500 includes the following steps: S510 vehicle following module: calculates the state equation and system equilibrium state of the convoy system according to the given vehicle following model; S520 system linearization module: Use the Aitelman method to perform cross-origin linearization on the nonlinear part of the system state equation containing state variables, obtain linearization coefficients, and define new state equations, coefficient matrices, and control matrices; S530 Optimal Control Module: defines the system energy functional, constrains the control vector, uses the minimum value principle to solve the functional extreme value, and obtains the optimal control of the system; S540 stability analysis module: Use Lyapunov's second method to analyze the stability of the system, select a general quadratic function as the system's Lyapunov function, substitute the linearization coefficient into the quadratic matrix, and obtain the specific form of the Lyapunov function. Use the negative definite property of the matrix to solve the range of values ​​of the linearization coefficient, and prove that the system can achieve optimal stability control according to the optimal control method.

5. An optimal stability control system for an autonomous driving vehicle fleet according to the optimal stability control method for an autonomous driving vehicle fleet as claimed in any one of claims 1 to 4, characterized in that: The optimal stability control system of the autonomous driving fleet includes: a fleet state acquisition module, a linearization module, a system optimal control module, a stability control module, and a simulation module; The convoy state acquisition module is used to obtain the convoy system state equation and system equilibrium state according to a specific vehicle following model; Linearization module, used to linearize the system using the Aitelman method; The system optimal control module is used to define the system energy functional, constrain the control vector, and use the minimum principle to solve the functional extreme value to obtain the system optimal control; Stable control module, used to take the solution of Riccati equation as the quadratic matrix in Lyapunov function, so that the system can achieve optimal stable control; The simulation module is used to design a vehicle following simulation program based on the optimal stability control method.

Citation Information

Patent Citations

  • Vehicle queue stability control method considering communication delay

    CN109591804A

  • Vehicle queue following stability control method considering communication topological time varying

    CN109656255A

  • Virtual marshalling train formation control method and system, electronic equipment and medium

    CN115718425A

  • Data-driven hybrid vehicle queue robust control method

    CN115857494A