Transverse and longitudinal coupling motorcade cooperative control method based on Frenet coordinate system

Abstract

The invention discloses a transverse and longitudinal coupled motorcade cooperative control method based on a Frenet coordinate system. According to the method, the initial state is given for each vehicle in the vehicle team, and the states of other vehicles are transmitted and received according to the communication topology, so that the vehicles in the whole vehicle team follow the target state, and the consistency of the vehicles in the vehicle team and the safety of the vehicle team under the non-constant straight road condition are ensured. Firstly, a three-degree-of-freedom dynamical model of a vehicle is introduced, then a Frenet coordinate system is introduced, and the two models are combined to obtain a control model of the single vehicles in a motorcade. Through the DMPC algorithm, each vehicle in the vehicle team is optimized and solved at each moment, so that the vehicle team keeps good vehicle team cooperation performance under the non-constant straight road condition, a good result is obtained in simulation, and the effectiveness of the transverse and longitudinal coupling distributed model prediction control method under the Frenet coordinate system is verified.

Description

Technical Field

[0001] The present invention relates to a vehicle fleet collaborative control method in the field of intelligent connected vehicle collaborative control, and in particular to a vehicle fleet collaborative control method and system based on a Frenet coordinate system based on a lateral and longitudinal coupled distributed model predictive control. Background Art

[0002] For autonomous vehicles, the coordinated control of intelligent networked vehicles has important social and economic value in terms of energy conservation, emission reduction and energy consumption reduction. The control of the fleet of intelligent networked vehicles has important research significance in improving traffic capacity, reducing traffic accidents and reducing driver fatigue. Previous fleet research has mostly focused on one-dimensional control, namely longitudinal control. In recent years, two-dimensional control, namely longitudinal and lateral control of the fleet, has become a hot topic, because the integrated control of the horizontal and longitudinal directions can allow the fleet to cope with more complex road conditions. Many current decoupled controls are based on the integrated horizontal and longitudinal control of a single vehicle. Such control methods simplify the design and implementation of the control system, facilitate development and debugging, and also reduce the computational burden. However, when the perspective of the problem shifts to fleet control, the decoupled control of the horizontal and longitudinal directions exposes a series of problems such as poor coordination and reduced overall performance of the fleet. The horizontal and longitudinal coupling control of the fleet can greatly improve the overall coordination of the fleet and the safety of the fleet form, and has better handling capabilities under complex road conditions.

[0003] The Frenet coordinate system is widely used in the field of autonomous driving, especially in the path planning system of autonomous driving in urban and highway traffic environments. The Frenet coordinate system uses the center line of the road as the base frame and uses the tangent vector and normal vector of the reference line to establish the coordinate system.

[0004] At present, for the control of the fleet, a large number of control methods have been verified by scholars for their feasibility, such as PID control, sliding membrane control (SMC), model predictive control (MPC), etc.; among them, MPC is one of the mainstream control methods currently used in fleet control. Compared with PID and SMC, MPC has obvious advantages in dealing with multi-variable coupling problems, while PID and SMC will greatly increase the design complexity of the controller when facing the problem of multi-input and output systems. At the same time, when dealing with constraint problems, MPC can explicitly deal with the constraints of system input, state and output. In each control cycle, the control action is ensured to be within the constraint range by solving the optimization problem. This is very important in actual fleet control, because the actual fleet system usually has physical or safety constraints, such as wheel angle.

[0005] At present, there are two main types of fleet control solutions:

[0006] 1. Centralized control scheme: The centralized control scheme relies on a central controller, which collects the status information of all vehicles in the entire fleet (such as position, speed, acceleration, etc.), performs global optimization calculations, and then sends control instructions to each vehicle to ensure the coordination and optimization of the overall behavior of the fleet. Since all information and calculations are concentrated in the central controller, the global optimal control strategy can be implemented to ensure the overall coordination and optimization effect of the fleet, and the central controller can make unified decisions to avoid suboptimal decisions due to incomplete local information.

[0007] 2. Distributed control scheme: The distributed control method distributes the control tasks to multiple subsystems in each vehicle or fleet. Each local controller performs local optimization independently and coordinates through vehicle-to-vehicle (V2V) communication. Each vehicle makes local control decisions based on its own status and information from other vehicles to achieve overall optimization in a collaborative manner. Compared with centralized control schemes, distributed control schemes can effectively reduce the computational burden. And with the development of V2V communication technology, the reliability of information transmission between vehicles has also been greatly improved, which has further promoted the development of distributed control schemes. Distributed model predictive control (DMPC), which combines model predictive control and distributed control schemes, has good results in solving constraint problems, reducing computational burdens, and handling multi-objective optimization problems, and has strong reliability, so it can achieve consistency and stability between fleets.

[0008] In summary, studying a convoy cooperative control method based on the Frenet coordinate system's horizontal and vertical coupled distributed model predictive control is not only of theoretical significance, but also of engineering significance. Summary of the invention

[0009] The purpose of the present invention is to address the shortcomings of the current fleet collaborative control method of distributed model predictive control with horizontal and vertical coupling, and to propose a fleet collaborative control method based on distributed model predictive control with horizontal and vertical coupling in Frenet coordinate system. The method integrates distributed model predictive control, a three-degree-of-freedom vehicle dynamics model and the coordinates of the vehicle in the Frenet coordinate system, takes into account the real vehicle dynamics constraints, controls the fleet in the horizontal and vertical directions, ensures the consistency and safety of the fleet, and verifies the feasibility and safety of the method through simulation.

[0010] The objective of the present invention is achieved through the following technical solutions:

[0011] 1. A platoon cooperative control method based on the lateral and longitudinal coupling of the Frenet coordinate system

[0012] S1: Combined with the Frenet coordinate system, the lateral and longitudinal coupling control model of the fleet is constructed using the vehicle three-degree-of-freedom bicycle dynamics model;

[0013] S2: After discretizing the lateral and longitudinal coupling control model of the convoy using the Euler forward difference method, a discrete nonlinear model of each following vehicle is obtained;

[0014] S3: Build the communication topology of the fleet;

[0015] S4: According to the communication topology of the convoy and the discrete nonlinear model of the follower vehicle, the distributed model predictive control method is used to optimize the state of each follower vehicle in the convoy to achieve convoy control.

[0016] In S1, the lateral and longitudinal coupling control model of the fleet satisfies the following formula:

[0017]

[0018] in, and are the longitudinal speeds of the following vehicle i Vehicle lateral speed The differential of is the longitudinal acceleration of the following vehicle i in the vehicle body coordinate system, m i represents the mass of the following car i, i∈1,2,…,N, N represents the number of following cars in the convoy; represents the yaw rate of the following vehicle i, is the arc length s of the following vehicle i i The time derivative, is the distance d between the center of mass of the following vehicle i and the projection point of the center of mass of the following vehicle i on the reference line i The differential of and denote the cornering stiffness of the front and rear wheels of the following vehicle i, δ i is the front wheel turning angle of the following vehicle i, γ i is the yaw rate of the following vehicle i, a i and b i are the distances between the front wheel and rear wheel turning wheelbase and the center of mass of vehicle i, is the moment of inertia of following vehicle i, represents the yaw angle of the following vehicle i, is the angle between the tangent line of the projection point of the following vehicle i on the reference line and the X-axis, k i is the road curvature corresponding to the projection point of the center of mass of the following vehicle i on the reference line.

[0019] The discrete nonlinear model of each following vehicle satisfies the following formula:

[0020]

[0021] Among them, X i (t+1) and X i(t) are the state sequences at time t+1 and time t, respectively, U i (t) is the control input at time t, y i (t) is the control output at time t, φ i (·) is the lateral and longitudinal coupling control model of the following vehicle i, T s is the sampling time, is the yaw angle of following vehicle i at time t.

[0022] In S3, the communication topology of the convoy is a directed graph including a directed spanning tree, wherein the leading vehicle of the convoy is taken as a root node, and the following vehicles are taken as other nodes in the directed spanning tree, and the root node is connected to the other nodes.

[0023] The S4 is specifically:

[0024] S4.1: Construct the optimization problem for each following vehicle, the formula is as follows:

[0025]

[0026] in, is the predicted input sequence of the following car i, J i (·) is the cost function, is the predicted state sequence of the following vehicle i, is the predicted output sequence of the following vehicle i, is the hypothetical output sequence of the following car i, is the hypothetical output sequence of vehicle j, is the predicted position sequence of the following vehicle i, s0(:|t) is the position sequence of the leading vehicle, is the state at time t+c predicted at time t, represents the unknown input variable to be optimized, X i (t) is the state of the following vehicle i at time t, φ i (·) is the nonlinear expression of the single-vehicle lateral and longitudinal coupling model, and are the lower and upper bounds of the input acceleration constraint, δ i,min and δ i,max are the lower and upper bounds of the input steering angle, is the input acceleration, is the input steering angle, |·| is the absolute value, N p is the prediction time domain, X i,des (c|t) is the ideal state sequence of the following vehicle i, Q i is the first positive definite weight matrix, s0(c|t) is the position sequence of the lead vehicle, is the predicted position sequence of the following vehicle i, G i is the second positive definite weight matrix, is the predicted output sequence of the following vehicle i, is the hypothetical output sequence of the following vehicle i, F i is the third positive definite weight matrix, is the adjacency matrix of the following car i, M i is the fourth positive definite weight matrix, R i is the fifth positive definite weight matrix; diag() is the matrix diagonalization operation;

[0027] S4.2: According to the communication topology of the convoy and the discrete nonlinear model of the follower vehicle, the distributed model predictive control method is used to optimize and solve the optimization problem of each follower vehicle, so as to control each follower vehicle, and the optimization problems of all follower vehicles in the convoy are traversed and optimized to control all follower vehicles, thereby realizing the control of the convoy.

[0028] In S4.2, the distributed model predictive control method is used to optimize and solve the optimization problem of each following vehicle, so as to control each following vehicle, specifically:

[0029] Step 1: At t = 0, the parameters are initialized:

[0030] Specifically, at time t=0, the assumed control input sequence of each following vehicle i is initialized and the assumed state sequence X i (t) is the preset initial state of each vehicle; assuming the output sequence φ i () is the lateral and longitudinal coupling control model of the following vehicle i;

[0031] Step 2: At time t>0, after optimizing and solving the optimization problem described in S4.1 for each following vehicle i, the optimal control input sequence is obtained and its own hypothetical output sequence is transmitted to the adjacent following vehicle j according to the communication topology of the convoy, and the hypothetical output sequence of the preceding vehicle and the leading vehicle is accepted;

[0032] Step 3: Calculate the hypothetical control input sequence, hypothetical state sequence and hypothetical output sequence at the next moment based on the current optimal control input sequence. The calculation formula is as follows:

[0033]

[0034] in, is the hypothetical control input sequence of the ith following vehicle, is the hypothetical output sequence of the i-th following vehicle, and are the hypothetical state sequences at time t+c and time t+c+1 predicted at time t, respectively, i (·) is the lateral and longitudinal coupling control model of the i-th following vehicle, is the optimal control input sequence predicted by the following vehicle i at time t-1 at time t+c, is the t+Nth prediction of the following vehicle i at time t-1 p -2 The optimal control input sequence at time, and are the hypothetical state sequence of the following vehicle i and the optimal state sequence of the following vehicle i respectively;

[0035] Step 4: Apply the first value of the current optimal control input sequence of each following vehicle to the following vehicle i as the current real control quantity of the following vehicle i to control the acceleration and front wheel steering angle of the following vehicle i;

[0036] Step 5: At the next moment, based on the current assumed control input sequence, assumed state sequence and assumed output sequence of other vehicles, repeat Step 2-Step 4 and start rolling optimization solution again, so as to continuously control the following vehicle i.

[0037] 2. A convoy cooperative control system based on the Frenet coordinate system with horizontal and vertical coupling

[0038] The fleet control model building module is used to build the lateral and longitudinal coupling control model of the fleet by combining the Frenet coordinate system and using the vehicle three-degree-of-freedom bicycle dynamics model;

[0039] The following vehicle discrete nonlinear model building module is used to discretize the lateral and longitudinal coupling control model of the convoy using the Euler forward difference method to obtain the discrete nonlinear model of each following vehicle;

[0040] A vehicle fleet communication topology structure generation module, used to generate a vehicle fleet communication topology structure;

[0041] The fleet controller is used to optimize the state of each follower vehicle in the fleet based on the communication topology of the fleet and the discrete nonlinear model of the follower vehicle by using the distributed model predictive control method, and control all follower vehicles in real time.

[0042] 3. A computer device

[0043] The device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the convoy cooperative control method based on the lateral and longitudinal coupling of the Frenet coordinate system when executing the computer program.

[0044] 4. A computer-readable storage medium

[0045] The medium stores a computer program, and when the computer program is executed by a processor, the steps of the convoy cooperative control method based on the lateral and longitudinal coupling of the Frenet coordinate system are implemented.

[0046] V. A computer program product

[0047] The product includes a computer program / instruction, which, when executed by a processor, implements the steps of the convoy cooperative control method based on the lateral and longitudinal coupling of the Frenet coordinate system.

[0048] The beneficial effects of the present invention are as follows:

[0049] 1. A three-degree-of-freedom vehicle dynamics model is introduced for vehicle control. Compared with the high-degree-of-freedom model, it greatly reduces the computational burden while ensuring a certain degree of accuracy of the model.

[0050] 2. The Frenet coordinate system of the vehicle is introduced and combined with the three-degree-of-freedom vehicle dynamics model to form a convoy control model, ensuring that the convoy can maintain good consistency of vehicles in the convoy and the safety of the convoy even on non-constant straight road conditions.

[0051] 3. Compared with traditional PID control and SMC, the DMPC control algorithm is adopted, which has a simpler design, can effectively constrain the state of the vehicle, and has a faster solution speed than the centralized MPC control algorithm.

[0052] 4. In view of the changes in road curvature in reality, the above-mentioned lateral and longitudinal coupled convoy control method was tested and good experimental results were obtained, which proved the effectiveness of the above-mentioned convoy control method. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A schematic diagram of a fleet provided by an embodiment of the present invention.

[0054] Figure 2 A schematic diagram of a three-degree-of-freedom vehicle dynamics model provided in an embodiment of the present invention.

[0055] Figure 3 A schematic diagram of the coordinates of a vehicle in the Frenet coordinate system provided by an embodiment of the present invention.

[0056] Figure 4 A schematic diagram of the PLF vehicle communication topology structure provided in an embodiment of the present invention.

[0057] Figure 5 A schematic diagram of a reference trajectory provided by an embodiment of the present invention.

[0058] Figure 6This is an experimental simulation effect diagram of a fleet of vehicles using a horizontally and vertically coupled distributed model predictive control traveling on a reference trajectory provided by an embodiment of the present invention.

[0059] Figure 7 A carsim simulation scene diagram provided by an embodiment of the present invention.

[0060] Figure 8 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0061] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0062] like Figure 8 and Figure 1 As shown, the embodiment provides a convoy cooperative control method based on the Frenet coordinate system for lateral and longitudinal coupling, including the following steps:

[0063] S1: Combined with the Frenet coordinate system, the lateral and longitudinal coupling control model of the fleet is constructed using the vehicle three-degree-of-freedom bicycle dynamics model;

[0064] S1 is specifically:

[0065] S1.1: If Figure 2 As shown, the three-degree-of-freedom bicycle dynamics model of the vehicle is given, and the formula is as follows:

[0066]

[0067] Among them, m i represents the mass of the following vehicle i, i∈1,2,…,N, is the moment of inertia of following vehicle i, a i and b i are the distances between the front wheel and rear wheel turning wheelbase and the center of mass of vehicle i, and They represent the longitudinal forces acting on the front and rear tires of the i-th vehicle respectively; and They represent the lateral forces received by the front and rear tires of the i-th following vehicle respectively; the three degrees of freedom of the vehicle are the longitudinal velocity of the vehicle Vehicle lateral speed and yaw rate and The vehicle longitudinal speed is Vehicle lateral speed The differential of .

[0068] Assuming that the slip angles of the front and rear tires are within a small range and the tire model is linear, the lateral forces of the front and rear tires are:

[0069]

[0070] in, and denote the cornering stiffness of the front and rear wheels of the following vehicle i, δ i is the front wheel steering angle of the following vehicle i.

[0071] Define the state x of the following car i i and control input u i ,satisfy T stands for transpose, is the longitudinal acceleration of the following vehicle i in the vehicle body coordinate system. Rewrite the three-degree-of-freedom bicycle dynamics model of the vehicle:

[0072]

[0073] in, represents the yaw rate of the following vehicle i The differential of .

[0074] S1.2: If Figure 3 As shown, the coordinates of the vehicle in the Frenet coordinate system are given:

[0075]

[0076] Among them, s i is the arc length of the following vehicle i, that is, the distance that vehicle i travels on the reference line from the starting point, For i The derivative with respect to time. is the distance d between the center of mass of the following vehicle i and the projection point of the center of mass of the following vehicle i on the reference line i The differential of . represents the yaw angle of the following vehicle i. i is the road curvature corresponding to the projection point of the center of mass of the following vehicle i on the reference line, is the angle between the tangent line of the projection point of the following vehicle i on the reference line and the X-axis. Since most actual roads have small curvatures, 1-k i d i ≠0.

[0077] S1.3: The lateral and longitudinal coupling control model can be obtained by combining the three-degree-of-freedom bicycle dynamics model of the above vehicle and the coordinates of the vehicle in the Frenet coordinate system. The lateral and longitudinal coupling control model of the fleet satisfies the following formula:

[0078]

[0079] in, and are the longitudinal speeds of the following vehicle i The lateral speed of the following vehicle i The differential of is the longitudinal acceleration of the following vehicle i in the vehicle body coordinate system, m i represents the mass of the following car i, i∈1,2,…,N, N represents the number of following cars in the convoy; represents the yaw rate of the following vehicle i, is the arc length s of the following vehicle i i The time derivative, is the distance d between the center of mass of the following vehicle i and the projection point of the center of mass of the vehicle on the reference line i The differential of and denote the cornering stiffness of the front and rear wheels of the following vehicle i, δ i is the front wheel turning angle of the following vehicle i, γ i is the yaw rate of the following vehicle i, a i and b i are the distances between the front wheel and rear wheel turning wheelbase and the center of mass of vehicle i, is the moment of inertia of following vehicle i, represents the yaw angle of the following vehicle i, is the angle between the tangent line of the projection point of the following vehicle i on the reference line and the X-axis, k i is the road curvature corresponding to the projection point of the center of mass of the following vehicle i on the reference line.

[0080] S2: After discretizing the lateral and longitudinal coupling control model of the convoy using the Euler forward difference method, a discrete nonlinear model of each following vehicle is obtained to be applied to the subsequent DMPC optimization solution;

[0081] S2 is specifically:

[0082] S2.1: At discrete time t, the model is discretized using the Euler forward difference method:

[0083]

[0084] Among them, T s is the sampling time;

[0085] Define the system status as:

[0086]

[0087] The system input is:

[0088]

[0089] The system output is:

[0090] y i (t) = diag(1,0,0,0,1,0)X i (t)

[0091] S2.2: The discrete nonlinear model of each following vehicle satisfies the following formula:

[0092]

[0093]

[0094] Among them, the function X i (t+1) and X i (t) are the state sequences at time t+1 and time t, respectively, U i (t) is the control input at time t, T is the transpose, y i (t) is the control output at time t, φ i (·) is the nonlinear expression of the lateral and longitudinal coupling control model of the following vehicle i, T s is the sampling time, is the yaw angle of the following vehicle i at time t.

[0095] S3: construct the communication topology of the fleet to determine the information exchange strategy in the subsequent fleet;

[0096] S3 is specifically:

[0097] Assume a convoy of N+1 vehicles, with indices i∈{0,1,…,N}, driving on a non-constant straight road. 0 represents the lead vehicle, and the indexes of the following vehicles range from 1 to N. The vehicles in the convoy exchange information according to the topological structure.

[0098] The communication topology of the fleet is a directed graph To represent the node set Edge Set Adjacency Matrix It means that there is a directed edge from node j to node i, that is, vehicle i can receive information from node j. ij The definition is as follows:

[0099] when a ij =1, it means that vehicle i can receive information from vehicle j, that is, vehicle j is the neighbor of vehicle i, so the neighbor set of vehicle i can be expressed as follows:

[0100]

[0101] If there is a root node such that any other node of the directed graph G can be reached through at least one path starting from the root node, then the directed graph G is said to contain a directed spanning tree. Figure 4 As shown, the present invention adopts a front-leader-follower (PLF) topology.

[0102] S4: According to the communication topology of the convoy and the discrete nonlinear model of the follower vehicle, the distributed model predictive control method is used to optimize the state of each follower vehicle in the convoy to achieve convoy control.

[0103] Assume that the leading car is uncontrolled and always drives according to the target state, so the other states of the leading car are always ideal values, but the speed will change over time. Define the position and speed information of the leading car as s0 and v0.

[0104] Based on the communication topology, each following vehicle receives information from the leading vehicle and the preceding vehicle, performs optimization and solution using DMPC at time t, and transmits the status information to the following vehicle, thereby achieving consistent control of the convoy. In the experiment of the present invention, the number of following vehicles is 2, that is, there are 3 vehicles in the convoy.

[0105] Define the ideal state X for each vehicle i,des (k):

[0106]

[0107] in, s i,des (t), d i,des (t) are the ideal lateral velocity, ideal longitudinal velocity, ideal yaw rate, ideal yaw angle, and ideal distance between the ideal position and the reference line of the following vehicle i. That is, each following car must follow the speed of the leading car. That is, the lateral speed of the vehicle is equal to 0, so that the vehicle will not shake excessively. s i,des (t) = s0(t) - i*dis, where dis represents the distance and is a constant. In the present invention, the value of this constant is 15. i,des (t) = 0. s0 is the arc length of the leading vehicle.

[0108] For each following vehicle, define its control input sequence U i (:|k), state sequence X i (:|k) and the output sequence.

[0109] U i (:|t)={U i (0|t),Ui (1|t),…,U i (N p -1|t)}

[0110] X i (:|t)={X i (0|t),X i (1|t),…,X i (N p |t)}

[0111] Among them, U i (0|t),U i (1|t),U i (N p -1|t) are respectively the 0th moment, the 1st moment, the Nth moment in the prediction time domain at time t. p -1 time control input value, X i (0|t),X i (1|t),X i (N p |t) are respectively the 0th moment, the 1st moment, the Nth moment in the prediction time domain at time t p -1 The vehicle status value at time N p is the prediction time domain, and in the present invention, N is set p =6.

[0112]

[0113] Among them, the road curvature k i (c|t) and the angle between the tangent line of the projection point and the X axis is a quantity known in advance. Define the output sequence y i (c|t):

[0114] y i (c|t)=diag(1,0,0,0,1,0)X i (c|t),c=0,1,…,N p

[0115] S4 is specifically:

[0116] S4.1: Construct the optimization problem for each following vehicle, the formula is as follows:

[0117]

[0118]

[0119] in, is the predicted input sequence of the following car i, J i (·) is the cost function, is the predicted state sequence of the following vehicle i, is the predicted output sequence of the following car i, is the hypothetical output sequence of the following car i, is the hypothetical output sequence of vehicle j, is the predicted position sequence of the following vehicle i, s0(:|t) is the position sequence of the leading vehicle, is the state at time t+c predicted at time t, represents the unknown input variable to be optimized, X i (t) is the state of the following vehicle i at time t, φ i (·) is the nonlinear expression of the single-vehicle lateral and longitudinal coupling model, and are the lower and upper bounds of the input acceleration constraint, 8 m / s respectively. 2 and -5m / s 2 , δ i,min and δ i,max are the lower and upper bounds of the input steering angle, 1rad and -1rad respectively, is the input acceleration, is the input steering angle, |·| is the absolute value, N p is the prediction time domain, X i,des (c|t) is the ideal state sequence of the following vehicle i, Q i is the first positive definite weight matrix, dis=15 is the ideal distance between two vehicles, s0(c|t) is the position sequence of the leading vehicle, is the predicted position sequence of the following vehicle i, G i is the second positive definite weight matrix, is the predicted output sequence of the following car i, is the hypothetical output sequence of the following vehicle i, F i is the third positive definite weight matrix, is the adjacency matrix of the following car i, M i is the fourth positive definite weight matrix, R i is the fifth positive definite weight matrix; diag() is the matrix diagonalization operation; and δ i,min ≤δ i (j|k)≤δ i,max is the input constraint. is a terminal constraint.

[0120] In the prediction time domain [0,N p -1], there are three control input sequences and state sequences.

[0121] Predict input; Assume control input; Optimal control input.

[0122] Prediction status; Assumption status; Optimal state.

[0123] S4.2: According to the communication topology of the convoy and the discrete nonlinear model of the follower vehicle, the distributed model predictive control method is used to optimize and solve the optimization problem of each follower vehicle, so as to control each follower vehicle, and the optimization problems of all follower vehicles in the convoy are traversed and optimized to control all follower vehicles, thereby realizing the control of the convoy.

[0124] In S4.2, the distributed model predictive control method is used to optimize and solve the optimization problem of each following vehicle to obtain the state of each following vehicle, specifically:

[0125] Step 1: At t = 0, the parameters are initialized:

[0126] Specifically, at time t=0, the initialization parameter N p , T s , Q i , G i , F i , M i , R i , and the assumed control input sequence for initializing each follower vehicle i and the assumed state sequence X i (t) is the preset initial state of each vehicle; assuming the output sequence

[0127] φ i () is the lateral and longitudinal coupling control model of the following vehicle i;

[0128] Step 2: At time t>0, after optimizing and solving the optimization problem described in S4.1 for each following vehicle i, the optimal control input sequence is obtained and its own hypothetical output sequence is transmitted to the adjacent following vehicle j, i.e., the following vehicle, according to the communication topology of the convoy, and the hypothetical output sequence of the leading vehicle and the leading vehicle is accepted;

[0129] Step 3: Calculate the hypothetical control input sequence, hypothetical state sequence and hypothetical output sequence at the next moment based on the current optimal control input sequence. The calculation formula is as follows:

[0130]

[0131] in, is the hypothetical control input sequence of the ith following vehicle, is the hypothetical output sequence of the i-th following vehicle, and are the hypothetical state sequences at time t+c and time t+c+1 predicted at time t, respectively, i (·) is the lateral and longitudinal coupling control model of the i-th following vehicle, is the optimal control input sequence predicted by the following vehicle i at time t-1 at time t+c, is the t+Nth prediction of the following vehicle i at time t-1 p -2 The optimal control input sequence at time, and are the hypothetical state sequence of the following vehicle i and the optimal state sequence of the following vehicle i respectively;

[0132] Step 4: Apply the first value of the current optimal control input sequence of each following vehicle to the following vehicle i as the current real control quantity of the following vehicle i to control the acceleration and front wheel steering angle of the following vehicle i;

[0133] Step 5: At the next moment, based on the current assumed control input sequence, assumed state sequence and assumed output sequence of other vehicles, repeat Step 2-Step 4 and start rolling optimization solution again, so as to continuously control the following vehicle i.

[0134] The present invention also proposes a fleet cooperative control system based on Frenet coordinate system horizontal and vertical coupling distributed model predictive control, including:

[0135] The fleet control model building module is used to build the lateral and longitudinal coupling control model of the fleet by combining the Frenet coordinate system and using the vehicle three-degree-of-freedom bicycle dynamics model;

[0136] A following vehicle discrete nonlinear model building module is used to discretize the lateral and longitudinal coupling control model of the convoy using the Euler forward difference method to obtain a discrete nonlinear model of each following vehicle;

[0137] A vehicle fleet communication topology structure generation module, used to generate a vehicle fleet communication topology structure;

[0138] The fleet controller is used to optimize the state of each follower vehicle in the fleet based on the communication topology of the fleet and the discrete nonlinear model of the follower vehicle by using the distributed model predictive control method, and control all follower vehicles in real time.

[0139] The simulation of this embodiment is performed by joint simulation of MATLAB and carsim. The simulation scene diagram of carsim is as follows: Figure 7 The simulation results are shown in Figure 6 As shown, Figure 6(a) Figure 6 (b) Figure 6 (c) Figure 6 (d) Figure 6 (e), Figure 6 (f) and Figure 6 (g) are respectively the simulation graphs of vehicle speed changing with time, vehicle spacing error changing with time, vehicle lateral error changing with time, vehicle lateral speed changing with time, vehicle position changing with time, vehicle yaw rate changing with time and vehicle yaw angle changing with time. It can be seen from the figure that the following vehicles in the convoy can follow the speed of the leading vehicle very well, and the spacing error between each following vehicle and the leading vehicle is within a very small range, which proves the safety of the convoy, and each vehicle is very close to the center of the reference line, which proves that the vehicle is on the reference trajectory.

[0140] According to the above experimental results, we can find that:

[0141] 1. A three-degree-of-freedom vehicle dynamics model is introduced for vehicle control. Compared with the high-degree-of-freedom model, it greatly reduces the computational burden while ensuring a certain degree of accuracy of the model.

[0142] 2. The Frenet coordinate system of the vehicle is introduced and combined with the three-degree-of-freedom vehicle dynamics model to form a convoy control model, ensuring that the convoy can maintain good consistency of vehicles in the convoy and the safety of the convoy even on non-constant straight road conditions.

[0143] 3. The DMPC control algorithm is adopted. Compared with the traditional PID control and SMC, the design is simpler, and it can effectively constrain the state of the vehicle, and the solution speed is faster than the centralized MPC control algorithm.

[0144] 4. In view of the changes in road curvature in reality, the above-mentioned lateral and longitudinal coupled convoy control method was tested and good experimental results were obtained, which proved the effectiveness of the above-mentioned convoy control method.

[0145] The above is only a preferred embodiment of the present invention. Although the present invention has been disclosed as a preferred embodiment, it is not intended to limit the present invention. Any technician familiar with the art can make many possible changes and modifications to the technical solution of the present invention by using the above disclosed methods and technical contents without departing from the scope of the technical solution of the present invention, or modify it into an equivalent embodiment of equivalent changes. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the scope of protection of the technical solution of the present invention.

Claims

1. A convoy cooperative control method based on the Frenet coordinate system with lateral and longitudinal coupling, characterized in that: The following steps are involved: S1: Combined with the Frenet coordinate system, the lateral and longitudinal coupling control model of the fleet is constructed using the vehicle three-degree-of-freedom bicycle dynamics model; S2: After discretizing the lateral and longitudinal coupling control model of the convoy using the Euler forward difference method, a discrete nonlinear model of each following vehicle is obtained; S3: Build the communication topology of the fleet; S4: According to the communication topology of the convoy and the discrete nonlinear model of the follower vehicle, the distributed model predictive control method is used to optimize the state of each follower vehicle in the convoy to achieve convoy control.

2. The method for cooperative control of a convoy based on the Frenet coordinate system according to claim 1, characterized in that: In S1, the lateral and longitudinal coupling control model of the fleet satisfies the following formula: in, and are the longitudinal speeds of the following vehicle i Vehicle lateral speed The differential of is the longitudinal acceleration of the following vehicle i in the vehicle body coordinate system, m i represents the mass of the following car i, i∈1,2,…,N, N represents the number of following cars in the convoy; represents the yaw rate of the following vehicle i, is the arc length s of the following vehicle i i The time derivative, is the distance d between the center of mass of the following vehicle i and the projection point of the center of mass of the following vehicle i on the reference line i The differential of and denote the cornering stiffness of the front and rear wheels of the following vehicle i, δ i is the front wheel turning angle of the following vehicle i, γ i is the yaw rate of the following vehicle i, a i and b i are the distances between the front wheel and rear wheel turning wheelbase and the center of mass of vehicle i, is the moment of inertia of following vehicle i, represents the yaw angle of the following vehicle i, is the angle between the tangent line of the projection point of the following vehicle i on the reference line and the X-axis, k i is the road curvature corresponding to the projection point of the center of mass of the following vehicle i on the reference line.

3. The method for coordinated control of a fleet based on the Frenet coordinate system according to claim 2, characterized in that: The discrete nonlinear model of each following vehicle satisfies the following formula: Among them, X i (t+1) and X i (t) are the state sequences at time t+1 and time t, respectively, U i (t) is the control input at time t, y i (t) is the control output at time t, φ i (·) is the lateral and longitudinal coupling control model of the following vehicle i, T s is the sampling time, is the yaw angle of the following vehicle i at time t.

4. The method for coordinated control of a fleet of vehicles based on the Frenet coordinate system according to claim 1, characterized in that: In S3, the communication topology of the convoy is a directed graph including a directed spanning tree, wherein the leading vehicle of the convoy is taken as a root node, and the following vehicles are taken as other nodes in the directed spanning tree, and the root node is connected to the other nodes.

5. The method for coordinated control of a fleet of vehicles based on the Frenet coordinate system in accordance with claim 1, characterized in that: The S4 is specifically: S4.1: Construct the optimization problem for each following vehicle, the formula is as follows: in, is the predicted input sequence of the following car i, J i (·) is the cost function, is the predicted state sequence of the following vehicle i, is the predicted output sequence of the following car i, is the hypothetical output sequence of the following car i, is the hypothetical output sequence of vehicle j, is the predicted position sequence of the following vehicle i, s0(:|t) is the position sequence of the leading vehicle, is the state at time t+c predicted at time t, represents the unknown input variable to be optimized, X i (t) is the state of the following vehicle i at time t, φ i (·) is the nonlinear expression of the single-vehicle lateral and longitudinal coupling model, and are the lower and upper bounds of the input acceleration constraint, δ i,min and δ i,max are the lower and upper bounds of the input steering angle, is the input acceleration, is the input steering angle, |·| is the absolute value, N p is the prediction time domain, X i,des (c|t) is the ideal state sequence of the following vehicle i, Q i is the first positive definite weight matrix, s0(c|t) is the position sequence of the lead vehicle, is the predicted position sequence of the following vehicle i, G i is the second positive definite weight matrix, is the predicted output sequence of the following car i, is the hypothetical output sequence of the following vehicle i, F i is the third positive definite weight matrix, is the adjacency matrix of the following car i, M i is the fourth positive definite weight matrix, R i is the fifth positive definite weight matrix; diag() is the matrix diagonalization operation; S4.2: According to the communication topology of the convoy and the discrete nonlinear model of the follower vehicle, the distributed model predictive control method is used to optimize and solve the optimization problem of each follower vehicle, so as to control each follower vehicle, and the optimization problems of all follower vehicles in the convoy are traversed and optimized to control all follower vehicles, thereby realizing the control of the convoy.

6. The method for coordinated control of a fleet of vehicles based on the Frenet coordinate system according to claim 5, characterized in that: In S4.2, the distributed model predictive control method is used to optimize and solve the optimization problem of each following vehicle, so as to control each following vehicle, specifically: Step 1: At t = 0, the parameters are initialized: Specifically, at time t=0, the assumed control input sequence of each following vehicle i is initialized and the assumed state sequence X i (t) is the preset initial state of each vehicle; assuming the output sequence φ i ( ) is the lateral and longitudinal coupling control model of the following vehicle i; Step 2: At time t>0, after optimizing and solving the optimization problem described in S4.1 for each following vehicle i, the optimal control input sequence is obtained and its own hypothetical output sequence is transmitted to the adjacent following vehicle j according to the communication topology of the convoy, and the hypothetical output sequence of the preceding vehicle and the leading vehicle is accepted; Step 3: Calculate the hypothetical control input sequence, hypothetical state sequence and hypothetical output sequence at the next moment based on the current optimal control input sequence. The calculation formula is as follows: in, is the hypothetical control input sequence of the ith following vehicle, is the hypothetical output sequence of the i-th following vehicle, and are the hypothetical state sequences at time t+c and time t+c+1 predicted at time t, respectively, i (·) is the lateral and longitudinal coupling control model of the i-th following vehicle, is the optimal control input sequence predicted by the following vehicle i at time t-1 at time t+c, is the t+Nth prediction of the following vehicle i at time t-1 p -2 The optimal control input sequence at time, and are the hypothetical state sequence of the following vehicle i and the optimal state sequence of the following vehicle i respectively; Step 4: Apply the first value of the current optimal control input sequence of each following vehicle to the following vehicle i as the current real control quantity of the following vehicle i to control the acceleration and front wheel steering angle of the following vehicle i; Step 5: At the next moment, based on the current assumed control input sequence, assumed state sequence and assumed output sequence of other vehicles, repeat Step 2-Step 4 and start rolling optimization solution again, so as to continuously control the following vehicle i.

7. A convoy coordinated control system based on the Frenet coordinate system with lateral and longitudinal coupling, characterized in that: include: The fleet control model building module is used to build the lateral and longitudinal coupling control model of the fleet by combining the Frenet coordinate system and using the vehicle three-degree-of-freedom bicycle dynamics model; The following vehicle discrete nonlinear model building module is used to discretize the lateral and longitudinal coupling control model of the convoy using the Euler forward difference method to obtain the discrete nonlinear model of each following vehicle; A vehicle fleet communication topology structure generation module, used to generate a vehicle fleet communication topology structure; The fleet controller is used to optimize the state of each follower vehicle in the fleet based on the communication topology of the fleet and the discrete nonlinear model of the follower vehicle by using the distributed model predictive control method, and control all follower vehicles in real time.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images