Interconnected autonomous vehicle formation dynamic constraint control method based on TSSZNN-DC1 / 2

Abstract

The invention discloses an interconnected autonomous vehicle formation dynamic constraint control method based on TSSZNN-DC1 / 2. The method comprises the following steps: constructing a dynamic constraint control system of an interconnected autonomous vehicle formation; and constructing a TSSZNN-DC1 discrete model and a TSSZNN-DC2 discrete model, and realizing more optimized dynamic constraint control of the interconnected autonomous vehicle formation under a certain condition. According to the method, a time-varying bounded attenuation coefficient based on topology feedback is designed and adopted in the design of a TSSZNN-DC1 discrete model and a TSSZNN-DC2 discrete model, so that the model has a higher convergence speed, meanwhile, the system burden caused by overlarge data is reduced, and meanwhile, a strong bounded constraint function based on error norm gain is designed; dynamic constraint control is carried out on the vehicle formation, the formation condition in an actual traffic scene is better met, the step-by-step gain coefficient based on errors is adopted, the anti-interference performance of the model is enhanced, and the model has the advantages of being high in convergence speed, high in precision, low in time delay, high in anti-interference capacity and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of formation control of interconnected autonomous vehicles, and particularly to a dynamic constraint control method for interconnected autonomous vehicle formations based on TSSZNN-DC1 / 2 Background Art

[0002] In the research of traditional interconnected autonomous vehicle formations, controllers are mainly designed as required and combined with specific observers to achieve the control of the position, speed, and acceleration of each vehicle in the interconnected autonomous formation. There are many optimized solutions for the formation control of a small number of vehicles on a straight road. However, when the number of vehicles designed in the vehicle formation is large, the vehicle controller will not only encounter external interference from the road when solving the system equations, but also be affected by noise interference generated by its own sensors and other problems. This is not desirable when implementing the formation control of interconnected autonomous vehicles. Some literature has proposed designing a nonlinear manifold and a distributed observer to solve the interference problem and thus obtain an accurate solution. However, this method mostly uses fixed parameters, which is a relatively inefficient process in experiments, and there will be problems such as slow system convergence speed and time delay

[0003] To solve this problem, the prior art literature has proposed a method of using time-varying parameters for the observer. After introducing time-varying parameters, the convergence time of the system is shortened. However, due to the introduction of time-varying increasing parameters in the observer, as time increases, the time-varying parameters will tend to infinity, resulting in damage to the observer equipment, increased computational cost, making it difficult to solve in real time, and there will be a situation where the magnitude of the acceleration is too large, which has an adverse impact on the comfort experience of passengers, and the magnitude of the acceleration does not conform to the actual road conditions and vehicle performance Summary of the Invention

[0004] Aiming at the above pain points in the prior art, the present invention provides a dynamic constraint control method for interconnected autonomous vehicle formations based on TSSZNN-DC1 / 2, where TSSZNN-DC1 / 2 represents the TSSZNN-DC1 discrete model or the TSSZNN-DC2 discrete model, and the specific meaning is: including a time-varying bounded attenuation coefficient based on topological feedback, a strong bounded dynamic constraint function based on error norm gain, and a nullifying neural network dynamic constraint discrete model 1 or 2 based on an error-based step-by-step gain coefficient. It solves the problems that it is difficult to effectively control the interconnected autonomous vehicle formation stably in the prior art and it is difficult to achieve effective acceleration dynamic constraints

[0005] In order to achieve the above invention purpose, the technical solution adopted by the present invention is as follows

[0006] Provided is a method for dynamic constraint control of an interconnected autonomous vehicle formation based on TSSZNN-DC1 / 2, which includes the following steps:

[0007] S1. Set up the longitudinal dynamics system of the interconnected autonomous vehicle;

[0008] S2. Use the feedback linearization technique to process the longitudinal dynamics system equation of the interconnected autonomous vehicle and add interference to obtain the linear longitudinal dynamics system of the interconnected autonomous vehicle;

[0009] S3. Based on the obtained linear longitudinal dynamics system of the interconnected autonomous vehicle, construct the longitudinal dynamics system of the virtual signal interconnected autonomous vehicle;

[0010] S4. Use the forward Euler discretization method to discretize the linear longitudinal dynamics system equation processed in S2 and the longitudinal dynamics system equation of the virtual signal interconnected autonomous vehicle obtained in S3 respectively, and correspondingly obtain the discretized longitudinal dynamics system equation of the vehicle fleet and the discretized virtual longitudinal dynamics system;

[0011] S5. According to the discretized longitudinal dynamics system equation of the vehicle fleet and the discretized virtual longitudinal dynamics system equation, construct the dynamic constraint control system equation of the interconnected autonomous vehicle formation and its error function;

[0012] S6. According to the discretized longitudinal dynamics system equation of the vehicle fleet and the discretized virtual longitudinal dynamics system, construct the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model for the undirected communication topology and the directed communication topology respectively;

[0013] S7. Through the constructed TSSZNN-DC1 discrete model and TSSZNN-DC2 discrete model, observe the interconnected autonomous vehicle in different topological situations and perform on-site dynamic constraint control to complete the dynamic constraint control of the interconnected autonomous vehicle formation.

[0014] The beneficial effects of the present invention are as follows: In the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model of the present invention, new distributed observers, controllers, nonlinear manifolds, time-varying bounded attenuation coefficients based on topological feedback, strongly bounded dynamic constraint functions based on error norm gain, and step-by-step gain coefficients based on errors are designed and adopted. The new distributed observer enables the present invention to track the signals of virtual signal vehicles in real time. The controller enables the present invention to converge within a finite time. The nonlinear manifold enables the model to enable the present invention to converge even under noise interference. The time-varying bounded attenuation coefficient based on topological feedback can prevent the overfitting of the observer parameters of the present invention. The strongly bounded dynamic constraint function based on error norm gain enables the present invention to effectively dynamically constrain the acceleration of the vehicle, optimizing the comfort experience of the vehicle formation for passengers and being more in line with the formation situation under actual traffic scenarios. The step-by-step gain coefficient based on error enhances the anti-interference ability of the controller, having the advantages of high precision and low latency. The present invention is superior to existing solutions in terms of convergence speed, accuracy, and robustness. Description of the Drawings

[0015] Figure 1 It is a schematic flowchart of the method;

[0016] Figure 2 It is an undirected communication topology diagram provided in the embodiment of the present application based on the method;

[0017] Figure 3 It is a following-leader mode diagram of the directed communication topology provided in the embodiment of the present application based on the method;

[0018] Figure 4 It is a virtual signal-following-leader mode diagram of the directed communication topology provided in the embodiment of the present application based on the method;

[0019] Figure 5 It is a comparison diagram of the convergence accuracy of the method based on topological feedback time-varying bounded attenuation coefficient and the use of exponential function coefficients, fixed coefficients, and linear function coefficients to solve the interconnected autonomous vehicle formation control system to achieve the ideal distance between vehicles under the conditions of bounded Gaussian white noise interference and road interference for the undirected communication topology formation;

[0020] Figure 6 It is a comparison diagram of the convergence accuracy of the method based on topological feedback time-varying bounded attenuation coefficient and the use of exponential function coefficients, fixed coefficients, and linear function coefficients to solve the interconnected autonomous vehicle formation control system to achieve the ideal distance between vehicles under the conditions of bounded Gaussian white noise interference and road interference for the following-leader mode of the directed communication topology formation;

[0021] Figure 7Comparison graph of the convergence accuracy of the time-varying bounded attenuation coefficient method based on topological feedback and the methods using exponential function coefficients, fixed coefficients, and linear function coefficients to achieve the ideal distance between vehicles in the virtual signal - leading vehicle following mode of the directed communication topology formation under bounded Gaussian white noise interference and road interference conditions;

[0022] Figure 8 Comparison graph of the actual position, speed, and acceleration of each vehicle in the interconnected autonomous vehicle formation dynamic constraint control system equation solved by this method with the expected values under bounded Gaussian white noise and road interference conditions in the undirected communication topology connection state;

[0023] Figure 9 Error graph of the actual position, speed, and acceleration of each vehicle in the interconnected autonomous vehicle formation control system solved by this method with the expected values under bounded Gaussian white noise and road interference conditions; where (a) is the position difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle; (b) is the position difference between adjacent interconnected autonomous vehicles; (c) is the speed difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle; (d) is the acceleration difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle;

[0024] Figure 10 Comparison graph of the actual position, speed, and acceleration of each vehicle in the interconnected autonomous vehicle formation dynamic constraint control system equation solved by this method with the expected values under bounded Gaussian white noise and road interference conditions in the leading vehicle following mode connection state of the directed communication topology;

[0025] Figure 11 Error graph of the actual position, speed, and acceleration of each vehicle in the interconnected autonomous vehicle formation control system solved by this method with the expected values under bounded Gaussian white noise and road interference conditions in the leading vehicle following mode connection state of the directed communication topology; where (a) is the position difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle; (b) is the position difference between adjacent interconnected autonomous vehicles; (c) is the speed difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle; (d) is the acceleration difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle;

[0026] Figure 12 Comparison graph of the actual position, speed, and acceleration of each vehicle in the interconnected autonomous vehicle formation dynamic constraint control system equation solved by this method with the expected values under bounded Gaussian white noise and road interference conditions in the virtual signal - leading vehicle following mode connection state of the directed communication topology;

[0027] Figure 13In the presence of bounded Gaussian white noise and road disturbances, the error graphs of the actual position, speed, and acceleration of each vehicle in the interconnected autonomous vehicle formation control system in the virtual signal - leading vehicle following mode connection state of the directed communication topology solved by this method compared with the expected values; where (a) is the position difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle; (b) is the position difference between adjacent interconnected autonomous vehicles; (c) is the speed difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle; (d) is the acceleration difference between the interconnected autonomous vehicle and the virtual interconnected autonomous vehicle.

[0028] Figure 14 In the presence of bounded Gaussian white noise and road disturbances, the comparison graph of the actual acceleration of each vehicle in the interconnected autonomous vehicle formation dynamic constraint control system equation in the undirected communication topology connection state solved by this method with the actual acceleration solved without the strong bounded constraint function term under the same other conditions.

[0029] Figure 15 In the presence of bounded Gaussian white noise and road disturbances, the comparison graph of the actual acceleration of each vehicle in the interconnected autonomous vehicle formation dynamic constraint control system equation in the leading vehicle following mode connection state of the directed communication topology solved by this method with the actual acceleration solved without the strong bounded constraint function term under the same other conditions.

[0030] Figure 16 In the presence of bounded Gaussian white noise and road disturbances, the comparison graph of the actual acceleration of each vehicle in the interconnected autonomous vehicle formation dynamic constraint control system equation in the virtual signal - leading vehicle following mode connection state of the directed communication topology solved by this method with the actual acceleration solved without the strong bounded constraint function term under the same other conditions. Detailed implementation manners

[0031] The following elaborates on the detailed implementation manners of the present invention to facilitate those skilled in the art of this technology to understand the core content of the present invention. However, it should be clear that the present invention is not limited to the scope of the detailed implementation manners. Based on the idea of the present invention, there will be changes in both the detailed implementation manners and the application scope. For those ordinary skilled in the art of this technology, as long as various changes are within the scope of the present invention defined and determined by the appended claims, these changes are obvious. In summary, all inventions created using the concept of the present invention are within the scope of protection.

[0032] As Figure 1 shown, the interconnected autonomous vehicle formation dynamic constraint control method based on TSSZNN - DC1 / 2 includes the following steps:

[0033] S1. Set up the longitudinal dynamics system of the interconnected autonomous vehicle.

[0034] In this embodiment, the expression of the longitudinal dynamics system equation of the vehicle is as follows:

[0035]

[0036] where and respectively represent the position, speed, actual driving / braking torque, and desired driving / braking torque of the th interconnected autonomous vehicle; and are respectively the first derivatives of and ; respectively represent the resultant external force, mass, mechanical efficiency of the power transmission system, tire radius, combined aerodynamic drag coefficient, gravity, rolling resistance coefficient, and inertial delay of longitudinal dynamics of the th interconnected autonomous vehicle; t represents time.

[0037] S2. Use the feedback linearization technique to process the longitudinal dynamics system equation of the interconnected autonomous vehicle and add disturbances to obtain the linear longitudinal dynamics system of the interconnected autonomous vehicle;

[0038] The specific method of using the feedback linearization technique to process the longitudinal dynamics system equation of the interconnected autonomous vehicle and add disturbances in step S2 includes the following sub-steps:

[0039] S21. Construct the expression of the feedback linearization technique:

[0040]

[0041] where respectively represent the gravitational acceleration and the linearized control input of the th interconnected autonomous vehicle;

[0042] S22. Apply the linearized expression to the longitudinal dynamics system equation of the interconnected autonomous vehicle to obtain the linearized longitudinal dynamics system equation of the vehicle, and its expression is:

[0043]

[0044] where is the acceleration of the th interconnected autonomous vehicle; is 's first derivative;

[0045] S23. Add interference to the longitudinal dynamics system equation of the linearized vehicle to obtain the longitudinal dynamics system equation after adding interference, and its expression is:

[0046]

[0047] where is the interference of the th connected autonomous vehicle;

[0048] S24. Let , and obtain the rewritten longitudinal dynamics system equation, and its expression is:

[0049]

[0050] where is the interference after the new control input of the th connected autonomous vehicle; is the dynamic constraint controller to be designed for the th connected autonomous vehicle.

[0051] S3. Construct the longitudinal dynamics system of the virtual signal connected autonomous vehicle based on the obtained linear longitudinal dynamics system of the connected autonomous vehicle;

[0052] The expression of the longitudinal dynamics system equation of the virtual signal connected autonomous vehicle in step S3 is:

[0053]

[0054] where , , and respectively represent the position, speed, acceleration and control input of the virtual signal connected autonomous vehicle; , and are respectively the first-order derivatives of , and .

[0055] S4. Use the forward Euler discretization method to discretize the linear longitudinal dynamics system equation after S2 processing and the longitudinal dynamics system equation of the virtual signal connected autonomous vehicle obtained in S3 respectively, and correspondingly obtain the discretized longitudinal dynamics system equation of the vehicle fleet and the discretized virtual longitudinal dynamics system;

[0056] The expression of the discretized longitudinal dynamics system equation of the vehicle fleet in step S4 is:

[0057]

[0058] Among them 、 、 、 and respectively represent the sampling values of the th connected and autonomous vehicle at moment for 、 、 、 and ; 、 and respectively represent the first derivatives of 、 and .

[0059] The discretized virtual longitudinal dynamics system equation in step S4, and its expression is:

[0060]

[0061] Among them 、 、 and respectively represent the sampling values of the virtual signal connected and autonomous vehicle at moment for 、 、 and ; 、 and respectively represent the first derivatives of 、 and .

[0062] S5. Construct the connected and autonomous vehicle formation dynamic constraint control system equation and its error function according to the discretized platoon longitudinal dynamics system equation and the discretized virtual longitudinal dynamics system equation;

[0063] The expression of the connected and autonomous vehicle formation dynamic constraint control system equation in step S5 is:

[0064]

[0065] Among them , the th element in this column vector is the sampling value of the th connected and autonomous vehicle at moment for ; , the The rd connected autonomous vehicle takes a sample value of at ; the th element in this column vector is the rd connected autonomous vehicle's sample value of at ; the th element in this column vector is the rd connected autonomous vehicle's sample value of at ; the th element in this column vector is the rd connected autonomous vehicle's sample value of at ; and are respectively the -th, -th and -th first-order derivatives; represents the transpose of a matrix; is the total number of connected autonomous vehicles; for the disturbance , its expression is:

[0066]

[0067] where represents taking the minimum of the two, represents taking the maximum of the two, is the set upper bound of the acceleration, is the set lower bound of the acceleration, represents Gaussian white noise, represents bounded Gaussian white noise;

[0068] The expression of the error function of the connected autonomous vehicle formation dynamic constraint control system equation is:

[0069]

[0070] where and and are respectively the column vectors composed of the -th, -th and -th observed values of each connected autonomous vehicle for the virtual signal connected autonomous vehicle; , Denote the column vector composed of the designed distances between each connected and autonomous vehicle and the virtual signal connected and autonomous vehicle. is the desired spacing between adjacent vehicles designed; 、 and are the column vectors composed of the error values of each connected and autonomous vehicle relative to the virtual signal connected and autonomous vehicle in terms of position, speed, and acceleration, respectively.

[0071] S6. Construct the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model respectively according to the discretized platoon longitudinal dynamics system equation and the discretized virtual longitudinal dynamics system and for the undirected communication topology and the directed communication topology;

[0072] Both the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model in step S6 include a distributed observer, a dynamic constraint controller, a nonlinear manifold, a time-varying bounded decay coefficient based on topological feedback, a strong bounded dynamic constraint function based on error norm gain, and a step-by-step gain coefficient based on error; among which the expression of the distributed observer of the TSSZNN-DC1 discrete model for the undirected communication topology is:

[0073]

[0074] where 、 and represent respectively 、 and the values at the th step, is the time step; , is the enhanced Laplacian matrix of the connected and autonomous vehicle, where is the Laplacian matrix of the undirected graph, , where is the adjacency matrix of the undirected graph, when the th vehicle maintains communication with the th vehicle, otherwise it is all 0, is the diagonal matrix of, , where n represents the number of vehicles, , , …, ; is the communication link matrix between the connected and autonomous vehicle and the virtual signal connected and autonomous vehicle, when the th signal connected and autonomous vehicle can communicate with the virtual signal connected and autonomous vehicle, , otherwise ; , , ; is the sign function, and its expression is:

[0075]

[0076] is the time-varying bounded attenuation coefficient based on topological feedback, and its expression is: , where , are constants to be set, represents the minimum eigenvalue of the matrix, represents the sampling time;

[0077] The dynamic constraint controller of the TSSZNN-DC1 discrete model has the following expression:

[0078]

[0079] where , , , are constants to be set; , , are the step gain coefficients based on the error, and the expressions are: , , , , , are constants to be set, represents the two-norm; As a whole, it represents a strongly bounded dynamic constraint function based on the error norm gain, where the error norm gain has the following expression: , represents the one-norm, where represents the settable boundary, , , where: , is the settable upper bound of the acceleration, is the settable lower bound of the acceleration; is the nonlinear manifold, and its expression is:

[0080] In step S6, the expression of the distributed observer of the TSSZNN-DC2 discrete model for the directed communication topology is:

[0081]

[0082] Among them 、 and respectively represent 、 and the value of the th step, and is the time step; among them is the Laplacian matrix of the directed communication topology graph of connected autonomous vehicles, and there are two topological cases as follows:

[0083]

[0084]

[0085] represents the leading vehicle following mode, represents the virtual signal - leading vehicle following mode; represents a constant to be set; , , ; is the sign function, and its expression is:

[0086]

[0087] is the time - varying bounded attenuation coefficient based on topological feedback, and its expression is: where 、 are constants to be set, , , , , represents obtaining the minimum eigenvalue of the matrix, n represents the number of vehicles, represents the sampling time;

[0088] The dynamic constraint controller of the TSSZNN - DC2 discrete model has the following expression:

[0089]

[0090] where , 、 、 are constants to be set; 、 、 are the step - by - step gain coefficients based on the error, and the expression is: 、 、 , , , is a constant to be set, represents the two-norm; as a whole represents a strongly bounded dynamic constraint function based on the error norm gain, where the error norm gain has the expression: , represents the one-norm, where represents the settable boundary, , , where: , is the settable upper bound of the acceleration, is the settable lower bound of the acceleration; is a non-linear manifold, and its expression is:

[0091] .

[0092] S7. By constructing the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model, observe interconnected autonomous vehicles in different topological situations and perform real-time dynamic constraint control to complete the dynamic constraint control of the interconnected autonomous vehicle formation;

[0093] In step S7, the expressions for the first-order derivatives of the positions, speeds, and accelerations of the interconnected autonomous vehicles in the interconnected autonomous vehicle formation dynamic constraint control system equation solved by the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model are:

[0094]

[0095] The specific method for solving the interconnected autonomous vehicle formation dynamic constraint control system equation by the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model is:

[0096] Let

[0097] where ; represents the iterative update value of; is the first-order derivative of; is the acceleration of the virtual signal interconnected autonomous vehicle; is the speed of the virtual signal interconnected autonomous vehicle; is represented as , is represented as , is represented as ;

[0098] By substituting into the TSSZNN-DC1 discrete model or the TSSZNN-DC2 discrete model, the position, speed, and acceleration of each interconnected autonomous vehicle in the case of an undirected communication topology or a directed communication topology can be obtained.

[0099] In the specific implementation process, the acceleration of the virtual signal interconnected autonomous vehicle is expressed as:

[0100]

[0101] In an embodiment of the present invention, the three communication topology diagrams of the vehicle are as shown in Figure 2 , Figure 3 and Figure 4 . To verify the effect of the present invention, this method is compared with methods using other function coefficients (such as exponential function coefficients, fixed coefficients, and linear function coefficients), and the results are as shown in Figure 5 , Figure 6 , Figure 7 . It can be intuitively seen from Figures 5 - 7 that the convergence speed of this method to the desired inter-vehicle distance is faster than that of methods using other function coefficients; in terms of constraint ability, the constraint ability of this method in solving the dynamic constraint control system of interconnected autonomous vehicle formations is better than other solutions that do not use a strong bounded constraint function.

[0102] Figure 8 , Figure 9 , Figure 10 , Figure 11 , Figure 12 , Figure 13 , Figure 14 , Figure 15 and Figure 16 show the comparison diagrams and error diagrams of the actual position, speed, and acceleration of each vehicle with the expected values in solving the equations of the dynamic constraint control system of interconnected autonomous vehicle formations under bounded Gaussian white noise and road interference. It can be seen from Figures 8 - 16 that when the parameters are moderate, the errors between the actual position, speed, and acceleration of each vehicle and the expected values can all reach the -4 order and converge within a finite time. Thus, it can be seen that this method has the advantages of high precision and low latency in solving the equations of the dynamic constraint control system of interconnected autonomous vehicle formations.

Claims

1. A dynamic constraint control method for interconnected autonomous vehicle formations based on TSSZNN-DC1 / 2, characterized in that, It includes the following steps: S1. Set up the longitudinal dynamics system of the connected and autonomous vehicle; S2. Use the feedback linearization technique to process the longitudinal dynamics system equation of the connected and autonomous vehicle and add disturbances to obtain the linear longitudinal dynamics system of the connected and autonomous vehicle; S3. Based on the obtained linear longitudinal dynamics system of the connected and autonomous vehicle, construct the longitudinal dynamics system of the virtual signal connected and autonomous vehicle; S4. Use the forward Euler discretization method to discretize the linear longitudinal dynamics system equation processed in S2 and the longitudinal dynamics system equation of the virtual signal connected and autonomous vehicle obtained in S3 respectively, and correspondingly obtain the discretized longitudinal dynamics system equation of the vehicle platoon and the discretized virtual longitudinal dynamics system; S5. According to the discretized longitudinal dynamics system equation of the vehicle platoon and the discretized virtual longitudinal dynamics system equation, construct the connected and autonomous vehicle platoon dynamic constraint control system equation and its error function; S6. According to the discretized longitudinal dynamics system equation of the vehicle platoon and the discretized virtual longitudinal dynamics system, construct the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model for undirected communication topology and directed communication topology respectively; S7. Through the constructed TSSZNN-DC1 discrete model and TSSZNN-DC2 discrete model, observe the connected and autonomous vehicles in different topological situations and perform on-site dynamic constraint control to complete the connected and autonomous vehicle platoon dynamic constraint control.

2. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 1, characterized in that, The expression of the longitudinal dynamics system equation of the vehicle in step S1 is: ； Among them 、 and respectively represent the position, speed, actual driving / braking torque, and desired driving / braking torque of the th connected and autonomous vehicle; and are respectively the first derivatives of and ; respectively represent the resultant external force, mass, mechanical efficiency of the powertrain, tire radius, combined aerodynamic drag coefficient, gravity, rolling resistance coefficient, and inertial delay of longitudinal dynamics acting on the th connected and autonomous vehicle; t represents time.

3. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 2, characterized in that, The specific method of using the feedback linearization technique to process the longitudinal dynamics system equation of the connected and autonomous vehicle and add disturbances in step S2 includes the following sub-steps: S21. Construct the expression of the feedback linearization technique; ； wherein respectively represent the gravitational acceleration and the control input after linearization of the -th connected and automated vehicle S22. Apply the linearized expression to the longitudinal dynamics system equation of the connected and autonomous vehicle to obtain the linearized longitudinal dynamics system equation of the vehicle, and its expression is: ； wherein is the acceleration of the th connected and autonomous vehicle; is the first derivative; S23. Add disturbances to the linearized longitudinal dynamics system equation of the vehicle to obtain the longitudinal dynamics system equation after adding disturbances, and its expression is: ； Among them is the interference of the th connected and autonomous vehicle; S24. Let , and obtain the rewritten longitudinal dynamics system equation, whose expression is: ； wherein is the interference after the new control input of the th connected and autonomous vehicle; is the dynamic constraint controller to be designed for the th connected and autonomous vehicle.

4. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 3, characterized in that, The expression of the longitudinal dynamics system equation of the virtual signal connected and autonomous vehicle in step S3 is: ； wherein , , and respectively represent the position, speed, acceleration and control input of a virtual signal interconnected autonomous vehicle; , and are respectively the , and first derivatives.

5. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 4, characterized in that The expression of the discretized longitudinal dynamics system equation of the vehicle platoon in step S4 is: ； Among them , , , and respectively represent the sampling values of the th connected and autonomous vehicle at the moment for , , , and ; , and respectively represent the first-order derivatives of , and . The discretized virtual longitudinal dynamics system equation in step S4, and its expression is: ； Among them , , and respectively represent the sampled values of the virtual signal interconnected autonomous vehicle at moment for , , and . , and respectively represent the first-order derivatives of , and .

6. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 5, characterized in that, The expression of the connected and autonomous vehicle platoon dynamic constraint control system equation in step S5 is: ； Among them , it means that the -th element in this column vector is the sampling value of the -th connected and autonomous vehicle at moment for ; , it means that the -th element in this column vector is the sampling value of the -th connected and autonomous vehicle at moment for ; , it means that the -th element in this column vector is the sampling value of the -th connected and autonomous vehicle at moment for ; , it means that the -th element in this column vector is the sampling value of the -th connected and autonomous vehicle at moment for ; , it means that the -th element in this column vector is the sampling value of the -th connected and autonomous vehicle at moment for ; , and are the first-order derivatives of , and respectively; represents the transpose of a matrix; is the total number of connected and autonomous vehicles; for the interference , its expression is: ； wherein represents taking the minimum value of the two, represents taking the maximum value of the two, is the set upper bound of the acceleration, is the set lower bound of the acceleration, represents Gaussian white noise, represents bounded Gaussian white noise; The expression of the error function of the connected and autonomous vehicle platoon dynamic constraint control system equation is: ； Among them 、 and are column vectors composed of the observed values of the 、 and for each connected and autonomous vehicle with respect to the virtual signal connected and autonomous vehicle; , represents the column vector composed of the designed distances between each connected and autonomous vehicle and the virtual signal connected and autonomous vehicle, is the desired spacing between adjacent vehicles designed; 、 and are column vectors composed of the error values of each connected and autonomous vehicle relative to the virtual signal connected and autonomous vehicle in terms of position, speed, and acceleration, respectively.

7. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 6, characterized in that, Both the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model in step S6 include a distributed observer, a dynamic constraint controller, a nonlinear manifold, a time-varying bounded attenuation coefficient based on topological feedback, a strongly bounded dynamic constraint function based on error norm gain, and a step-by-step gain coefficient based on error. The specific method includes the following sub-steps: The expression of the distributed observer of the TSSZNN-DC1 discrete model for undirected communication topology is: ； Among them 、 and respectively represent 、 and the value of the th step, being the time step; , is the Laplacian matrix enhanced for connected autonomous vehicles, where is the Laplacian matrix of the undirected graph, , where is the adjacency matrix of the undirected graph. When the th vehicle maintains communication with the th vehicle, otherwise it is all 0, is the diagonal matrix of , , where n represents the number of vehicles, , ,…, ; is the communication link matrix between connected autonomous vehicles and virtual signal connected autonomous vehicles. When the th signal connected autonomous vehicle can communicate with the virtual signal connected autonomous vehicle, , otherwise ; , , ; is the sign function, and its expression is: ； is the time-varying bounded attenuation coefficient based on topological feedback, and its expression is: , where and are constants to be set, represents the minimum eigenvalue of the obtained matrix, represents the sampling time; Dynamic Constraint Controller for TSSZNN-DC1 Discrete Model The expression of which is as follows: ； Among them , 、 、 are constants to be set; 、 、 are step gain coefficients based on the error, and the expression is: 、 、 , 、 、 are constants to be set, represents the two-norm; as a whole represents a strongly bounded dynamic constraint function based on the error norm gain, where the error norm gain has the expression: , is represented as the one-norm, where represents the settable boundary, 、 , where: , is the settable upper bound of the acceleration, is the settable lower bound of the acceleration; is a non-linear manifold, and its expression is: ； The expression of the distributed observer for the TSSZNN-DC2 discrete model for the directed communication topology is: ； Among them , and respectively represent , and the value of the th step, where is the Laplacian matrix of the directed communication topology graph of the connected and autonomous vehicles, and is divided into two topological cases as follows: ； ； It is expressed as the following-vehicle following mode, It is expressed as the virtual signal - following-vehicle following mode; It is expressed as a constant to be set; , , ; is the sign function, and its expression is: ； is the time-varying bounded attenuation coefficient based on topological feedback, and its expression is: , where 、 are constants to be set, , , , , represents the minimum eigenvalue of the obtained matrix, n represents the number of vehicles, represents the sampling time; Dynamic Constraint Controller for the TSSZNN-DC2 Discrete Model The expression of which is: ； Among them , , , are constants to be set; , , are step gain coefficients based on the error, and the expression is: , , , , , are constants to be set, represents the two-norm; as a whole represents a strongly bounded dynamic constraint function based on the error norm gain, where the error norm gain has the expression: , is represented as the one-norm, where represents the settable boundary, , , where: , is the settable upper bound of the acceleration, is the settable lower bound of the acceleration; is a non-linear manifold, and its expression is: 。 8. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 7, characterized in that, In step S7, the expressions of the first-order derivatives of the positions, velocities, and accelerations of the interconnected autonomous vehicles in the interconnected autonomous vehicle formation dynamic constraint control system equation solved by the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model are: 。 9. The interconnected autonomous vehicle formation dynamic constraint control method based on the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model according to claim 8, characterized in that, The specific method for solving the interconnected autonomous vehicle formation dynamic constraint control system equation by the TSSZNN-DC1 discrete model and the TSSZNN-DC2 discrete model is: Let ; Among them ; represents the iterative update value of is the first derivative of the acceleration of the virtual signal interconnected autonomous vehicle; the speed of the virtual signal interconnected autonomous vehicle; is expressed as , is expressed as , is expressed as ; Substitute into the TSSZNN-DC1 discrete model or the TSSZNN-DC2 discrete model to obtain the position, speed, and acceleration of each interconnected autonomous vehicle in the case of an undirected communication topology or a directed communication topology.

Images