CN-122018551-A - Automatic driving fleet transverse dynamics inclusion control method based on output feedback

CN122018551ACN 122018551 ACN122018551 ACN 122018551ACN-122018551-A

Abstract

The invention relates to an automatic driving fleet transverse dynamics inclusion control method based on output feedback, which comprises the steps of constructing a transverse dynamics model of an automatic driving vehicle and abstracting the model into a leading following multi-agent system, constructing a communication topological graph, designing an open-loop reference model and a closed-loop reference model for a follower agent, constructing an open-loop reference model inclusion controller, solving a reference control input of the follower agent, enabling the open-loop reference model to be converged into a convex hull formed by a system state of the leading vehicle, constructing a decentralization model reference self-adaptive output controller, updating a gain matrix, solving a control input of the follower vehicle, enabling the follower vehicle to track the open-loop reference model, and realizing inclusion control of the automatic driving vehicle. The invention can still ensure the stability of the closed loop system when uncertain system parameters, unknown input gains and directed communication topology exist, realize the inclusion control target of multiple intelligent agents and improve the engineering realizability and robustness of the method.

Inventors

YUE DONGDONG
LI YUNLONG
SHI JIANTAO
CHEN CHUANG
BAO DAN

Assignees

南京工业大学

Dates

Publication Date: 20260512
Application Date: 20260320

Claims (10)

1. An automatic driving fleet lateral dynamics inclusion control method based on output feedback, comprising: constructing a transverse dynamics model of the automatic driving vehicle; Abstracting a transverse dynamics model of the automatic driving vehicle into a leader following multi-agent system by taking a leader vehicle and a follower vehicle as agents, wherein the leader following multi-agent system comprises a follower vehicle dynamics model and a leader vehicle dynamics model; Constructing a communication topological graph based on the leader following multi-intelligent system by taking intelligent agents as nodes and communication relations among the intelligent agents as edges; Designing open-loop reference models for follower intelligent agents, updating the system state and control output of the open-loop reference models of the follower intelligent agents, and calculating the control output errors of the open-loop reference models; introducing an observer gain matrix to design a closed-loop reference model for the follower intelligent agent, solving the estimated control output and the estimated system state of the follower intelligent agent, and calculating the control output error of the closed-loop reference model; Constructing a follower state estimator, and updating the estimated state of each follower intelligent agent by combining the control output of the open-loop reference model; constructing a leader state estimator, and updating the estimated state of each leader intelligent agent; The method comprises the steps of constructing an open-loop reference model, wherein the open-loop reference model comprises a controller, and calculating reference control input of a follower agent by combining a communication topological graph, estimated states of the leader agent and the follower agent, so that the open-loop reference model is converged into a convex hull which is formed by the system states of a leader vehicle; The method comprises the steps of constructing a de-centralized model reference self-adaptive output controller, based on a closed-loop reference model control output error and an open-loop reference model control output error, combining a reference control input of a follower intelligent body, solving a control input of a follower vehicle, enabling the follower vehicle to track an open-loop reference model, and realizing inclusion control of an automatic driving vehicle.
2. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 1, characterized in that: the model of the lateral dynamics of the autonomous vehicle is described by the following equation: ; Wherein, the For the lateral offset of the vehicle relative to the lane centerline, As the heading angle error of the vehicle, 、 Respectively is 、 Respectively describe first order time derivatives of (a) 、 Dynamic changes over time; 、、 Respectively the cornering stiffness, the mass and the longitudinal speed of the front wheel of the vehicle, Representing the unknown input gain, as a diagonal matrix with positive elements, Is the control input to be designed.
3. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 1, characterized in that: The follower vehicle dynamics model is described by the following equation: ; Wherein, the The representation number is Including lateral offset and heading angle errors of the follower vehicle, Is that Description of first order time derivative of (2) Dynamic changes over time; Is a transposition operation; a system matrix for the follower vehicle; for unknown input gain, for diagonal matrix with positive elements; 、 respectively an input matrix and an output matrix; 、 Respectively denoted by the number Control input and control output of the follower vehicle; the leader vehicle dynamics model is described by the following equation: ; Wherein, the The representation number is The system status of the leader vehicle, including the lateral offset and heading angle error of the leader vehicle, Is that Description of first order time derivative of (2) Dynamic changes over time; A system matrix for the leader vehicle; 、 Respectively denoted by the number Control input, control output of the leader vehicle, and Is bounded.
4. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 3, characterized in that: the leader follows the multi-agent system, triplets Is controllable and considerable, and the transfer function Is strictly true, wherein Representing the identity matrix of the cell, Representation of Is used for the inverse matrix of (a), Complex frequency domain variables representing the laplace transform; Is a transposition operation; The leader follows the multi-agent system, there is an ideal matching gain So that And (2) and , Is a set of known uncertainties.
5. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 1, characterized in that: The communication topology is a non-directional connection between the follower agents, and for each follower agent, there is at least one leader agent that is able to reach the follower agent via a directional path.
6. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 1, characterized in that: the designing of the open-loop reference model for the follower agents includes designing an open-loop reference model for each follower agent, described by the following equation: ; Wherein, the Representing follower agent Is provided for the system state of the open loop reference model, Representing follower agent Is provided with a control output of an open loop reference model, Representing follower agent Is provided with a reference control input to the control system, For a system matrix of leader vehicles, In order to input the matrix of the data, Is an output matrix; Is that Description of first order time derivative of (2) The method comprises the steps of designing a closed-loop reference model corresponding to the open-loop reference model for each follower agent, wherein the closed-loop reference model is described by the following equation: ; Wherein, the Representing follower agent Is provided for the estimation of the state of the system, Representing follower agent Is provided with an estimated control output of (c), Is that Description of first order time derivative of (2) Dynamic changes over time; Representing an observer gain matrix; the representation number is Control outputs of follower vehicles, i.e. follower agents A control output of (2); the open loop reference model control output error calculation mode is as follows: Wherein Representing follower agent An open loop reference model control output error; The calculation mode of the closed loop reference model control output error is as follows: Wherein Representing follower agent The closed loop reference model of (c) controls the output error.
7. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 1, characterized in that: the follower state estimator is described by the following equation: ; Wherein, the Representing follower agent Is used to determine the estimated state of (1), Is that Description of first order time derivative of (2) Dynamic changes over time; for a system matrix of leader vehicles, In order to input the matrix of the data, Is an output matrix; Representing follower agent Reference control input of (a); Is a transposition operation; Representing follower agent Control output of an open loop reference model; The leader state estimator is described by the following equation: ; Wherein, the Representing leader agents Is used to determine the estimated state of (1), Is that Description of first order time derivative of (2) The dynamic change over time is that, Representing leader agents Is provided with a control output of (a), Representing leader agents Is used for the input of the (c) to be processed, Representing an output error feedback gain matrix; The open loop reference model includes a controller described by the following equation: ; Wherein, the Representing a feedback gain matrix that linearly weights and maps neighbor differential information, Representing the coupling gain factor that adjusts the strength of action of the linear differential feedback term, Representing follower agents in a communication topology adjacency matrix With the leader agent Is a connection weight of (2); representing follower agents in a communication topology adjacency matrix With follower agent Is a connection weight of (2); Representing an adjusting nonlinear function Coupling gain coefficient corresponding to action intensity of feedback term, and , ; The total number of vehicles of the follower, namely the total number of the intelligent agents of the follower; As a nonlinear function to Nonlinear function as an argument The definition is as follows: ; Wherein, the Is a constant, used to characterize the width of the boundary layer; Representing euclidean norm calculations.
8. The method for controlling the lateral dynamics of an automatic driving fleet based on output feedback as set forth in claim 6, wherein the constructing the de-centralized model reference adaptive output controller, based on the closed-loop reference model control output error and the open-loop reference model control output error, updates the gain matrix in combination with the reference control input of the follower agent and solves the control input of the follower vehicle comprises: Designing a decentered model for a follower vehicle dynamics model with reference to an adaptive output controller is described by the following equation: ; Wherein, the The representation number is Is provided with a control input of the follower vehicle, Indicating the ideal state of the feedback gain, Indicating the ideal coupling gain is to be achieved, Representing preselected adjustments The adaptive adjustment matrix of the velocity is updated, Representing preselected adjustments The adaptive adjustment matrix of the velocity is updated, And Respectively are intelligent agents For ideal matching gain And Is used for the estimation of (a), Representing unknown input gain Is used for the inverse matrix of (a), The time of day is indicated as such, Is that Description of first order time derivative of (2) The dynamic change over time is that, Is that Description of first order time derivative of (2) The dynamic change over time is that, Is a transposition operation; Representing follower agent An open loop reference model control output error; Representing follower agent The closed loop reference model of (c) controls the output error.
9. The output feedback-based automatic fleet lateral dynamics inclusion control method according to claim 7, characterized in that: the observer gain matrix Designed as Wherein, the method comprises the steps of, A positive scalar adjustment coefficient representing observer gain, an 。
10. An output feedback based autopilot fleet lateral dynamics inclusion control system for applying the output feedback based autopilot fleet lateral dynamics inclusion control method of any one of claims 1-9, comprising: the dynamics model building module is used for building a transverse dynamics model of the automatic driving vehicle and abstracting the transverse dynamics model into a leading following multi-agent system; the reference model construction module is used for designing an open-loop reference model for the follower intelligent agent and introducing an observer gain matrix to design a closed-loop reference model; the reference model controller module is used for constructing an open-loop reference model comprising a controller, combining a communication topological graph, the estimated states of a leader intelligent agent and a follower intelligent agent, resolving a reference control input of the follower intelligent agent and sending the reference control input to the vehicle output controller module; The vehicle output controller module is used for constructing a decentralization model reference self-adaptive output controller, controlling output errors based on a closed-loop reference model and an open-loop reference model, updating a gain matrix and resolving control input of a follower vehicle by combining reference control input of a follower agent from the reference model controller module, and sending the control input to the vehicle containing control module; the vehicle comprises a control module for receiving control input of the follower vehicle from the vehicle output controller module and realizing the inclusion control of the automatic driving vehicle.

Description

Automatic driving fleet transverse dynamics inclusion control method based on output feedback Technical Field The invention relates to the technical field of unmanned cluster control, in particular to an automatic driving fleet transverse dynamics inclusion control method based on output feedback. Background In the applications of automatic driving fleet cooperative control, unmanned system formation, group robots and the like, dynamics of multiple vehicles/machines are usually abstracted into a multi-agent system, and information is exchanged through a communication network to realize cooperative behavior, wherein 'including control' is used as a leader-following an important branch of cooperative control, and the state or output of a follower is usually required to be finally converged and enter a convex hull formed by a plurality of leader tracks, so that the goals of fleet shape maintenance, queue safety constraint, cooperative maneuver and the like are realized. Taking the automatic driving vehicle transverse dynamics as an example, the states of the vehicle transverse offset, the course angle error and the like can form a follower system state vector, a leader can correspond to a piloted vehicle (or a plurality of key vehicles) in a motorcade, and the states of the leader form a reference boundary for the motorcade cooperative control. Under the communication topology constraint, each follower does not need to precisely track a specific pilot vehicle, and states such as self transverse offset, course angle and the like gradually enter and are kept in a convex hull safety domain formed by a plurality of leader tracks through local information interaction, so that the formation keeping and safety constraint of keeping up and not crossing the boundary are realized. It is noted that under the influence of factors such as road adhesion coefficient change, speed measurement error, load change and the like, the model often has problems such as parameter uncertainty and even unknown input gain, and certain challenges are brought to the transverse dynamics of a motorcade including control problems. In the existing research, various distributed protocols and output feedback methods are proposed for control and consistency control, but a few schemes depend on known system dynamics and measurable full state, when the communication topology is a directed graph and information is limited, the methods often need additional global topology information, conservative prior estimation or more complex hierarchical structures, and engineering realization cost and parameter setting difficulty are obviously increased. Meanwhile, under the scene without leaders or multiple leaders, the complexity of closed loop analysis and stability demonstration is amplified by the system uncertainty propagation through network coupling, so that the applicability and robustness of the existing method under the superposition condition of directional communication, output availability and parameter uncertainty still have defects, and the further research on a control method which is more convenient for decentralization realization and is more robust to the uncertainty is necessary. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an automatic driving fleet transverse dynamics inclusion control method based on output feedback, which solves the problems of insufficient applicability, robustness and the like caused by the fact that the traditional method depends on the conditions of known system dynamics and measurable full state and the accumulation and superposition of system uncertainty. The invention designs a decentralization control and parameter self-adaptive updating mechanism, so that when uncertain system parameters, unknown input gains and directed communication topology exist in the system, the system can still ensure the stability of a closed-loop system, realize the inclusion control target of multiple intelligent agents, and simultaneously avoid the strong dependence on the whole state measurability, thereby improving the engineering realizability and robustness of the method. In order to achieve the technical aim, the invention provides the following technical scheme that the automatic driving fleet transverse dynamics inclusion control method based on output feedback comprises the following steps: constructing a transverse dynamics model of the automatic driving vehicle; Abstracting a transverse dynamics model of the automatic driving vehicle into a leader following multi-agent system by taking a leader vehicle and a follower vehicle as agents, wherein the leader following multi-agent system comprises a follower vehicle dynamics model and a leader vehicle dynamics model; Constructing a communication topological graph based on the leader following multi-intelligent system by taking intelligent agents as nodes and communication relations among the intelligent agents as edges; Designing open-l