CN-121982893-A - Network-connected vehicle queue collaborative optimization safety control method and system

CN121982893ACN 121982893 ACN121982893 ACN 121982893ACN-121982893-A

Abstract

The invention discloses a network connection vehicle queue collaborative optimization safety control method and system. Firstly, a model of a pilot vehicle and following vehicle queue system is established, vehicle state information is acquired through a sensor, and a distributed communication time delay elastic observer is designed for possible abnormality of inter-vehicle communication links so as to estimate the pilot vehicle state. And the system makes a difference between the following vehicle state and the pilot vehicle real state or the observer state according to the judging result of the communication link state, and an error system is constructed. And deducing Hamiltonian quantity and an HJB equation based on the value function, and solving the optimal control law. And finally, respectively approximating a value function and a control law by utilizing the judgment and execution neural network in reinforcement learning, obtaining an optimized optimal control law through iterative calculation, and outputting the optimal control law to the following vehicle. The invention effectively improves the stability and the robustness of the system under malicious attack and communication abnormality while realizing the collaborative optimization control of the multiple vehicle queues.

Inventors

LI YONGMING
WANG MINGSHUO
LI KEWEN
Yu Kunting
DONG GUOWEI
WU WEI
Shao xinfeng
HU JUN
CHENG SHUAI

Assignees

辽宁工业大学

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (6)

1. The networking train queue collaborative optimization safety control method is characterized by comprising the following steps of: The method comprises the steps of establishing a pilot vehicle system model and a following multi-vehicle queue system model, acquiring a following vehicle state by combining data acquired by sensors carried by the following vehicle with the following multi-vehicle queue system model, and acquiring a pilot vehicle real state by combining data acquired by the sensors carried by the pilot vehicle with the pilot vehicle system model; Based on the pilot vehicle system model, designing a distributed communication time delay elastic observer, wherein the distributed communication time delay elastic observer is used for estimating the pilot vehicle state under an inter-vehicle communication link abnormal scene and outputting the observer state; If the communication link is abnormal, the following vehicle state is differed from the observer state to obtain an error system formed by an error state and an error dynamics equation, and the error state and the corresponding error system are output; According to the error state and the error system, firstly defining a value function of the error system, constructing a Hamiltonian volume based on the value function, deriving the Hamiltonian volume value as 0 to obtain an HJB equation, and then solving a bias derivative of the Hamiltonian volume with respect to a control input under the constraint of meeting the HJB equation to obtain an optimal control law and outputting the optimal control law; And taking the optimal control law as input, adopting a judging neural network and an executing neural network to respectively approximate the value function and the optimal control law, carrying out iterative computation to obtain the optimal control law after optimization, and outputting the optimal control law after optimization to the following vehicles in the following multi-vehicle queue system to realize the collaborative optimization control of the multi-vehicle queue.
2. The network connection queue collaborative optimization security control method according to claim 1, wherein the inter-vehicle communication link anomaly comprises two situations of communication delay and DoS attack.
3. The networked vehicle queue collaborative optimization safety control method according to claim 1, wherein the pilot vehicle system model is generated by the following equation: ; ; Wherein the variables are Is in the state of a pilot vehicle, Is the output of the pilot vehicle and is used for controlling the pilot vehicle, And A system matrix and an output matrix of the pilot vehicle system, respectively.
4. The network-connected-train-queue collaborative optimization safety control method according to claim 1, wherein the following multi-train-queue system model is described by the following differential equation: ; Wherein intermediate variables are defined 、 And Respectively the first Vehicle following vehicle Displacement, velocity and acceleration of moment, continuous function And The drift dynamics and input gains of the follower respectively, Is the system of Time of day (time) Control law and intermediate variable of vehicle following vehicle Wherein air density, vehicle cross-sectional area and air resistance coefficient are respectively used 、 And The representation is made of a combination of a first and a second color, Representing the engine constant and, Wherein Represents the rolling resistance and is used to control the rolling resistance, Representing the force of gravity and, In order to be a coefficient of rolling resistance, Is the acceleration of gravity and, It is the road gradient that is to be determined, Is the first The mass of the vehicle following the vehicle.
5. The network connection queue collaborative optimization security control method according to claim 1, wherein the form of the distributed communication delay elastic observer is as follows: ; ; Wherein, the Represents the first The state of the pilot vehicle observed by the vehicle following the vehicle, Represents the first The vehicle following the vehicle corresponds to the observer gain, Is present in the first The vehicle follows the communication delay between the vehicle and the pilot vehicle, Is present in the first Vehicle following vehicle and the first The vehicles follow the communication time delay between the vehicles, Is the first A safe state between the vehicle following vehicle and the pilot vehicle, Is the first A safe state between the vehicle following vehicle and the pilot vehicle, Is the pilot vehicle and the first The vehicles follow the contiguous elements between the vehicles, Is the first Vehicle following vehicle and the first The vehicles follow the contiguous elements between the vehicles, In order to follow the total number of vehicles, Represents the first The state of the pilot vehicle observed by the vehicle following the vehicle, The current time is indicated as such, Representing a state of the pilot vehicle subject to a communication delay, Is the first The pilot vehicle output signal observed by the vehicle following vehicle, The system matrix of the pilot vehicle system is C, and the output matrix of the pilot vehicle system is C.
6. The networking train queue collaborative optimization safety control system is characterized by comprising: The system comprises a vehicle queue system, a pilot vehicle real state and a vehicle queue system, wherein the vehicle queue system is used for establishing a pilot vehicle system model and a following multi-vehicle queue system model; The distributed observer module is used for designing a distributed communication time delay elastic observer based on the pilot vehicle system model, wherein the distributed communication time delay elastic observer is used for estimating the pilot vehicle state under the abnormal scene of an inter-vehicle communication link and outputting the observer state; The system conversion module is used for judging the state of an inter-vehicle communication link, and if the communication link is normal, making a difference between the state of the following vehicle and the real state of the pilot vehicle; the optimal control law calculation module is used for defining a value function of the error system according to the error state and the error system, constructing Hamiltonian quantity based on the value function, deriving the Hamiltonian quantity value as 0 to obtain an HJB equation, and solving a bias derivative of the Hamiltonian quantity about control input under the constraint of meeting the HJB equation to obtain an optimal control law and outputting the optimal control law; And the reinforcement learning module takes the optimal control law as input, adopts a judging neural network and an executing neural network to respectively approximate the value function and the optimal control law, and performs iterative computation to obtain an optimized optimal control law, and outputs the optimized optimal control law to the following vehicles in the following multi-vehicle queue system so as to realize the collaborative optimization control of the multi-vehicle queue.

Description

Network-connected vehicle queue collaborative optimization safety control method and system Technical Field The invention belongs to the technical field of network-connected multi-vehicle cooperative formation control, and particularly relates to a network-connected vehicle cooperative optimization safety control method and system. Background With the rapid development of intelligent traffic systems, a multi-vehicle queue system is a core component of the intelligent traffic system, and is attracting attention because of remarkable advantages in the aspects of improving road traffic efficiency and safety, improving resource utilization rate and the like. However, the rapid increase of the quantity of the motor vehicles is beneficial to the travel of people, and a series of social problems such as frequent traffic accidents, aggravation of road congestion, environmental pollution and the like are also derived. The cooperative control of the distributed multi-vehicle queue system is highly dependent on real-time information interaction of the internet of vehicles, however, in a complex dynamic network environment, the inherent transmission delay characteristic of the internet of vehicles communication link and malicious attack (such as denial of service attack) will seriously destroy the real-time property and consistency necessary for cooperative control of the multi-vehicle queue. In particular, communication delays often result in unsynchronized updates of state information between vehicles, which can cause performance degradation of the queue controller or even create a safety hazard. Denial of service attacks (Denial of Service, doS attacks) directly result in the failure of conventional cooperative adaptive cruise control algorithms based on continuous communication assumptions by periodically or randomly interrupting critical data transmissions. Therefore, under the mixed model of simultaneously considering the coexistence of communication delay and DoS attack, how to design a controller with a distributed architecture, attack resistance robustness and control optimality simultaneously so as to ensure the stability, safety interval maintenance and energy consumption minimization of a multi-vehicle queue under malicious attack, which is a key scientific problem to be solved in the field of intelligent transportation system network safety and cooperative control. At present, research on the cooperative control problem of a multi-vehicle queue system has attracted attention of students in the fields of control theory and control engineering, and a series of important research results are obtained on the aspects of theoretical research and technical realization. In the field of longitudinal motion control, the multi-vehicle-queue system can automatically adjust the opening degree and the braking strength of the accelerator of the vehicle, so that the automatic control capability of the vehicle in the longitudinal direction is improved, and traffic accidents caused by fatigue driving caused by long-time operation of the vehicle by a driver are reduced. Currently, there are many control techniques for multi-vehicle queuing systems. However, the existing control method still has the following problems: 1. existing research results are all above the ideal assumption that instant messaging is not delayed and that inter-vehicle communication links are continuously stable. However, in the actual operation scene of the automatic driving multi-vehicle formation, the system stability is affected and even potential safety hazards are generated due to the unexpected conditions such as inter-vehicle communication link interruption or communication delay; 2. most of the existing control strategies are designed only by considering the control target and the system stability, but huge energy consumption is usually caused. Therefore, how to apply the optimal control method to the multi-vehicle queue system, and achieving the minimization of the energy consumption while ensuring the queue control target is a key problem to be solved by the present study. Disclosure of Invention The invention aims to provide a network connection queue collaborative optimization safety control method and system aiming at the technical problems that the stability of the existing control method of a multi-vehicle queue is poor due to inter-vehicle communication interruption or communication delay and the energy consumption is huge. In order to achieve the technical purpose, the following technical scheme is adopted in the embodiment of the application. In a first aspect, an embodiment of the present application provides a network link queue collaborative optimization security control method, including: The method comprises the steps of establishing a pilot vehicle system model and a following multi-vehicle queue system model, acquiring a following vehicle state by combining data acquired by sensors carried by the following vehicle with the foll