Research On Control Strategy Of Multi-vehicle Cooperative Driving On The Highway

The transition from autonomous vehicle intelligence to vehicle collaborative multi-agents is a huge advantage of vehicle networks and intelligent transportation.In order to reduce road congestion,improve fuel efficiency,and enhance driving safety,it is necessary to maintain the ability of vehicles to communicate with each other and exchange information in real time.From the perspective of development trends,highways have the characteristics of network and intelligence.The road structure is simpler and the traffic capacity is strong,which is suitable for the early implementation of multi-agent technology.Transportation is the forerunner of economic development,and new technological changes on the highway are in the ascendant.This article uses the highway as a research scenario,and studies multi-vehicle cooperative driving.The core algorithms of the system,the control of actuators,and test and verification methods are discussed.The following are the main contents:(i).CACC collaborative adaptive cruiseCACC is the most basic module in platoon driving.The interaction between intelligent driving vehicles is mainly controlled by the car formation.Information sharing is a key task in multi-agent systems.The article analyzes the formation rules of several topological structures,and finally identifies the "pilot-leading vehicle" as the best communication topology as the basis of CACC research.By adopting the linear superposition of speed and distance,the shortcomings of single speed or distance tracking effect are overcome.Stability is one of the important indicators to measure the performance of the queue.When the system is disturbed,it should converge as soon as possible,and not diverge with the expansion of the queue size.In order to explore the queue’s driving effect under ideal conditions,communication delay,and actuator delay,the control law in the time domain is converted to the complex domain to determine the stability and errors under steady state.When the system cannot always be stable,calculate the parameter range when it can be stabilized.Then define the design queue size and communication method.(ii).Collaborative Collision AvoidanceVehicle collision risk identification and early warning is one of the key technologies of intelligent driving systems.By taking the front and rear vehicle distance as the boundary research scope,the concepts of early warning distance and minimum early warning distance were proposed by comparing the advantages and disadvantages of TTC,THW,and RF indicators.According to the current risk value,it is further subdivided into front-end collision avoidance and rear-end collision avoidance.For rear-end collision avoidance,a strategy TRM based on the total relative kinetic energy of the queue is designed,which takes the mass and speed difference as the direct control objects and the vehicle number as the weight coefficient.The advantage of this method is that the braking distance and response time are fully utilized,and the control target is no longer a single individual but the entire system.The purpose of the system is to make the objective function reach the minimum value at each sampling point.When the degree of danger further increases and the current conditions of the host vehicle are not sufficient for safe braking,a front-end collision avoidance strategy is adopted.That is,the host vehicle assigns a coordinated acceleration to the preceding vehicle,and a calculation formula for the coordinated acceleration is designed.(iii).Research on Control Strategy of Multi-vehicle Cooperative Lane Change Based on MPCDifferent from the first two conditions,multi-vehicle cooperative lane change involves longitudinal and horizontal coupling control.Firstly,the synchronous lane change mode was determined,which is shorter than the lane change time of the distributed lane change.It is more suitable for a simple structure such as a highway.According to the time sequence,the entire process is divided into a preparation phase,a start phase,a lane change,a lane change end,and an adjustment phase.The cause of the lane change motivation and the switching logic are analyzed.And a two-lane cooperative lane change strategy is proposed.A kinematics analysis was performed for each vehicle to solve the equations of longitudinal and lateral motion,so as to satisfy the safety conditions.Next,the linear time-varying MPC strategy is used to implement trajectory tracking.Compared with the optimal preview algorithm,MPC has better simulation performance.Finally,the safety,comfort,stability and lane changing efficiency of the decision and algorithm based on the queue are verified.