基于MPC的智能車輛換道重規(guī)劃研究

【摘要】為解決智能車輛換道過程中，由于其他車輛的行駛狀態(tài)發(fā)生改變導致碰撞風險的問題，需對主車進行實時局部路徑規(guī)劃。基于模型預(yù)測控制（MPC）提出換道軌跡重規(guī)劃策略，根據(jù)碰撞風險將其劃分為換道軌跡修正策略、換道折返策略及前向主動避撞策略，建立基于MPC的橫向控制和基于雙比例積分微分控制器（PID）的縱向控制，并以縱向車速為聯(lián)合點的橫縱向綜合軌跡跟蹤控制器。將軌跡規(guī)劃模塊、軌跡重規(guī)劃模塊及軌跡跟蹤模塊分層集成，對換道軌跡重規(guī)劃策略進行仿真驗證。仿真結(jié)果表明：基于MPC的換道軌跡重規(guī)劃和軌跡跟蹤控制可在不同場景下實現(xiàn)車輛安全避撞。

主題詞：智能車輛 換道重規(guī)劃 模型預(yù)測控制 軌跡跟蹤

中圖分類號：U463.3" "文獻標志碼：A" "DOI： 10.19620/j.cnki.1000-3703.20240704

Study on Lane-Change Replanning for Intelligent Vehicle Based

on Model Predictive Control

Li Yaohua1， Chong Guochen1， Fan Jikang1， Zhai Dengwang1， Lan Fenlong2

（1. School of Automotive， Chang’an University， Xi’an 710064; 2. College of Mechanical and Vehicle Engineering， Chongqing University， Chongqing 400044）

【Abstract】To address the issue of collision risk during lane changing of intelligent vehicles caused by changes in the driving status of other vehicles， real-time local path planning is required for the host vehicle. Based on Model Predictive Control （MPC）， a lane change trajectory replanning strategy is proposed， which is divided into lane change trajectory correction strategy， lane change reentry strategy， and forward active collision avoidance strategy according to the collision risk. A lateral control based on MPC and a longitudinal control based on dual PID are established， and a comprehensive lateral-longitudinal trajectory tracking controller is designed with the longitudinal speed as the joint point. The trajectory planning module， trajectory replanning module， and trajectory tracking module are integrated in layers， and simulations are conducted to verify the lane change trajectory replanning strategies. The simulation results show that the lane change trajectory replanning and trajectory tracking control based on MPC can achieve safe collision avoidance for vehicles in different scenarios.

Key words： Intelligent vehicle， Lane-change replanning， Model predictive control， Trajectory tracking

【引用格式】 李耀華， 種國臣， 范吉康， 等. 基于MPC的智能車輛換道重規(guī)劃研究[J]. 汽車技術(shù)， 2024（10）： 20-27.

LI Y H， CHONG G C， FAN J K， et al. Study on Lane-Change Replanning for Intelligent Vehicle Based on Model Predictive Control[J]. Automobile Technology， 2024（10）： 20-27.

1 前言

換道軌跡規(guī)劃要求智能車輛根據(jù)周圍的交通環(huán)境及自車運動狀態(tài)，規(guī)劃出從換道起始點到換道目標點的安全無碰撞路徑，并且要兼顧換道效率和換道舒適性[1]。文獻[2]提出基于碰撞風險評估的局部路徑規(guī)劃算法，在多種道路場景中實現(xiàn)對靜態(tài)和移動障礙物的有效規(guī)避，但未對規(guī)劃軌跡進行跟蹤控制研究。文獻[3]提出基于代價函數(shù)的局部路徑規(guī)劃算法，使用一元三次方程對路徑進行曲線擬合，未考慮實際場景中其他車輛行駛狀態(tài)改變的情形。文獻[4]僅考慮跟車和換道兩種策略，提出具有兩層結(jié)構(gòu)的協(xié)同控制策略，建立了基于五次多項式換道軌跡模型和固定車頭時距跟車模型。……