Research On Decision-making Planning And Control Of Intelligent Connected Vehicles For Energy-saving Drivin

Intelligent connected vehicles(ICVs)have significant application value in energy conservation,traffic safety,and transportation efficiency.As a strong dynamic and highreal-time system,ICVs face complex traffic scenarios and variable driving tasks.The primary challenge for this technology application lies in how to achieve energy-efficient,safe,and comfortable driving of ICVs in complex scenarios.This paper focuses on the systematic research of behavior decision-making,trajectory planning,and tracking control algorithms for ICVs.Furthermore,the feasibility and efficacy of the proposed decisionmaking,planning,and control algorithms are validated through joint simulation.The main research content and innovation points include:1)A coupled framework of energy-efficient driving and real-time vehicle planning and control is proposed.The framework utilizes high-precision map information from the Carla to construct a road map,and employs the A* algorithm to obtain a globally optimized path as input for global speed planning.The speed planning problem is transformed into an optimal control problem with the minimum vehicle energy consumption as the optimization objective,and a robust three-stage energy-efficient speed planning algorithm is proposed.The energy-saving effect of the speed planning algorithm is verified through joint simulation of Carla and ROS.The simulation results demonstrate that compared to the algorithms of Dynamic Programming Method(DPM)and Modified Intelligent Driver Model(M-IDM),the proposed energy-saving speed planning method significantly reduces computational costs and improves energy utilization by 32%,while also demonstrating robustness,energy efficiency,and high efficiency.2)A hierarchical decision-making framework based on a layered finite state machine is proposed for driving tasks in complex scenarios.The upper level analyzes the decisionmaking mechanism between overtaking and following states,while the lower level analyzes the internal stage decision-making mechanism within the overtaking state.The PID feedback control algorithm is employed for speed decision-making in the following state.To address the issues of non-holonomic constraints and the single form of collision avoidance responsibility allocation in the Optimal Reciprocal Collision Avoidance(ORCA)algorithm,a modified ORCA(MORCA)algorithm is proposed,which optimizes the preference speed and collision avoidance responsibility allocation and is applied to the lanechanging and overtaking scenario.Simulation results show that the proposed hierarchical decision-making framework can make efficient and reasonable behavior decisions.The speed decision-making based on the PID feedback algorithm can ensure that ICVs maintain a reasonable safety distance from the preceding vehicle during the driving process.The average time required for MORCA in the overtaking task is reduced by 15% compared with ORCA.In addition,the collision avoidance trajectory generated by MORCA is more accurate and safer than that of ORCA.3)A comprehensive control strategy combining pure tracking and active disturbance rejection control(ADRC)is proposed.Pure Pursuit algorithm is adopted for lateral trajectory tracking,while ADRC is utilized for longitudinal speed tracking.Simulation results show that the pure pursuit controller achieves excellent lateral trajectory tracking with a root mean square error(RMSE)of 0.0354 m at vehicle turning.For longitudinal speed tracking,the ADRC algorithm is able to track the speed reference trajectory effectively,and achieves an RMSE of 0.46 m,which is 21% better than the RMSE of 0.58 m achieved by the PID algorithm.The ADRC can effectively follow the speed reference trajectory,allowing the vehicle to approach the energy-efficient driving speed.4)A simulation platform for intelligent driving is built by integrating Carla and ROS to simulate realistic driving scenarios.The behavior decision-making algorithm,trajectory planning algorithm,and trajectory tracking control algorithm for intelligent connected vehicles are developed using the C++ programming language.The feasibility and effectiveness of the proposed algorithms are verified in the simulation environment,showcasing their potential for deployment in real vehicles.