Decision-making And Local Path Planning Of Intelligent Car Based On Driver Behavior

As the human material demand is gradually increasing,intelligent cars are gaining more and more attention.The three key technologies of the car’s environmental perception,decision planning and tracking control jointly guarantee the autonomous driving of intelligent cars.As a traffic environment with simple working conditions,the highway has a simple road structure and no unpredictable traffic conditions.It is one of the easiest road environments to complete automatic driving.As one of the three key technologies,decision-making planning of intelligent cars determines the car’s lane-changing behavior and driving path to ensure the safety,comfort and legality of driving.At present,for the decision planning of smart cars,most scholars rarely consider behavioral decisions when studying path planning,which may cause the planned path to be inconsistent with human driving characteristics.This paper designs a set of decisions that simultaneously consider lane change decisions,path planning,and speed planning algorithms to ensure that the driving trajectory of intelligent cars meet human driving habits.The specific research contents are as follows:This article firstly found through a questionnaire survey that the main reason for the driver’s lane change is the presence of obstacles in front of the lane,and the relative speed and relative distance of the vehicle in the current lane and the target lane are mainly considered when changing lanes,and the driving simulator is analyzed.Change of behavior.In order to accurately judge the car’s behavior decision under different working conditions,a lane-changing decision algorithm is designed based on the idea of decision tree,which quantifies the satisfaction of car driving,and the lane-changing behavior can be quickly determined by judging the satisfaction of different lanes.Secondly,the path planning algorithm is designed based on the sampling method,and the traditional sampling method is improved.The driver’s lane change time,speed and acceleration are used to determine multiple longitudinal distances.Compared with the previous fixed sampling distance,the algorithm has stronger robustness.Dangerous potential field function is introduced to judge the collision risk of vehicles and obstacles,which can effectively analyze the impact of different obstacles on the safety of the vehicle.Then,based on the discrete optimization method,the optimal speed is solved in the ST space,which can avoid the problem that the speed planning solution space of the car is a non-convex space under certain working conditions.In ST space,the discrete speeds that meet the constraints are sampled,and the optimal speed is solved by the cost function to complete the speed planning.Finally,this paper establishes the overall framework of the decision planning algorithm,and simulates and verifies the designed algorithm on the MATLAB platform.It can achieve good decision planning effects under different working conditions,proving the rationality and reliability of the algorithm.