Research On Lateral Obstacle Avoidance Of Intelligent Vehicle Based On Lidar

With the rapid increase in car ownership and non-professional drivers and the intensive traffic flow,a series of problems such as frequent traffic congestion,increased traffic accidents,serious environmental pollution and energy consumption are caused.The study of intelligent driving vehicles(including advanced driving assistance systems and automatic driving)is exactly an important means to solve these problems.The intelligent vehicle system integrates the functions of environmental awareness,planning decision and control execution.The research on lateral obstacle avoidance of intelligent vehicle based on Lidar in this thesis was carried out according to these three parts.(1)A single-line Lidar was used as the main sensor to carry out the target motion state recognition research.Firstly,the accuracy of the obstacle detection was improved by using the distance and laser reflection intensity coupling algorithm.Then,an adaptive Kalman filter estimation algorithm based on the improved current statistical model was designed.Finally,the accuracy of the algorithm under multiple operating conditions was verified by simulation.(2)The problem of target association matching in target motion state recognition was studied.According to the characteristics of Lidar,a method of data matching between the measurement and the realistic target was proposed by constructing the"difference function" of distance,size and reflection intensity.By adjusting the weight of each weighting coefficient in the experiment,the two sets of weighting coefficients were determined to fit the distance between the target and Lidar.The accuracy of the target association matching in the range of 30m was 95%,and the accuracy rate was above 80%in the range of 30～60m.(3)The problem of lateral obstacle avoidance and motion control was studied.The obstacle avoidance path planning was completed through the polynomial trajectory function,and then the path tracking algorithm based on the preview driver model as well as the modification of steering by wire system and the corresponding angel close-loop control algorithm(based on the Fuzzy PID)were designed and accomplished.Finally,the simulation and experiment results showed that the lateral control on the obstacle avoidance had a good effect on path tracking accuracy and performed good transient and steady state performance in the steering response.