Dual-objective intelligent vehicle lane changing trajectory planning based on polynomial optimization

Abstract

Listen

Translate article

With the diversification and deepening of intelligent technology research, intelligent vehicles are gradually involved in entering the transportation life. When intelligent vehicles enter the actual road driving, the success of changing lanes directly determines the rate of safety problems occurring when driving. Therefore, combining with the current background of advocating low-carbon travel, the relevant technology of intelligent vehicles is improved and supplemented in terms of trajectory planning and the degree of impact on the environment, and a trajectory optimization method is proposed that considers the traffic impact and low-carbon of the vehicle when changing lanes. The method combines a vehicle dynamics model with intelligent vehicle operating conditions, the corresponding positional relationship is used to analyze and establish the corresponding safety domain. At the same time, the quintic polynomial model in the ideal state is improved, and the hexagonal polynomial model in the longitudinal direction is established. For the lane change is difficult to get the completion time and end position of the lane change in the PSO-BP neural network to solve the problem, to get the intelligent vehicle lane change trajectory cluster a dual-objective performance evaluation function is established, optimization of evaluation results using the properties of genetic algorithms. Taking car No. 689 in the NGSIM data as an example, the established trajectory planning model and algorithm are used to solve the problem, and the final lane-changing trajectory equation is obtained. After the simulation of the obtained data, a more suitable trajectory curve is obtained when the vehicle is running in the lane change, which shows that the lane change trajectory optimization method can solve the lane change problem well. Then the study is mainly applied to the planning of trajectories required by intelligent vehicles when lane changing occurs on real roads. After knowing the parameters of the lane-changing vehicle before and after changing lanes, which can plan a safe and efficient lane change route for vehicles in real time. The two comprehensive indicators of sexuality reached a better value.