Evaluation of Lane Change Trajectory Based on Group Decision and Entropy Weight Method

LI Zhenlong,DONG Aihua,YANG Lei

DOI: https://doi.org/10.3969/j.issn.1674-0696.2023.01.17

2023-01-01

Abstract:Aiming at the information overlap between indicators in the lane change trajectory evaluation and the subjective problem in the setting of indicator weights, a method for selecting evaluation indicators based on correlation analysis and factor analysis was proposed. Firstly, 11 evaluation indicators that characterized comfort, efficiency, smoothness, ecology and safety were selected from 16 parameters. Then, analytic hierarchy process was used to calculate subjective weights of indicators, group decision theory was used to calculate expert weights, and entropy weight method was used to calculate objective weights of indicators. Therefore, an index weight setting method combining analytic hierarchy process, group decision theory and entropy weight method was proposed. The weights of safety, efficiency, comfort, ecology and smoothness were calculated to be 0.73, 0.15, 0.06, 0.04 and 0.02 respectively. Moreover, the lane change trajectory evaluation system was established and used to evaluate the lane change trajectory obtained by driving simulation. The results show that the proposed lane change trajectory evaluation system is relatively comprehensive and reduces the overlap of information between indicators, which combines subjective expert experience and objective data information to make the evaluation results more reasonable. Sorting the importance of trajectory performance can provide a reference for the tendency to focus when setting constraints in trajectory planning; the constructed lane change trajectory evaluation system and the proposed evaluation method provide a reference for lane change trajectory evaluation.