An evaluation of lane changing process based on cloud model and incentive-punishment variable weights

Abstract

Aiming at the problems of making fuzziness precise and fixed weight in the fuzzy comprehensive evaluation of lane changing process, a lane change evaluation method based on cloud model and incentive-punishment variable weights was proposed. Based on a driving simulator, the highway driving simulation experiment was conducted to obtain the vehicle's parameters in the lane changing process. Five single index evaluation cloud models of safety, smoothness, comfort, efficiency and ecology in the lane changing process were established. The average membership degree of each index belonging to the four levels of excellent, good, fair and poor was obtained. Considering the subjectivity of experts and the objectivity of data samples, the combination weights were obtained. According to the balance and prominence of the indexes, the combination weights were appropriately rewarded and punished to obtain the final weights. Finally, the evaluation method based on the cloud model and incentive-punishment variable weights was used to evaluate 35 lane change processes, and the proposed method was compared with the method based on the constant weights. There are 32 processes for which the results of the two evaluation methods are the same, while there are 3 processes for which the results of the two evaluation methods are different. Through the further analysis of the indexes, it is found that the evaluation results of incentive-punishment variable weights are more reasonable and accurate.