自動運転システムにおける人間要因の評価と制御モデル設計
Evaluation and Control Model Design of Human Factors for Autonomous Driving Systems
乗員の心理的な安心感を考慮し、機械学習で主観評価モデルを構築して人工ポテンシャル場とモデル予測制御による障害物回避の経路計画・追従アルゴリズムを設計した研究。
著者: Weishun Deng, Fan Yu, Zhe Wang, Dengbo He
分類: cs.RO
原文アブストラクト
With the fast development of driving automation technologies, user psychological acceptance of driving automation has become one of the major obstacles to the adoption of the driving automation technology. The most basic function of a passenger car is to transport passengers or drivers to their destinations safely and comfortably. Thus, the design of the driving automation should not just guarantee the safety of vehicle operation but also ensure occupant subjective level of comfort. Hence this paper proposes a local path planning algorithm for obstacle avoidance with occupant subjective feelings considered. Firstly, turning and obstacle avoidance conditions are designed, and four classifiers in machine learning are used to respectively establish subjective and objective evaluation models that link the objective vehicle dynamics parameters and occupant subjective confidence. Then, two potential fields are established based on the artificial potential field, reflecting the psychological feeling of drivers on obstacles and road boundaries. Accordingly, a path planning algorithm and a path tracking algorithm are designed respectively based on model predictive control, and the psychological safety boundary and the optimal classifier are used as part of cost functions. Finally, co-simulations of MATLAB/Simulink and CarSim are carried out. The results confirm the effectiveness of the proposed control algorithm, which can avoid obstacles satisfactorily and improve the psychological feeling of occupants effectively.
