摘要
针对自动驾驶汽车在转向时为防止横向偏差和角度偏差数据超出物理论域而导致决策错误,在模糊控制系统中设计了横向偏差和角度偏差量化因子模块.模糊控制系统中采用了阿克曼转向原理的汽车运动学模型,实时获取自动驾驶汽车坐标和航向角用于计算汽车基于路点的横向偏差和角度偏差,偏差通过量化因子作用,限定在了模糊论域之中,避免了在一些特殊路段由于数据超出论域而导致决策错误的情况,通过Car Sim和Simulink联合仿真的结果表明:在25 km/h以下的车速下,通过量化因子模块的作用,模糊控制器能够正常的输出转角。
In order to prevent the lateral deviation and angular deviation data beyond the domain of discourse and lead to poli-cy mistakes when the self-driving cars turning, the lateral deviation and angular deviation quantitation factor module are de-signed in the fuzzy control system. Fuzzy control system adopts the principle of Ackerman steering vehicle kinematic model,which could acquire the real-time coordinate and course angle of self-driving cars so as to calculate the lateral deviation andangular deviation based on the road point. By means of the quantitation factor, the deviation is limited to the fuzzy domain, soas to avoid some special sections that the controller would make wrong decision due to data beyond the domain of discourse.Through the joint simulation results of CarSim and Simulink, it is shown that under the speed of 25 km/ h, the fuzzy controllercan output angle normally with function of the quantitation factor module.
出处
《机械研究与应用》
2017年第2期56-57,61,共3页
Mechanical Research & Application
