基于行为融合的无人自动驾驶车辆的智能控制

Action Amalgamation-based Intelligent Control of Automatic Driving For Unmanned Vehicles

针对动态环境下无人自动驾驶车辆控制的非线性、时变的特点,提出并设计了一种基于行为融合的无人驾驶车辆的智能控制策略。根据车辆行驶基于模糊逻辑方法设计了一系列的基本行为模式,用模糊控制的方法分别建立各行为模式控制器,进而对车辆的方向和速度进行控制。在行为选择机制设计中,对常用的行为竞争和行为融合2种方法进行分析比较后,提出限制各行为模式的使用范围,通过各行为的控制和融合,既达到有效避障,又能完成行驶目标的目的。通过几种典型障碍物环境下的避障仿真实验,结果显示设计达到了预期效果。

Given the non-linear, time varying operational characteristics of unmanned vehicles in the dynamic driving environment, this paper studies an action amalgamation-based intelligent control strategy of automatic driving. Five kinds of basic action models are designed according to fuzzy logic and the controllers corresponding to the actions are also devel- oped to control the vehicle＇s speed and direction. In the design of action selection mechanism the actions are integrated but the range of allowable actions is limited to support the goal of test. Through simulation, it has been shown that by action models control and amalgamation the unmanned vehicles can automatically drive to the destination in the dynamic environ ment, and the design is effective in several representative obstacle conditions.