基于激光雷达的移动机器人实时避障策略

Real-time obstacle avoidance for mobile robots strategy based on laser radar

以激光雷达为主要传感器,对移动机器人设计一种实时避障算法。该算法考虑到机器人的非完整约束,利用基于圆弧轨迹的局部路径规划和控制使之能够以平滑的路径逼近目标位置。采用增强学习的方法来优化机器人的避障行为,利用激光雷达提供的报警信息形成刺激反应式行为,实现了动态环境下避障行为,具有良好的实时反应能力。该控制算法采用分布式软件设计方法,各功能模块异步运行,较好地实现了局部规划与全局导航目标的结合。该策略针对移动机器人MORCS在未知环境下实现了实时、有效避障,动作稳定流畅,轨迹平滑,具有良好的效果。

A real-time obstacle avoidance algorithm for mobile robot which took laser radar as its main sensor was put forward. This algorithm takes into account the nonholonomic restriction of robot. Robot can smoothly reach the goal through local planning based on circle locus controlling. At the same time a reinforcement method is used to optimize its obstacle avoidance algorithm. The stimulate-action behavior is formed by using the alarm information of laser radar, and it can realize the obstacle avoidance in the dynamic environments. This algorithm adopts a distrib- uted system, each module is asynchronously operated and the system can be easily combined with global planning. This algorithm is successfully implemented on mobile robot MORCS. It can be used to avoid the unknown obstacles without collision and to simultaneously steer the mobile robot toward the target in smooth, stable and continuous motion.