摘要
对自主引导车辆或自主移动机器人来说,在未知环境中能避开障碍而成功到达预定的目的地是一项基本要求。Bug及其衍生算法是实现机器人在未知环境中避开障碍到达指定目的地的基于传感器的著名导航算法,它们认为机器人是平面上的一个点,具有360°障碍检测范围,具有贴着障碍物边缘绕行的能力,对通常的非360°障碍检测范围且具有物理尺寸的机器人来说,Bug类算法就不能直接应用。提出一种适用于这类机器人的改进的Bug算法,包括结合DistBug算法、VisBug算法得到较短路径的'相遇点'与'脱离点'的确定方法,基于按分段直线与速度空间法结合得到圆滑自然路径的绕行障碍物边缘算法。还提出'虚拟触角'的概念来分析利用传感器数据,机器人依靠多种'虚拟触角'实现基于传感器的改进Bug算法。编写出完整的程序实现该算法,并以Pioneer3-AT室内机器人为对象对改进算法进行验证。结果表明,机器人可以利用提出的Bug改进算法实现未知环境下定目标点无碰撞导航。
A basic requirement for autonomous guided vehicles or indoor autonomous mobile robots is that they should be able to avoid obstacles in unknown environment and successfully arrive at the predetermined destination.Bug and its derivative algorithms are famous sensor-based navigation algorithms which can meet the above-mentioned requirement.They hold that a robot is a point on a plane,having a 360° sensing range,capable of skirting any obstacle.However,Bug algorithms cannot be directly applied to the ordinary robot with physical dimensions and a non-360° sensing range.A novel and practical method to accomplish inherently collision-free boundary following is presented so that DistBug algorithms,a derivative from Bug algorithms,can be implemented on these mobile robots,which features the following：A straight path technique reducing robot＇s motion to pure turning and pure translation;Pure turning is further divided into three classes to serve different purposes;virtual antennae are used to utilize sensory data to provide inputs to the whole algorithm for decision making.Based on this boundary following method a practical goal directed navigation approach for robots equipped with non-360° sensing system is implemented,which is tested with a Pioneer 3-AT mobile robot where a 180° laser range finder fits.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2010年第19期44-52,共9页
Journal of Mechanical Engineering
基金
教育部留学人员科研启动基金
国家高技术研究发展计划(863计划
2006AA10A304
2006AA10Z255)资助项目
关键词
避障
自动导航
非360°传感器
移动机器人
虚拟触角
Obstacle avoidance Autonomous navigation Non-360° sensor Mobile robots Virtual antenna
