摘要
针对木材加工车间的地面环境复杂多变的特点,利用CCD摄像头、辅助电路和控制器构成机器人的视觉系统,对多传感器智能清洁机器人进行动态避障路径规划。控制器之间选用并行主从式通讯方式,根据木材加工车间地面非结构化的特点,应用遗传规划对智能清洁机器人的路径规划进行研究,达到运动路径的全局最优化;在避障方面,利用视觉系统的信息反馈,对静态的障碍物进行避开,对复杂多变的动态障碍物进行分析,包括速度和运动趋势,进行智能避障和最优路径的规划。现场试验验证表明:设计能够使机器人实现预期的各项功能指标(如,自主避障及避障后最优路径的选择、无人干预时自主工作等);试验过程中,整个系统的工作稳定,能够很好地实现木材加工车间地面的智能清洁。
Because of the complicated environment of wood processing workshop, we used CCD camera, auxiliary circuit and controller and a robot vision system to establish dynamic obstacle avoiding path plan for multi-sensor intelligent cleaning robot. We uses parallel master-slave way to communicate with the system between the controllers. According to the unstructured characteristics of wood processing workshop ground, we used genetic programming algorism to plan the path, and do the global optimization of motion path. In obstacle avoidance, the static obstacles are avoided based on information feed- back of vision system, and the dynamic obstacles are avoided based on analysis of speed and movement trend. Therefore, we realized the intelligent obstacle avoidance and optimal path planning. By field test results, the robot system can achieve all the expected function indexes （e. g. , autonomous obstacle avoidance and optimal path selection, autonomous work without any human intervention, etc. ）. The working system is stable with good ground intelligent cleaning.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2014年第6期143-147,共5页
Journal of Northeast Forestry University
基金
东北林业大学大学生创新项目(JD201301)
关键词
木材加工车间
清洁机器人
动态避障
路径规划
Wood processing workshop
Cleaning robot
Dynamic obstacle avoidance
Path planning
