摘要
针对移动机器人路径规划与导航的实际应用,设计了一个基于全景视觉的移动机器人路径规划导航系统.首先,对导航系统的体系结构和功能进行描述.然后,分别就如何采用全景视觉传感器进行环境探索与地图创建,基于回归神经网络的广度优先搜索法和Voronoi骨架图法两种路径规划算法原理,以及如何实现按规划路径实施导航这三个方面进行了详细阐述.最后,结合实际机器人进行导航实验,评估导航系统的性能和路径规划算法的有效性.
For application development of path planning and navigation techniques for mobile robots, a navigation system based on omni vision is developed. Firstly, we briefly introduce the architecture and functionality of the navigation system. Secondly, the method of environment sensing and mapping based on omni vision is proposed. Meanwhile two path planning algorithms based on RNN (recurrent neural network) are also described in detail followed by the implementation of the navigator according to the planned path. Lastly, the performance of navigation system and the feasibility of the path planning algorithms are evaluated through a case study using a real mobile robot.
出处
《机器人》
EI
CSCD
北大核心
2005年第2期173-177,182,共6页
Robot
基金
国家自然科学基金资助项目(60105005).
关键词
全景视觉
回归神经网络
路径规划
导航
omni vision
RNN(recurrent neural network)
path planning
navigation
