An Obstacle Cluster Based Data Fusion Method for Mobile Robot

首先给出了一种通过融合多个超声波传感器和一台激光全局定位系统的数据建立机器人环境地图的方法 ,并在此基础上 ,首次提出了机器人在非结构环境下识别障碍物的一种新方法 ,即基于障碍物群的方法 .该方法的最大特点在于它可以更加简洁、有效地提取和描述机器人的环境特征 ,这对于较好地实现机器人的导航、避障 ,提高系统的自主性和实时性是至关重要的 .大量的实验结果表明了该方法的有效性 .

In this paper, a novel practical approach, named obstacle cluster based method, is introduced for the problem of obstacle recognition of mobile robot in completely unknown environments. The method is based on an improved grid map produced by fusing the data from onboard ultrasonic sensors and a laser global localization system in real-time. The main advantage of this approach lies in its concise and efficient representation of the environments, which is crucial for robot navigation and obstacle avoidance. Extensive experiments illustrate the effectiveness of the proposed approach.

国家高技术研究发展计划资助项目