基于多传感器融合的机器人障碍物检测和识别  

Obstacle Detection and Recognition of Robotics Based on Multi-Sensor Fusion

作　　者：戈新良[1] 杨杰[1] 

机构地区：[1]上海交通大学图像处理与模式识别研究所,上海200030

出　　处：《计算机仿真》2005年第3期168-170,共3页Computer Simulation

摘　　要：该文研究了在移动机器人中基于神经网络的多种传感器信息融合技术。通过由CCD摄像机获得的彩色图像与超声波传感器组获得的距离作为辅助视觉系统实现整个视觉系统功能。移动机器人的障碍物检测和识别的可靠性与精度比在任何单一传感器所获得的信息都有很大的提高。This paper presents the study of multi-sensor fusion technology based on neural network in mobile robot. Through image obtained by CCD and distance by the group of ultrasonic sensors as assistant vision system, the whole vision system function is realized. The reliability and precision in detection and recognition of obstacle based on multi-sensor information fusion are better than that based on any single sensor.

关 键 词：移动机器人 多种传感器融合 障碍物检测 神经网络 

分 类 号：TP314[自动化与计算机技术—计算机软件与理论]