摘要
为提高移动机器人的工位定位精度,通过实验分析了超声波传感器的测量距离d和入射角α对测量精度的影响。基于代数神经网络能实现样本空间的精确映射并具有较好非线性逼近能力,设计了一种移动机器人侧向定位融合模型。经实验比较,该定位融合模型具有较高的精度,使得位置误差小于0.9957mm,角度误差小于0.2966°。将该定位融合模型应用于自主研发的移动机器人的定位实验中,定位位置精度可达到±2.5mm,姿态角精度可达到±0.42°,满足定位要求。
To improve positioning precision of a mobile robot, the effect of measuring distance d and incident angle a of ultrasonic sensor was analyzed on measuring accuracy by experiments. Lateral positioning model was designed using algebra neural networks fusion algorithm because the fusion algorithm can precisely map in sample space and has better nonlinear approximation ability. By compari- son experiments, the positioning fusion model has higher precision, location error is decreased to less than 0. 9957mm and angle error is decreased to less than 0. 2966°. Finally, the positioning fusion model was applied to a self--developed mobile robot. The experimental results show the robot has location error of ±2. 5mm and angle error of ±0. 42°, and meets the positioning needs.
出处
《中国机械工程》
CAS
CSCD
北大核心
2008年第17期2102-2107,共6页
China Mechanical Engineering
基金
国家自然科学基金资助项目(50075069)
陕西省教育厅产业化培育计划资助项目(03JC17)
关键词
移动机器人
信息融合
代数神经网络
侧向定位
mobile robots information fusion
algebra neural networks lateral position
