摘要
针对移动机器人定位中航向角精确测量问题,设计了径向基函数 (RBF)神经网络来实时获取精确的航向角。使用正交最小平方 (OLS)算法训练神经网络,确定构建RBF网络所需的相关参数。基于RBF神经网络的组合传感器测量系统不仅能消除测量误差,使机器人工作过程中的定位精度提高近 8倍,且具有一定的工程实用性。实验结果表明:构建的RBF神经网络能够实时获取精确航向角,保证移动机器人在户外工作环境中完成指定任务。
Aiming at the successful localization of mobile robot,the problem of acquiring accurate heading angle is resolved.The RBF network is designed for mobile robot to acquire the accurate and real-time heading angle.Using the OLS algorithm,several designs related to the network architecture and training have been made to construct the RBF network.The proposed metrical system with combined sensors based on the RBF network can make the accuracy of the localization well improve (near 8 times),also be of engineering practicality.The results of experiment show that the designed neural network can acquire the accurate heading angle and guarantee the robot to complete assigned tasks in outdoor operational environment.
出处
《传感器技术》
CSCD
北大核心
2005年第2期62-65,共4页
Journal of Transducer Technology
关键词
径向基函数神经网络
定位
传感器
正交最小平方算法
移动机器人
RBF(radial basis function) neural network
localization
sensor
OLS(orthogonal least square) algorithm
mobile robot