摘要
分析了超声测距原理及其存在误差的原因,提出了一种采用BP神经网络的超声测距误差补偿算法。该算法可对给定的输入向量和目标向量进行样本训练,在训练过程中不断调整权值、阈值,最终达到一定的映射关系以修正误差。仿真结果验证了该算法的有效性。
The paper analyzed principle of ultrasonic ranging and its main error causes and put forward an error compensation algorithm of ultrasonic ranging based on BP neural network.The algorithm trains samples of given input vector and target vector,and adjusts weights and threshold in process of training constantly,so as to reach a mapping relationship to correct error.The simulation result verified effectiveness of the algorithm.
出处
《工矿自动化》
北大核心
2012年第12期62-65,共4页
Journal Of Mine Automation
基金
陕西省自然科学基础研究计划项目(SJ08A26)
中小企业技术创新基金项目(08C26226102372)
关键词
超声波传感器
超声测距
测距误差
补偿算法
BP神经网络
ultrasonic sensor
ultrasonic ranging
ranging error
compensation algorithm
BP neural network
