摘要
针对移动机器人定位问题,以自主搭建的复合式机器人为基础,提出一种基于飞蛾火焰优化-反向传播(MFO-BP)算法的移动机器人定位预测方法。将移动机器人视为一个“黑箱”,不单独考虑系统和非系统误差的影响,输入理论坐标值,输出预测坐标值。试验结果表明,MFO-BP算法预测模型能有效进行移动机器人定位预测,并且精度远高于传统反向传播(BP)神经网络预测模型。为了验证模型结构对预测结果的影响,将MFO-BP算法预测模型分为单隐含层和双隐含层这两种。试验结果显示,MFO-BP算法双隐含层与单隐含层相比,前者平均绝对误差更小、误差波动范围也更小、预测误差趋势更平稳。MFO-BP算法双隐含层预测效果更优,可以应用于复合式机器人末端定位。
Aiming at the problem of mobile robot localization, based on an autonomous composite robot, a moth-flame optimization-back propagation (MFO-BP) algorithm-based is proposed as a prediction method for mobile robot localization.The mobile robot is regarded as a “black box”, and the influence of systematic and non-systematic errors is not considered separately, the theoretical coordinate values are input, and the predicted coordinate values are output.The experimental results show that the MFO-BP algorithm prediction model can effectively predict the positioning of mobile robots, and the accuracy is much higher than that of the traditional back propagation(BP) neural network prediction model.To verify the influence of the model structure on the prediction results, the MFO-BP algorithm prediction model is divided into two kinds: single hidden layer and double hidden layer.The experimental results show that compared with the MFO-BP algorithm between double hidden layer and single hidden layer, the average absolute error of the former is smaller, the range of error fluctuation is also smaller, and the trend of prediction error is smoother.The prediction effect of the MFO-BP algorithm double hidden layer is better, and it can be applied to composite robots ’ end localization.
作者
陈泉
王湘江
CHEN Quan;WANG Xiangjiang(School of Mechanical Engineering,University of South China,Hengyang 421001,China)
出处
《自动化仪表》
CAS
2024年第7期40-44,共5页
Process Automation Instrumentation
基金
中央引导地方科技发展基金资助项目(2022ZYQ015)。
关键词
移动机器人
定位
预测模型
飞蛾火焰优化算法
反向传播神经网络
隐含层
Mobile robot
Localization
Prediction model
Moth-flame optimization(MFO)algorithm
Back-propagation(BP)neural network
Hidden layer
