基于神经网络的移动机器人对路径形式的识别与分类

The Identification and Classification on Road Curvatures for a Mobile Robot Based on Neural Network

在分析传统的３种路径跟踪方式（切线跟踪方式、弦线跟踪方式和圆弧跟踪方式）具有跟踪误差较大、运动不稳定等不足的基础上，提出了一种基于神经网络的移动机器人对其所跟踪路径进行实时识别和分类的新方法。

In this paper, by analyzing mobile robot classical road following approaches (e.g. tangential line and chordal line as well as unchanged arc road following methods) which have some disadvantages of relatively big road following error and unstable motion, a new method of real time identification and classification′s road curvature for a mobile robot based on ART 2 neural network is presented for the first time. The road following precision for mobile robot can be improved to a great extent by using this approach.