论文部分内容阅读
为了提高移动机器人地形分类的准确率,提出基于原始数据时域幅值分析的特征提取方法,利用LIBSVM中的一对一支持向量机(SVM)程序,采用投票决策法实现分类,给出票数相同情形下的新算法.在四轮移动机器人左前轮轮臂上安装x、y、z向加速度计和z向传声器,使之在沙、碎石、草、土、沥青地面上分别以6种速度行驶,提取车轮与地面相互作用的加速度和声压信号.根据本文的算法,分别对每种速度下的5种地形进行分类,平均准确率为88.7%.
In order to improve the accuracy of terrain classification of mobile robot, a feature extraction method based on time-domain amplitude analysis of raw data is proposed. By using the one-to-one support vector machine (SVM) program in LIBSVM, voting is used to achieve classification. The new algorithm in the case of the four-wheel mobile robot mounted on the front left wheel arm x, y, z accelerometer and z to the microphone to sand, gravel, grass, soil, asphalt ground respectively 6 species According to the algorithm of this paper, the five kinds of terrain under each speed are classified respectively, with the average accuracy of 88.7%.