滑移转向机器人不确定度分析及路面识别方法

Uncertainty Analysis of Slip Steering Robot and Pavement Identification Method

传感器的不确定度是移动机器人定位中的关键问题;文章对Pineer3-AT滑移转向机器人在转弯时运动学状态进行分析,发现了滑动偏差受地面与轮的摩擦系数及左右两轮的速度影响,通过Adams与Matlab/Simulink联合仿真实验得到了在不同地面不同轮速的滑动偏差的大小多组数据;为了使仿真结果对研究里程计不确定度有一定的参考价值,在仿真轨迹重复性偏差中选取最大值,并对结果拟合,建立了滑动偏差模型,并通过实验进行了验证对比;对仿真结果分析可以得出,摩擦系数越大,在相同速度下滑动偏差越小,根据这一特性,提取小车在不同轮速下的滑动偏差作为地面分类的原始数据;通过k-近邻(KNN)方法,对地面进行分类,识别率达到70%以上。

Sensor uncertainty is a key issue in mobile robot positioning.The paper analyzes the kinematics of the Pioneer3-AT sliding-steering robot during cornering and finds that the sliding amount is affected by the friction coefficient of the ground and the wheels and the speed of the left and right wheels.Adams and Matlab/Simulink co-simulation experiments have obtained multiple sets of data for different sliding speeds at different ground speeds.In order to make the simulation results have a certain reference value for the study of odometer uncertainty,the maximum value is selected in the repetitive deviation of the simulation trajectory,and the results are fitted.The sliding deviation model is established and verified by experiments.Based on the analysis of the results,it can be concluded that the larger the friction coefficient,the smaller the sliding amount at the same speed.According to this characteristic,the sliding amount of the trolley at different wheel speeds is extracted as the raw data of the ground classification.The k-nearest neighbor(KNN)method was used to classify the ground and the recognition rate reached over 70%.