基于机器人运动模型的EKF-SLAM算法改进

摘要

针对两轮驱动机器人运动模型定向误差的累积问题,提出改进的三轮驱动机器人运动模型,对EKF-SLAM算法的一致性进行改进;该模型通过对机器人车轮线速度的解耦运算,将机器人运动过程中的旋转角速度提取出来并作为系统的控制输入之一,从而可以直接得到各个控制时间间隔内的机器人姿态角变化,很好地避免了机器人定向误差的累积;量后,基于归一化估计方差的检验标准进行实验,验证了三轮驱动机器人运动模型有效提高了EKF- SLAM算法的一致性.%Aiming at the problem of orientation error accumulation of two wheels driving robot motion model*,improved three wheels driving robot motion model is presented to improve the consistency of EKF-SLAM algorithm. This model extracts the angular velocity in the robot movement process with the decoupling operation of wheels' linear velocity as one of the control inputs of the system, and thus the change of robot attitude angular in each control time interval can be got directly, and avoids the accumulation of robot orientation error well. Finajly, doing experiments based on the distinguished rules and testing the three wheels driving robot motion model has improved the consistency of EKF-SLAM algorithm efficiently.