基于粒子滤波的移动机器人目标追踪控制改进算法

Improved Algorithm of Mobile Robot target Tracking Control Based on Particle Filter

针对移动机器人目标追踪控制问题,以三轮全向移动智能足球机器人为研究对象,结合图像区域分割技术建立机器人运动模型和测量模型,提出基于高斯和粒子滤波的移动机器人目标追踪改进算法应用于足球机器人比赛环境。该算法将高斯和滤波中的并行扩展卡尔曼滤波用并行高斯粒子滤波代替,充分考虑当前时刻观测量对状态分布的影响,提高系统精度的同时降低滤波过程中使用的粒子数目,减轻系统计算量;仿真结果表明,该算法有效的减少足球机器人目标追踪过程中的误差和修正时间,特别是目标物体曲线行驶时,可以有效的将x轴与y轴方向的误差控制在一个较稳定的范围内,提高系统的稳定度。

Focusing on the problems of mobile robot target tracking, taking the intelligent mobile soccer robot with three omni- direc tional wheels as the research object, the Gaussian sum particle filter （GSPF） improved algorithm is provided for mobile robot target tracking based on the robot motion model and measurement model combined with the image of regional segmentation technology for soccer robot game environment. Parallel extended kalman filtering is replaced by Parallel Gaussian particle filter of Gaussian sum filtering in the G-SPF, and full consideration is given to the influence of the current observed quantity to the state distribution, which improves the precision and relieves the computation of system by lowering the number of particles used in filtering. The simulation results show that this algorithm effectively re duces the error and correction time in the process of soccer robot target tracking and controls the x axis and y axis error in a more stable range especially when the target objects curve driving, improves the stability of system.