基于粒子滤波的智能机器人定位算法

自主定位是智能机器人的关键性技术。针对轮式智能机器人在使用里程计、激光雷达进行定位过程中存在较大误差的问题,联合双目摄像机和激光雷达数据,提出基于粒子滤波的自适应蒙特卡洛(AMCL)优化定位算法。预测阶段,利用双目摄像机和激光雷达数据改善提议分布,减少滤波过程中重采样的粒子数,用更少的粒子数来估计机器人的后验概率分布。在激光雷达匹配点云时,提出一种分组阶梯式阈值判断法,在不降低点云匹配效果的情况下,有效降低现有的迭代最近点(ICP)匹配算法的计算量。为了验证改进算法的性能,在四轮智能机器人平台上进行实验。结果表明:改进的AMCL优化定位算法可以有效提高机器人的定位精度,具有较好的实用性。

INTELLIGENT ROBOT POSITIONING ALGORITHM BASED ON PARTICLE FILTER

Autonomous positioning is the key technology of intelligent robot.Aiming at the problem of large errors in the positioning process of the robot using the odometer and lidar,an adaptive monte carlo(AMCL)optimal positioning algorithm based on particle filter is proposed by combining the data of binocular camera and lidar.In the prediction stage,the binocular camera and lidar data were used to improve the proposed distribution,reducing the number of resampled particles in the filtering process to estimate the posterior probability distribution of the robot with fewer numbers of particles.In the case of lidar matching point cloud,a grouped stepped threshold judgment method was proposed,which could effectively reduce the computational load of the existing iterative nearest point(ICP)matching algorithm without reducing the matching effect of the point cloud.In order to verify the performance of the improved algorithm,experiments were carried out on a four-wheel intelligent robot platform.The results show that the improved AMCL algorithm can effectively improve the positioning accuracy of the robot and has good practicability.