一种基于超声波和AMCL的室内机器人定位改进方法  

An improved indoor localization method for robot based on ultrasonic and AMCL

作　　者：李阳 吴明晖 LI Yang;WU Minghui(College of Mechanical and Automotive Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学机械与汽车工程学院,上海201620

出　　处：《农业装备与车辆工程》2023年第6期117-121,146,共6页Agricultural Equipment & Vehicle Engineering

摘　　要：针对自适应蒙特卡洛定位算法(Adaptive Monte Carlo Localization, AMCL)在全局定位方面的不足,提出一种融合超声波双信标信息的改进AMCL定位方法。在全局定位过程中,首先利用聚类算法求解出超声波定位点群质心,从而消除障碍物干扰带来的误差,然后依靠安装在机器人上的成对信标,解算航偏角,获得完整的全局定位信息。将超声波定位结果作为AMCL的初始状态,对机器人进行位姿跟踪。实验表明,该方法无须机器人在空间中运动探索,可以直接完成AMCL算法的全局粒子收敛,缩短定位时间,并提供了精准的初始定位状态。Aiming at the shortcomings of Adaptive Monte Carlo Localization(AMCL)algorithm in global localization,an improved AMCL localization method was proposed by integrating ultrasonic dual beacon information.In the process of global positioning,the centroid of ultrasonic anchor point group was firstly solved by clustering algorithm so as to eliminate the error caused by obstacle interference.The complete global positioning information can be obtained by solving the yaw angle by the pair of beacons mounted on the robot.The ultrasonic positioning results were used as the initial state of AMCL to track the robot's position and pose.Experiments show that this method can directly achieve global particle convergence of AMCL algorithm without robot motion exploration in space,shorten positioning time,and provide precise initial positioning state.

关 键 词：移动机器人 全局定位 超声波定位 自适应蒙特卡洛 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]