基于激光成像雷达的未知目标相对位姿估计算法

LIDAR-Based Relative Position and Attitude Filtering for Unknown Targets

为解决对空间未知目标的相对位置、姿态估计问题,以激光成像雷达作为测量敏感器,提出了基于扩展Kalman滤波(EKF,Extended Kalman Filter)的相对位姿估计算法。采用迭代最近点算法(Iterative Closest Point,ICP)对激光雷达的点云测量数据进行解算,得到相对位姿粗值并将其作为位姿估计算法的测量输入。以相对姿态、角速度、惯量比、相对位置、相对速度和目标测量参考系的位姿作为滤波状态,算法在对相对位置和姿态估计的同时,可辨识出目标的未知参数。为提高数值仿真的可信度,用Geomagic软件模拟点云测量。采用Matlab进行数值仿真,验证了新算法的有效性。

A LIDAR-Based Extended Kalman Filter （EKF） for relative position and attitude estimation of unknowns target was proposed. The relative position and attitude between the target and a servicing spacecraft was solved by the Iterative Closet Point （ICP） using LIDAR point cloud data, which served as the EKF＇s measurement input. The system states of EKF include the relative attitude, angular velocity, inertia ratios, relative position, relative velocity, and the position/attitude of target measurement reference frame with respect to target principle frame. The proposed filter estimated the relative position and attitude as well as the unknown parameters of the target. To improve the confidence of numerical simulation, geomagic was used to simulate the point cloud data of LIDAR. A simulation based on Matlab verifies the proposed algorithm.