论文部分内容阅读
针对空间机械臂对运动目标的抓捕提出了一种分段的在线运动规划方法,重点解决机械臂可能遮挡立体视觉相机视线的问题,并满足抓捕时间要求.基于扩展卡尔曼滤波器(EKF)建立了目标运动状态估计器.对立体视觉相机的视线遮挡约束进行了数学建模.将机械臂向目标的接近过程分成2个阶段,在2个阶段中分别使用多约束环境下的滚动RRT(快速扩展随机树)方法和能够快速接近目标的比例导引算法,并根据对目标运动状态的估计精度自主切换运动段.同时考虑组合体的动力学耦合特性,在运动规划中限制了平台姿态角速度.利用数学仿真验证了本文的目标运动状态估计方法和运动规划方法.比例导引方法可能由于机械臂遮挡立体视觉相机观测视线而抓捕失败,而本文的分段运动规划方法对全部仿真情况都能成功抓捕目标.本文的分段运动规划方法能够对各个方向运动的目标进行有效的运动状态估计并快速可靠地抓捕,避免了因遮挡立体视觉相机观测视线引起的抓捕失败.基于目标运动状态估计的切换策略能够根据实际的目标运动情况在线自主地切换2个运动段,对运动状态未知的目标具有鲁棒性.
Aiming at the capture of moving target by space manipulator, a segmented online motion planning method is proposed, which focuses on solving the problem that the manipulator may occlude the line of sight of stereo vision camera and meets the arresting time requirements.Based on the extended EKF ) Established a target motion state estimator.Mostly, the mathematical model was established for the line-of-sight occlusion constraint of the stereo camera.The robot’s approach to the target was divided into two stages, using the rolling RRT (Fast expanding random tree) method and a proportional navigation algorithm that can quickly approach the target, and automatically switch the motion segments according to the estimation accuracy of the target motion state. At the same time, considering the dynamic coupling characteristics of the combination, the platform attitude is limited in the motion planning Angular velocity.The mathematical model is used to verify the target state of motion estimation and motion planning method in this paper.Pilot navigation method may fail to capture due to the occlusion of stereo camera by the manipulator.However, Can successfully capture the target.In this paper, the staged motion planning method can move in all directions The target can effectively estimate the state of motion and capture quickly and reliably, which avoids the capture failure caused by obstructing the view of the occluded stereoscopic camera.The switching strategy based on the target state of motion estimation can switch 2 autonomously according to the actual target motion Exercise segment, the state of motion is unknown goals with robustness.