基于运动轨迹预测控制的机械臂目标捕捉策略
Target Capture Strategy of Robotic Arm Based on Dynamic Trajectory Predictive Control
李双圻 1朱天启 1冒建亮1
作者信息
- 1. 上海电力大学自动化工程学院,上海 200090
- 折叠
摘要
利用机械臂捕捉运动中的物体是一项极具挑战性的任务,目标的轨迹预测、落点计算以及机械臂的运动控制需要在毫秒级内完成.为此提出一种基于运动轨迹预测控制的机械臂目标捕捉策略.首先,结合长短时记忆神经网络和卡尔曼滤波算法,设计一种改进的运动轨迹预测模型,实现动态目标的落点计算.其次,采用基于模型预测方法的位置视觉伺服控制,用于实时修正机械臂的姿态,从而实现运动物体的捕捉.最后,建立基于UR5 机械臂的仿真与实验环境,验证所提方案的有效性.
Abstract
Capturing moving objects with a robotic arm is a challenging task since the trajectory prediction of the target,the calculation of the landing point,and the motion control of the robotic arm need to be completed within milli-seconds.To this end,this paper proposes a target capture strategy for robotic arms based on motion trajectory predic-tion control.First,an improved motion trajectory prediction model was designed by combining the long short-term memory neural network and the Kalman filter algorithm to realize the calculation of the dynamic target's landing point.Second,the position-based visual servo control based on the model prediction method was constructed to correct the posture of the robotic arm in real-time,so as to realize the capture of moving objects.Finally,a simulation and experi-mental environment based on the UR5 manipulator was established to verify the effectiveness of the proposed scheme.
关键词
运动轨迹预测/模型预测控制/运动目标捕捉Key words
Motion trajectory prediction/Model predictive control/Motion target capture引用本文复制引用
出版年
2024
NETL
NSTL
万方数据