A rail type mobile robot introduced in this paper is used to carry out automatic welding on the T, Y and K-joints in offshore platform jackets. This robotic system is composed of a mobile platform and a serial manipul...A rail type mobile robot introduced in this paper is used to carry out automatic welding on the T, Y and K-joints in offshore platform jackets. This robotic system is composed of a mobile platform and a serial manipulator. The structure of manipulator was designed according to the mathematical modeling of single brace tubular joint ; some improvements of the robot mechanism were considered to improve the flexibility" of manipulator and increase the workspace. The inverse kinematics analysis was solved through two steps calculation. The collision avoidance strategy used in the K-joint automatic welding was discussed. Moreover, the simulation results were presented to demonstrate the feasibility of the robot mechanism used in tubular joint welding and the validity of the joint value calculation method.展开更多
针对冗余机械臂运动规划过程中的避碰问题,提出一种基于深度学习的避碰运动规划方法。首先,建立冗余机械臂的运动学模型,借助Gilbert Johnson Keerthi算法进行碰撞检测,确定机械臂连杆与障碍物之间的安全距离;其次,将避碰运动规划问题...针对冗余机械臂运动规划过程中的避碰问题,提出一种基于深度学习的避碰运动规划方法。首先,建立冗余机械臂的运动学模型,借助Gilbert Johnson Keerthi算法进行碰撞检测,确定机械臂连杆与障碍物之间的安全距离;其次,将避碰运动规划问题转换成多元函数优化问题,以循环神经网络为主体框架,用天牛须搜索机制来防止算法陷入局部最优值和保证解的多样性;最后,在仿真环境下验证文中所提算法的有效性,结果表明:该算法的预测精度比改进人工势场法平均高出10.89%、算法时间节约2.44 s,其可以帮助冗余机械臂规划一条平滑、柔顺、无碰撞的运动轨迹,具有一定的工程参考价值。展开更多
基金This work was supported by the Tianjin Scientific and Technological Development Fund of Higher Education of China ( No. 20090406 ) , Tianjin Municipal Natural Science Foundation (No. 09JCYBJC04900 ) and the Key Technologies R&D Program of Tianjin (No. 10ZCKFGX03900).
文摘A rail type mobile robot introduced in this paper is used to carry out automatic welding on the T, Y and K-joints in offshore platform jackets. This robotic system is composed of a mobile platform and a serial manipulator. The structure of manipulator was designed according to the mathematical modeling of single brace tubular joint ; some improvements of the robot mechanism were considered to improve the flexibility" of manipulator and increase the workspace. The inverse kinematics analysis was solved through two steps calculation. The collision avoidance strategy used in the K-joint automatic welding was discussed. Moreover, the simulation results were presented to demonstrate the feasibility of the robot mechanism used in tubular joint welding and the validity of the joint value calculation method.
文摘针对冗余机械臂运动规划过程中的避碰问题,提出一种基于深度学习的避碰运动规划方法。首先,建立冗余机械臂的运动学模型,借助Gilbert Johnson Keerthi算法进行碰撞检测,确定机械臂连杆与障碍物之间的安全距离;其次,将避碰运动规划问题转换成多元函数优化问题,以循环神经网络为主体框架,用天牛须搜索机制来防止算法陷入局部最优值和保证解的多样性;最后,在仿真环境下验证文中所提算法的有效性,结果表明:该算法的预测精度比改进人工势场法平均高出10.89%、算法时间节约2.44 s,其可以帮助冗余机械臂规划一条平滑、柔顺、无碰撞的运动轨迹,具有一定的工程参考价值。