A Data-Driven Sparse Motion Planning Scheme for Redundant Manipulators
Long Jin, Jinchuan Zhao, Shuai Li
Abstract
A large number of motion control schemes have been developed for redundant manipulators in the past few decades to solve their control problems. These resolutions are often based on the structure information of a manipulator being precisely known and require the manipulator to adopt a full-level joint actuation way when performing a given trajectory tracking task. To solve the problem of controlling redundant manipulators with model unknown in a sparse manner, a data-driven sparse motion planning (DSMP) scheme and the corresponding dynamic neural network (DNN) are proposed in this brief. Simulative experiments confirm the effectiveness and superiority of the proposed scheme solved by DNN.
Topics & Concepts
Scheme (mathematics)Computer scienceTrajectoryRobot manipulatorControl theory (sociology)Motion (physics)Motion planningTracking (education)Artificial intelligenceTask (project management)Manipulator (device)Artificial neural networkControl (management)RobotMathematicsEngineeringAstronomySystems engineeringPedagogyMathematical analysisPsychologyPhysicsRobotic Mechanisms and DynamicsRobot Manipulation and LearningProsthetics and Rehabilitation Robotics