Redundancy-Based Motion Planning with Task Constraints for Robot Manipulators

Abstract

Finding realistic motions for redundant manipulators is essential for complex jobs such as home care and industrial assembly. Motion planning is complex when a task requires standing upright or moving through restricted spaces. This work provides an effective motion-planning strategy for 7-DOF manipulators that improves connections via redundancy. The analytic Cartesian-space-to-joint-space kinematic mapping models for 7-DOF redundant manipulators with diverse configurations are constructed first, and the feasible nodes are determined by sampling the Cartesian space without barriers to satisfy the task requirements. Each Cartesian-space sampling node can provide numerous feasible joint-space nodes because of the redundancy of the robot manipulators. To remove additional valid nodes from a singular position, joint configurations with the same end-effector position orientation are modified iteratively. Finally, we find the nearest nodes in the joint-space constraint manifold and build collision-free smooth pathways. The task constraint levels were varied for a 7-DOF manipulator in simulations and experiments. The proposed planner finds more viable nodes at the same end-position attitude than one-to-one projection. It does not require numerical iterations and achieves high planning efficiency and a high motion-planning success rate.