Managing Temporal Robot Constraints using Reachable Volumes

This project focuses on planning the motion for high degree of freedom manipulator robots under dynamic (or temporal) constraints. Manipulator robots are widely used in industry and are important because they can do jobs that are either too tedious or too dangerous for humans. An example would be picking up toxic waste or exploring underwater archeological sites. Motion planning for high degree of freedom (DOF) manipulators under task constraints is challenging because it gives rise to high dimensional configuration spaces (C-space) that are complex in structure. Our approach reduces the complexity by re-parameterizing the manipulator robots DOFs into a space that contains the valid regions that the end effector of the robot can reach, known as the Reachable Volume space (RV-space). In this way, we can sample valid configurations in Cspace in linear time with the number of DOFs of the manipulator. Current Reachable Volume theory only handles permanent constraints and cannot adapt to scenarios that require constraints that are enabled at certain times in the problem and disabled at other times. For example, when a manipulator grabs an object, closure constraints on the grasper must be satisfied, but when the object is to be dropped, these constraints must be ignored. Additionally, certain scenarios require the cooperation of multiple robots. This is obvious if we consider problems that involve objects that are too large for a single robot to handle. In this work, we produce a working computational framework for efficient motion planning of high degree of freedom manipulator arms under dynamic constraints through the extension of existing work in Reachable Volume spaces