Kinematic Control of Redundant Manipulators

Abstract

While most manipulators have enough degrees of freedom (DOFs) to perform tasks in their end-effector task space (desired position and orientation), their workspace is limited due to mechanical joint constraints and obstacles that may be present in the work space. Redundant manipulators have extra DOFs than required for reaching the desired position and orientation of the end-effector. This allows the redundant manipulators to use the extra DOFs to avoid their joint limits and obstacles in the workspace, while still reaching a desired end-effector pose in the task space. The objective of the thesis is to implement and analyze the performance of most common methods for redundancy resolution with respect to algorithmic and kinematic singularity, and incorporation of different additional desired tasks, e.g., joint limit avoidance, singularity avoidance. Moreover, wrist mounted force/torque sensor is calibrated and used to detect the contact with the objects in the environment and for the robust extraction of the object during picking and stowing operations. The thesis is a part of work done for Amazon Robotics Challenge 2017.