Task-based servoing in quaternion space

This paper introduces and analyzes a method based on vector-quaternion pairs for accomplishing task-based control. The optimal control paradigm has been extended into quaternion space with a complete definition of the control law required to control a plant in such a space. A complete set of derivations of the quaternion algebra used to accomplish the optimal control have also been developed. The advantage of this approach is that it decouples the position and orientation states, thereby reducing the complexity of the control algorithms. The intended application of this research is task-level control of the five-fingered Anthrobot mounted on a PUMA 760.