Finger inverse kinematics using error model analysis for gesture enabled navigation in virtual environments

Abstract

In this paper we provide a new method for solving the hand fingers inverse kinematics problem. Given the finger's end-effector position with respect to the finger's metacarpal joint, the finger's four degrees of freedom joint angles are uniquely solved directly without iterations. The solution of a closely related, simpler inverse kinematics problem is used as a rough estimate of the finger's MCP and abduction angles. The error model of the estimate is used to correct the prediction. The error analysis is done a priori and is used directly in real-time. The method provides accurate results and is computationally efficient