This paper presents an efficient numerical formulation for the prediction of realistic postures. This problem is defined by the method (or procedure) used to predict the posture of a human, given a point in the reachable space. The exposition addresses (1) the determination whether a point is reachable (i.e., does is it exist within the reach envelope) and (2) the calculation of a posture for a given point. While many researchers have used either statistical models of empirical data or the traditional geometric inverse kinematics method for posture prediction, we present a method based on kinematics for modeling, but one that uses optimization of a cost function to predict a realistic posture. It is shown that this method replicates the human mind in selecting a posture from an infinite number of possible postures. We illustrate the methodology and an accompanying experimental code through a planar and a spatial example, and validation using commercial human modeling and simulation code.
Realistic Posture Prediction for Maximum Dexterity
Sae Technical Papers
Digital Human Modeling For Design And Engineering Conference And Exhibition ; 2001
2001-06-26
Conference paper
English
NTRS | 1976
|Robotic dexterity via nonholonomy
Springer Verlag | 1998
|Dexterity analysis of robot for neurosurgery
British Library Online Contents | 2004
|On the dexterity of robotic manipulators
TIBKAT | 1986
|Increasing The Dexterity Of Redundant Robots
NTRS | 1990
|