Dynamic analysis of anthropomorphic manipulators in computer animation

Loizidou, Stephania M. (1992) Dynamic analysis of anthropomorphic manipulators in computer animation. Doctoral thesis, University of North London.

Abstract

The thesis examines the motion control and motion coordination of articulated bodies, in particular human models, through the use of dynamic analysis.

The work concentrates particularly on the use of dynamic analysis as a tool for creating an animated walking sequence. The aim is to relieve the animator of the tedious specification of unknown parameters of motion control, and instead, to investigate the possibility of shifting this specification and much of the work to the animation software. Using this idea. a higher level of control could be used to coordinate the movement.

This has been examined through the use of a Hybrid Inverse and Direct Dynamics System, HIDDS, based on Featherstone's formulation. The inverse system is used to get an initial estimate of the values of the forces and torques needed to move the model to the required position. Once the values of these forces/torques are determined. a direct dynamics system is called to compute the accelerations produced from the influence of the input forces/torques. Double integration over time then gives the new positions. Applications of the hybrid technique can create animation sequences through the use of the graphics editor AnthroPI (Anthropomorphic Programming Interface) especially created for the implementation of the desired results.

The problem of ground reaction forces has also been studied and formulated. An algorithm proposed by Kearney for a one-legged hopping machine was extended to the two-legged anthropomorphic model by the introduction of the virtual-leg concept.

The dynamics approach showed promise in creating motion sequences which are both natural and realistic and HIDDS proved to be a useful experimental tool on which further research can be based. At its early stage. it provides a middle-level of control. which we believe has the potential to be upgraded to a higher level.

Documents
7464:38604
[img]
Preview
314630.pdf - Published Version

Download (6MB) | Preview
Details
Record
Statistics

Downloads

Downloads per month over past year



Downloads each year

View Item View Item