COMPUTER SYNTHESIS AND OPTIMIZATION OF JUMPING MOTIONS VIA NONSTATIONARY CONSTRAINTS

Abstract

Introduction: This report considers the results of the authors’ research on the goal-oriented computer synthesis of human motions in support and non-support phases. The main attention is paid to the synthesis of the pushing phases. In particular, an analysis is made of the results of a sequential optimization of running long jumps and acrobatic jumps. The computer modeling of complex coordination motions is based on the development of an adequate anthropomorphic model. Methods and Results: Most effective in the developed modeling system proved to be the employment of differentiated non-stationary holonomic and nonholonomic constraints equations in order to model goal-oriented motions [1]. For descriptions of additional non-stationary items in constraints equations we used parametrically controlled smooth approximation functions which allowed us to synthesize the desired motion trajectories, ground reaction force and kinetic moment increment. Due to the non-stationary nature of constraints equations, any experimental data on kinematics and/or the dynamics of real motion can fulfill their function. For the analysis of modeling results we consider estimates of interelement control motions distribution in the support phase of jumping motion. A number of anthropomorphic model (AM) elements can change with respect to the level of AM adequacy to real human motion. For example, we used a 15-element AM for modeling the support and flying phases of the running long jump. Analysis of synthesized inter-element control moments values showed that the most significant influence on the value of the ground reaction and, therefore, on the pushingoff velocity was the motion of the swinging nonsupport leg. Variation of the parameters values of ground reaction and the resulting value of the kinetic moment allowed us to synthesize the AM motion in the support phase so that it would ensure the desired trajectory of the AM motion in the flying phase of acrobatic motions. Conclusions: Research showed the necessity of employment of non-stationary constraint equations in the synthesis of complex coordination human motions. Such an approach to motion control synthesis minimizes the number of parameters to be varied and gives a relatively stable solution with respect to small variations of AM structure. REFERENCES: 1. Zinkovsky, A.V., Sholuha, V.A., Ivanov, A.A. (1997). Mathematical Modeling and Computer Simulation of Biomechanical Systems, WSP, Singapore, 216.