Fuzzy-stochastic functor machine for general humanoid-robot dynamics

Abstract

In this paper the fuzzy-stochastic-Hamiltonian functor-machine is proposed as a general model for the humanoid-robot dynamics, including all necessary degrees of freedom to match the "realistic" human-like motion. Starting with the continual-sequential generalization of the standard state equation for the linear MIMO-systems, the "meta-cybernetic" model of the "functor-machine" is developed as a three-stage nonlinear description of humanoid dynamics: (1) dissipative, muscle-driven Hamiltonian dynamics, (2) stochastic fluctuations and discrete jumps, and (3) fuzzy inputs, parameters and initial conditions. An example of symmetrical three-dimensional (3-D) load-lifting is used to illustrate all the phases in developing the functor-machine model.