Optimization of Wheg Robot Running with Simulation of Neuro-Fuzzy Control

Abstract

This paper presents laboratory simulator for wheellegged (Wheg) robot running and application for collecting measurement data. Data is used as a basis for modelling and optimization of energy consumption of running Wheg. The laboratory setup includes instrumented measurement treadmill (IMT) and Wheg drive. The laboratory experimental setup also includes the sensors, drives and software application. Intelligent modelling and optimization of energy usage during Wheg's running is based on a combination of neural networks and genetic algorithms. Neural network has established a correlation between the parameters of running. Using genetic algorithm optimal parameters for running are found. Simulation of neuro-fuzzy control system for minimization of energy usage during running was developed as a function of the angle and Wheg running speed.