Control of hexapod with static-stable walking using artificial intelligence

Abstract

The objective of article is to describe a method for signal processing and decision making process of six-legged robots. An important characteristic of six-legged robot is its ability to maintain static stability while walking through terrain with obstacles. That is accomplished by evaluation of its leg position and body state and their adjustment realised by the decision making process. Control system of six legged robot model is composed of three main parts: walking pattern classification algorithm, fuzzy expert system, neural network for decision-making. This approach combines stand-alone method to solve problems of acyclic walking and methods of AI to speed up the decision-making process. The composite structure enables to apply the control law of the locomotion with help of the parallel co-operating control and learning. The control system of the hexapod allows to adopt the hexapod on the specific conditions of terrain. The new technique was tested on terrain with obstacles by selection of hexapod models. The described hierarchical technique based on coordination of mentioned methods has not been tested before. The process of determining the gait phases selects the safe action for each state to overcome the obstacles in the terrain.