Design and implementation of a complex agent using DFA for the MPR vision system

Compared with the existing mobile robot, the Mobile Parallel Robot (MPR) has more powerful climbing ability and more adaptive behaviors. This paper expands the complex agent using Deterministic Finite Automaton (DFA) from the wheeled mobile robot to a MPR successfully, and the monocular vision part is improved to stereo vision. The action status model with DFA is improved for the MPR structure. Besides that, the environment understanding module and the stereo vision module are added to the module kit. The model is used in the real-time object tracking experiment successfully and the MPR can adjust the height to avoid the obstacle in the air with stereo vision navigation.