A. Abbas and M. El-Samadony (Egypt)
Behaviour based robotics,logicObstacle avoidance, Fuzzy
This paper describes the development of a control architecture suitable for the obstacle avoidance autonomous navigation of mobile robots. A reactive multi-behaviour based control architecture has been developed which generates control actions in response to stimulated inputs obtained from the robot sensors. It is a modular architecture which can have behaviours added or removed based on required operating environment conditions. It also allows for the definition of relationships between the various behaviours. Uncertainties of the sensory information associated with obstacle avoidance navigation have been identified. A fuzzy logic algorithm has been defined to be able to cater for these uncertainties. Membership functions for the fuzzy variables have also been defined. A suitable rulebase and inference mechanisms have been setup encapsulating the human intrinsic navigational decision making procedural skills. A series of scenarios have been setup and executed in experiments in order to attest the feasibility of the developed architecture. The results of the trials demonstrated the success of the behaviour based architecture in achieving autonomous obstacle avoidance.
Important Links:
Go Back
