The development of fuzzy logic based navigation algorithm for multi-featured autonomous robot

Abstract

Successful autonomous robot exploration and navigation requires control algorithms that are capable of navigation through complex unknown environments. These control algorithms must be able to successfully avoid newlinedensely spaced obstruction while navigating a relatively efficient but safe path. Navigation through an indoor and outdoor autonomously attracts special interest newlinefrom various fields. Behavior-based control systems address this task without the need for prior information, these systems deviate from the general mapping approach by eliminating modelling and attempting to make the most direct matching between perception and action. The major components of behavior based systems are a set of independent parallel behaviors and an arbitration or fusion method for combining the behavioral newlinereactions to from a single desired control output. The arbitration problem is the newlinemost challenging problem to the behavior-based navigation, mainly due to the increased demand of complex environments. The major arbitration and fusion approaches are the sub-assumptive and the motor schema approaches. Subsumptive systems switch between behaviors based on the context while motor schema-based architectures combine behavior reactions with vector addition. newlineThe early approaches of behavior based systems used arbitration behaviors in newlinewhich behaviors respond to its stimulus by evoking a single control command. Poor arbitration in these systems led to the development of fusion based behavior newlinesystems in which each behavior respond to its stimulus by combining several control commands. Two approaches to fusion based behavior systems have been proposed the crisp logic approach and the fuzzy logic approach. The fuzzy logic approach is not only simple but also adds robustness to unreliable sensor newlineinformation and encourages smooth control.