Estudio comparativo basado en métricas para diferentes arquitecturas de navegación reactiva

Abstract

This study compares four autonomous navigation architectures for mobile robots, in order to determinate which has a greater efficiency in tasks solving. These architectures are: reactive, neuronal network, fuzzy logic and neuro-fuzzy hybrid architecture. The contrast within architectures was assessed collecting data of performance metrics based on safety, smoothness of curve (energy consumption), speed and distance. The comparison was done using the robotic platform, Pioneer 3-DX within a simulated environment of MobileSim, the official software of MobileRobot®. The collected data were statistically analyzed, generating, therefore a formal base of comparison. The outcomes of this procedure indicate that neuro-fuzzy architecture presents higher levels of security and energy efficiency in contrast with other architectures. Thus, using synergic architectures allows achieving a greater efficiency.