Path optimization for multiple humanoid robot using TLBO based ANFIS controller in obscure environment

Abstract

Humanoid robot NAO is an intelligent reprogrammable agent designed to execute desired work using wireless control system by sensing the working environment at the same time. This article concentrates on the analysis of hybrid intelligent navigation approaches for single and multiple humanoid NAOs. It describes the optimization of a collision-free path in a static and dynamic terrain. The hybridization introduced is designed in two steps. ANFIS (Adaptive Network-based Fuzzy Inference System) controller produces a transitional driving angle (TDA) using a robot location with respect to obstacle distance in all directions. It is fed to Teacher Learner based Optimization (TLBO) approach that produces optimum driving angle (ODA) for the humanoid robot to guide along with its predefined target. It optimizes the path selection by taking Euclidean distance and ODA as the key factor. The hybridized controller is examined in an environment that consists of single and multiple NAOs. Case of inter-collision may occur in the path planning of multiple humanoid NAOs. It is eliminated with the integration of the dining philosopher controller in the base algorithm to prioritized one robot towards the assigned target. Simulation results demonstrate the proposed controller's efficacy and explain that it gives optimized path length and travel time. In addition, a comparative study with previously applied effective algorithms is done.