Navigational Analysis of Artificial Intelligence based Control Strategy for Humanoid Robot

Abstract

To accomplish difficult tasks in cluttered environments and to reduce human effort in performing a task, humanoid robots are employed and have gained immense research interest. While performing any hazardous or tedious task, it is important for the humanoid robot to have correct navigational trajectory. Hence, the investigation about the navigational strategies of humanoids may be treated as a subject of significant concern. The current investigation is focused on path optimisation and obstacle avoidance capabilities for humanoid robots using the artificial intelligence techniques. By performing the kinematic analysis of the humanoid and a standard designed biped robot, the fundamental constraints and important parameters in the motion analysis can be conceived. The artificial intelligent techniques used here are Adaptive Particle Swarm Optimisation (APSO) and Adaptive Ant Colony Optimisation (AACO) for path planning by avoiding static or dynamic obstacles. The path length and the time required to reach the target position from the initial point are the important aspects, which are considered while designing a particular navigational controller. After application of standalone techniques, a hybridization of different techniques has been performed that reveals better results as compared to individual methods. The algorithms are first simulated in the VREP 3.1.3 software then for the validation of the same real time experiments has been carried out by taking the NAO and Biped Robot. Multiple NAO robots have also been maneuvered in various environments for which each NAO acts as a dynamic obstacle for the other in addition to the static obstacles. Finally, a comparison among all the techniques reveals facts about the efficiency of each technique in terms of path length and time required. Some future scopes are also suggested in this field of research in terms of qualitative and quantitative approach.