• PDF / 6,702,566 Bytes
• 13 Pages / 595.276 x 790.866 pts Page_size
• 13 Downloads / 209 Views

DOWNLOAD

REPORT

---

RESEARCH PAPER

A Hybrid Artificial Immune System for Mobile Robot Navigation in Unknown Environments Prases K. Mohanty1

•

A. A. Kodapurath1 • Rishi Kumar Singh1

Received: 26 September 2018 / Accepted: 14 January 2020  Shiraz University 2020

Abstract To solve the problem of obstacle avoidance for mobile robot in an unknown environment, an algorithm is proposed in this article which uses both the idiotypic network theory and clonal selection theory of the artificial immune system. While the former is used for navigation in general, the latter comes into effect during local minima situations. Also, a modified version of Farmer’s equation for Jerne’s idiotypic network model has been used to determine the antibody concentrations. The simulation results for navigation using the proposed algorithm are presented. The results proved that our mechanism is capable of successfully avoiding various types of obstacles including the local minima traps. A comparative study between the proposed algorithm and various other algorithms is also presented. Finally, experimental validation of the simulation results for the proposed algorithm is carried out using a physical robot. Keywords Artificial immune system  Mobile robot  Navigation  Obstacle avoidance  Path planning

1 Introduction Autonomous mobile robots have potential applications in all kinds of environments such as indoor, outdoor, industrial, underwater exploration and many more due to their superior mobility. Obstacle avoidance and path planning are important factors of mobile robot navigation. In the present work, a hybridized variant of AIS has been deployed for navigation of mobile robot that alternates between the idiotypic network theory and the clonal selection theory when the necessary situation arises. This acts as an effective countermeasure to escape from concave-shaped obstacles like U traps and other such local minima situations. The objective of any robot navigational system is to carry the mobile robot to its target through an optimal path, without colliding with any obstacles. Navigation of mobile robot is carried out in two ways: a. Online navigation In this method, the robot is not aware of the obstacle’s position. They perceive their

& Prases K. Mohanty [email protected] 1

Department of Mechanical Engineering, National Institute of Technology, Arunachal Pradesh, Yupia, India

surroundings and locate the obstacles using sensors or cameras mounted on the mobile robot. This mode of navigation finds its application in warehouses, factories, space exploration, military service and in places beyond human reach such as extraterrestrial research. b. Offline Navigation In this method, the robot is already aware of the obstacle’s position. It then decides the optimal path and follows the same to reach its goal. Various algorithms have emerged over the years for the purpose of mobile robot navigation, some of which are heuristic and others classical (Mohanty and Parhi 2013). The heuristic algorithms include algorithms like the artificial neural network,