The Study Of Robot Society Using Habituation Algorithm

Rosafifah, Mohamad Ali (2012) The Study Of Robot Society Using Habituation Algorithm. Project Report. UTeM, Melaka, Malaysia. (Submitted)

	PDF (24 Pages) The_Study_Of_Robot_Society_Using_Habituation_Algorithm_-_24_Pages.pdf - Submitted Version Download (256kB)
	PDF (Full Text) The_Study_Of_Robot_Society_Using_Habituation_Algorithm_-_Full_Text.pdf - Submitted Version Restricted to Registered users only Download (622kB)

Abstract

There are many methods to study about multi agent robots. The problem is can a group of robots develop learning behavior that is affected by the states of environment? In order to do this, habituation algorithm is applied to these robots in order to make them imitate human social behavior. Simulation on Robot Society using Habituation Algorithm will be conducted using Matlab. As a result, a swarm of robot will adapt and learn independently with the state of environment and will learn the behavior based on equation used. Besides, robots can learn not only from environment but also from dynamic behavior of other robots. All robots are identical but their behaviors are different due to reaction from the state of environment. The method used here is Artificial Intelligence, focusing on Artificial Neural Network. There are three environments in the simulation. Two methods that are used to be tested in the environments are communication between robots and sensing of charging location. Environment 1 is tested with only communication between robots, Environment 2 is sensing of the charging location, while Environment 3 is tested with both methods. Environment with both methods show the most stable energy or survivability of robots, as the robot can charge their energy faster than the other.

Actions (login required)

View Item