The development of swarm-based exploration algorithm with the expanded square pattern using quadcopter

Abstract

Exploration algorithm is one of the most important roles in searching mechanism. In robotics field, exploration algorithm deals with the implementation of the robot to enlarge the information over a particular environment. In other words, the implementation of exploration algorithm into the robot is intended to survey the situation or condition of a specific area. Based on that comprehension, exploration is applicable to various field such as search and rescue, monitoring conservation, scientific space exploration, etc. Although the field of exploration algorithm on robotic has become a major research area and been studied since the 1950s, the exploration problem has always been an interesting topic for investigation. A variety of techniques has been developed, even the biological systems have also become an inspiration to be reckoned. In this thesis, we propose a swarm-based exploration algorithm with the expanded square pattern using the quadcopter to explore an unknown area. In this algorithm, the expanded square pattern is conducted by a series of the distance around a fixed reference point. We simulate the swarm-based exploration algorithm with the expanded square pattern in the VREP simulator. The existing exploration algorithms namely, the frontier baseline and the cellular automata are also simulated to be compared with the proposed algorithm. All algorithms are simulated with the same setup. In order to analyse and evaluate the performance of all algorithms, the data of the simulation are documented. Some comparisons are conducted such as the performance of all algorithms, the performance of a group of the quadcopter, the covered spaces and the cooperation among groups. According to the simulation results, the swarm-based exploration algorithm with the expanded square pattern can explore better and faster compared to the frontier baseline and the cellular automata as the number of robots increased. This is supported by the statistical analysis that is conducted at the end of this research.