Search Target Swarm Robot Algorithm

Chai, Guan Chi (2019) Search Target Swarm Robot Algorithm. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text ChaiGuanChi.pdf Restricted to Registered users only Download (2MB)

Abstract

In this project, the aim is to provide a better background research and scenario result for further study with different kind of searching algorithm, integrated with obstacle avoidance method target seeking method, and also communication method. This project proposed a method that involve of spiral pattern search with Bug2 obstacle avoidance method, light sensor of target seeking, and radio frequency communication method. The method will be tested with different algorithm such as Random search algorithm, Region Random search algorithm and also Particle Swarm Optimization search algorithm with several scenario from previous project and been modify to fit the real life scenario. Three different characteristic experiments based on the objectives have been carried out in this project, which consist of to test the efficiency of each algorithms in term of time of all targets found, optimized the number of robots in same experiment setting, and effectiveness of each algorithm in term of targets found in limited time. Total 112 experiments is carried out in this project, each experiment will be carried out with 4 different type of obstacle arrangement with 3 times to get the average time. The result of this project shows the spiral search algorithm used up the lesser time and constantly found targets in the environment while other used up more time and easy to lost target in the given time limit. The optimized number of robots for spiral and region random is found to be 3 robots in the same experimental settings. In a nutshell, this project achieved all the objectives and prove the efficient and effective of spiral compare to all others algorithms mentioned above in term of full coverage area, time constraint and obstacle avoidance. Webots simulation is limited to perform robot fully random movement and also unable to perform data centralized processing. The result of the project is used to contribute the research of comparison between pattern search and random search and solution to improve the search operation regardless the incident mentioned above an earthquake which happened in Mexico.