Investigating Virtual Environment and Optimal Behavior of Swarm Robots
MetadataShow full item record
This research encapsulated an innovative approach of implementing A.I based behavior trees in the research of swarm robotics. This work covers the manual development of Behavior Trees and their evolution. Behavior Trees make the system intelligent, robust and flexible for various applications. A simulation based optimization was carried out in order to achieve a optimized generalized behavior tree which will perform efficiently in any environment. The final evolved tree was tested in a V-Rep based simulation environment, which was later implemented on physical swarm bot. The work in thesis wraps up with the Behavior Tree implementation on real platform and the analysis of reality gaps associated with simulation vs real environments.