Exploiting semantic knowledge in swarm robotic systems for target searching

Robotic systems have long been used for search and rescue tasks in hazardous environments. The prevailing solutions which utilize delicate units for sensing and positioning show their reliance on globalized information when multiple robots are deployed. To employ multiple robots (especially swarm robots in this thesis) in a searching task, the local perceptual ability and local communication range demand a new strategy for environmental information recording and exchanging, to promote searching efficiencies of the robots.

This thesis presents a semantic knowledge-based mechanism for environmental information storage and communication in swarm robotic systems. Human expert knowledge about the environment can be utilized by such a mechanism for promoting searching efficiency. Robots without the knowledge provided in advance could learn knowledge in a task-oriented way, and help other robots in the swarm find the target faster by sharing the knowledge.