Reward-based evolutionary swarm UAVs on search and rescue mission

Search and rescue missions (SARs) are challenging in a harsh environment especially on a sea surface. Currently, swarm robotics incorporating group of autonomous aerial vehicle (UAVs) can cover greater search area than one autonomous vehicle more efficiently if correctly programmed. An ongoing study applying a rewards-based system to an evolutionary swarm (ES) to enhance its searching capability in a harsh environment is reported in this paper. However, the evolutionary method utilising fitness function to tackle tasks in such environments often gives a deceptive result, because the predefined objective function is fixed. The fitness function for the evolutionary mechanism should not be fixed for a useful result to be obtained. In this study a rewards-based system is used to investigate a method to overcome these deceptive results. Results and comparison between fitness-based and rewards-based ES on SAR mission are discussed in this paper and preliminary results reported on.