A Multi-Objective Optimization Approach to Robot Localization of Single and Multiple Emission Sources

摘要

A growing research axes in the field of artificial intelligence understudies the corporate behavior that emerges as a result of interactions among non-complex multi-agent robots and the potential of this behavior in providing sustainable solutions to the multi-source localization problem. While locating a single emission source or single target location has been treated extensively, there has been comparatively little emphasis on locating multiple target locations. Although some partial solutions to the general problem of localizing multiple emission sources do exist, they however are not integrated in an adhesive synergy and also lack clear recommended strategies to proceed with a search after a source/target is found. The missing piece of the multisource localization puzzle is the absence of theoretical foundations for progress (after a source is found), convergence, and termination of the search operational algorithm. Cooperation among multi-agents could be active (agents acknowledging each other) or inactive (agents oblivious of each other). In this paper, the authors investigate the potency of multi-objective optimization on a machine learning algorithm (the artificial neural network) for a swarm of robots by leveraging on their basic yaw and thrust actuators in order to achieve collaborative control (group behavior) amidst localization of multiple emission sources. The empirical results from our simulations are promising.