Fast dual control of exploration and exploitation for autonomous search of hazardous sources

This paper presents a new control algorithm designed for addressing the challenges posed by autonomous search problems, where the true source location and environment remain unknown. Utilizing a novel reward function, we formulate a fast dual control approach within the realm of optimization. This approach attains an optimal balance between exploration and exploitation by navigating the searcher towards the estimated source target while maximizing the exploration capability of control actions. The proposed algorithm not only demonstrates excellent search performance but also exhibits a high level of computational efficiency, as evidenced by two numerical examples provided for illustration.