In this paper, an advanced distributed exploration strategy of multiple agents based on one of the supervised learning algorithms, K-Nearest Neighbor (KNN), is proposed This strategy is based on the frontier-based exploration method and is used to make decisions in the process of assigning the next exploration node. Through KNN, the next exploration node can be assigned to a suitable agent in consideration of distance and exploration tendencies. The proposed strategy is compared with the performance of the Voronoi method through various indices such as total exploration time, agent's total moving distance, and exploration coverage. Through numerical simulation, the proposed strategy shows better performance than the Voronoi method.
