A Data Mining Approach for Indoor Navigation Systems in IoT Scenarios

Indoor positioning is an important aspect in internet of things which plays a crucial role in various scenarios. Meanwhile, depending on the scenario, Indoor Navigation Systems (INS) generates considerable amount of data which can be used in navigation as a data-driven approach. In this paper, a navigation method has been proposed based on data mining techniques which enables value added services for end-users in different IoT scenarios such as airports, shopping mall and hospitals. The proposed heuristic algorithm is based on combining the greedy and random forest algorithms. Toward that end, we have collected data from a real world scenario (a large hospital) and used them to our implementation. According to our results regarding passage of time and accumulated history in the central server, we were able to suggest better routes while the proposed method shows reduction in both traveled distance and elapsed time. It also improved routing by decreasing the number of turns and encounters with obstacles. Thus, the proposed method provides better solution as an intelligent indoor navigation.