The Power of Many: Multi-User Collaborative Indoor Localization for Boosting Standalone User-Based Systems in Different Scenarios

In today's smartphone-driven era, our lives revolve around smartphones, resulting in high smartphone densities in various contexts. Furthermore, modern smartphones can exchange inter-user measurements via different sources, such as WiFi and BLE. In turn, the concept of online multi-user collaboration has recently emerged as a promising solution to improve the performance of standalone user-based indoor positioning systems (SU-IPSs). On this basis, this research proposes a real-time multi-user collaborative indoor positioning (RT-MUCIP) scheme. This scheme aims at boosting the indoor positioning performance of users with unavailable position information or low position confidence. The procedures of the proposed scheme can be summarized as follows: First, it checks the density and position confidence of the surrounding users. Users with high position confidence are identified. Subsequently, the proposed scheme adapts to the changing density of users in the following manner: In scenarios with sparse users, a nearby user with high confidence is explored and exploited to boost a near-neighbor with low position confidence. In areas with dense users, the weight of surrounding users' positions and inter-user measurements are determined, and the RT-MUCIP solution is estimated using a weighted non-linear least squares algorithm. Additionally, inspired by wireless sensor networks, RT-MUCIP scheme proposed method to upgrade users observed in static mode with high positioning confidence to act as temporary anchor points. As a result of this upgrade, anchor node density increases, and overall positioning performance can be improved. To evaluate the performance of the proposed scheme, several tests were conducted in three scenarios. In light of the tests results, we can conclude that the proposed collaborative localization scheme can improve the localization accuracy of collaborated users without the need to use external resources.