Hierarchical RSS-Based Indoor Positioning Using a Markov Random Field Model

Locating indoor object's position is a fundamental application in many industries. As a low cost and universal implementation in wireless sensor networks (WSN), received signal strength (RSS) based positioning faces challenges of interfered measurements introduced by multiple sources. In order to improve the location prediction accuracy, we proposed a Markov random field model for indoor positioning applications. A conditional distribution is adopted to quantify the RSS measurement quality used in prediction. A hierarchical algorithm is presented to lower the computational complexity. Experiments illustrated that the proposed approach rendered promising positioning accuracy.