Collaborative Positioning Mechanism Using Bayesian Probabilistic Models for Industry Verticals

In this paper, we develop a collaborative positioning mechanism which uses Bayesian probabilistic models to combine multidimensional sensory data and localize target nodes over the network deployment area. Herein, heterogeneous anchor nodes with distinct radio access technologies and experiencing various radio channel features implement a joint sensor fusion and positioning system for industry verticals. The proposed mechanism also relies on a modern network architecture whereby devices offload high-demand computation to more capable edge servers which then estimate the target node position after gathering anchors measurements and prior history. Kernel density estimation results are used to show that edge servers implementing Bayesian-based sensor fusion and positioning system effectively estimate the target node location when using hybrid metrics and combining past and current sensory inputs.