DPFTT: Distributed Particle Filter for Target Tracking in the Internet of Things

A novel distributed particle filter algorithm for target tracking is proposed in this paper. It uses new metrics and addresses the measurement uncertainty problem by adapting the particle filter to environmental changes and estimating the kinematic (motion-related) parameters of the target. The aim is to calculate the distance between the Gaussian-distributed probability densities of kinematic data and to generate the optimal distribution that maximizes the precision. The proposed data fusion method can be used in several smart environments and Internet of Things (IoT) applications that call for target tracking, such as smart building applications, security surveillance, smart healthcare, and intelligent transportation, to mention a few. The diverse estimation techniques were compared with the state-of-the-art solutions by measuring the estimation root mean square error in different settings under different conditions, including high-noise environments. The simulation results show that the proposed algorithm is scalable and outperforms the standard particle filter, the improved particle filter based on KLD, and the consensus-based particle filter algorithm.