Consensus-based Distributed Variational Multi-object Tracker in Multi-Sensor Network

TL;DR

This paper introduces a distributed variational Bayesian multi-target tracking method in sensor networks that matches the accuracy of centralized approaches and outperforms simpler distributed fusion strategies.

Contribution

It proposes a novel distributed variational Bayesian tracker using consensus algorithms, achieving centralized-level accuracy with only local communication.

Findings

Distributed tracker performs on par with centralized tracker in accuracy.

Proposed method outperforms suboptimal distributed sensor fusion strategies.

Simulation results validate the effectiveness of the consensus-based approach.

Abstract

The growing need for accurate and reliable tracking systems has driven significant progress in sensor fusion and object tracking techniques. In this paper, we design two variational Bayesian trackers that effectively track multiple targets in cluttered environments within a sensor network. We first present a centralised sensor fusion scheme, which involves transmitting sensor data to a fusion center. Then, we develop a distributed version leveraging the average consensus algorithm, which is theoretically equivalent to the centralised sensor fusion tracker and requires only local message passing with neighbouring sensors. In addition, we empirically verify that our proposed distributed variational tracker performs on par with the centralised version with equal tracking accuracy. Simulation results show that our distributed multi-target tracker outperforms the suboptimal distributed sensor fusion strategy that fuses each sensor's posterior based on arithmetic sensor fusion and an average consensus strategy.