Related Experiment Video
Updated: Apr 21, 2026
11:54
Microfluidic Platform with Multiplexed Electronic Detection for Spatial Tracking of Particles
Published on: March 13, 2017
10.0K
Online variational Bayesian filtering-based mobile target tracking in wireless sensor networks
Bingpeng Zhou1, Qingchun Chen2, Tiffany Jing Li3
1School of Information Science & Technology, Southwest Jiaotong University, Chengdu 610031, China. zhoubingpeng@163.com.
Sensors (Basel, Switzerland)
|November 14, 2014
Summary
This study introduces a multi-layer dynamic Bayesian network (MDBN) for mobile node tracking in wireless sensor networks (WSNs). The developed variational Bayesian filtering (VBF) algorithm enhances tracking accuracy using Received Signal Strength (RSS) data.
Area of Science:
- Wireless Sensor Networks (WSNs)
- Mobile Target Tracking
- Bayesian Inference
Background:
- Received Signal Strength (RSS) based tracking is crucial for mobile nodes in WSNs.
- Traditional Bayesian filtering struggles with nonlinear observations and time-varying RSS precision.
- Characterizing target mobility and measurement precision is essential for accurate tracking.
Purpose of the Study:
- To develop an advanced model for Received Signal Strength (RSS)-based online tracking of mobile nodes in wireless sensor networks (WSNs).
- To propose a novel algorithm capable of handling nonlinear observations and time-varying measurement precision.
- To enhance the accuracy and robustness of mobile target tracking systems.
Main Methods:
- Introduction of a multi-layer dynamic Bayesian network (MDBN) to model target mobility (directional/undirected).
- Application of the Wishart distribution to approximate time-varying RSS measurement precision.
- Development of a mean-field variational Bayesian filtering (VBF) algorithm for online tracking.
- Joint optimization of real-time velocity and its prior expectation for online velocity tracking.
Main Results:
- The MDBN model provides a generalized framework by incorporating statistical information of target movement and observations.
- The proposed VBF algorithm effectively performs online tracking despite nonlinear observations and time-varying RSS precision.
- Numerical simulations and Bayesian Cramer-Rao Lower Bound (BCRLB) analysis validate the algorithm's performance.
- Tracking accuracy demonstrates a linear scaling with expectation under time-varying RSS measurement precision.
Conclusions:
- The MDBN model and VBF algorithm offer a promising solution for online mobile node tracking in WSNs.
- Exploiting potential state information via the MDBN model significantly improves tracking capabilities.
- The developed methods enhance the reliability and accuracy of tracking systems in dynamic environments.
