Abstract:A novel method named DS‐UPF is developed to improve the accuracy of target state estimation in a distributed multi‐station system. The suggested weighted Dempster–Shafer theory and unscented particle filtering effectively address challenges such as signal‐to‐noise ratio bias and particle degradation, which enhance fusion centre performance without demanding prior knowledge of local interference. In a distributed multi‐station system, the observations received by local radar nodes for a single target will have a large signal‐to‐noise ratio (SNR) bias due to inconsistent radar cross‐sections from distinct angles, different distances from the target, various local interference such as harsh weather, and dissimilar background noise. Integrating heterogeneous information in dynamic and uncertain environments can be challenging for the fusion centre. Moreover, the particles in the basic particle filter (PF) may degrade after many iterations, making it difficult to achieve accurate target state estimation in the local tracking process. To address these issues, the authors propose a novel method named DS‐UPF based on the Dempster–Shafer (DS) theory and the unscented particle filter (UPF). By updating the important density function, the UPF efficiently suppresses particle degradation. The weighted Basic Probability Assignments (BPAs) are proposed and integrated under the new synthesis formula. The weight‐modified DS method restrains the impact of significant local estimation errors on weighted BPAs fusion result, improving robustness without local interference prior knowledge. The experimental results demonstrate that the DS‐UPF outperforms the unscented Kalman filter, PF, and UPF in tracking tasks under various local interference. This indicates that the proposed algorithm can improve estimation precision in dynamic and uncertain environments.

Unscented Particle Double Layer Filter.

An Improved Unscented Particle Filter with Global Sampling Strategy.

The ensemble unscented particle filter

Research of STUPF Filter Algorithm on GPS/SINS

A Particle Filter Algorithm Based on a Spherical Simplex Unscented Kalman Filter

Fast Gaussian particle filtering algorithm

Novel Roughening Method for Reentry Vehicle Tracking Using Particle Filter

Improved Unscented Particle Filter for Nonlinear Bayesian Estimation

Improved Particle Filter Based on Fine Resampling Algorithm

Robust Information Unscented Particle Filter Based on M-estimate

Adaptive Particle Filtration for State Estimation

Threshold-Based Resampling for High-Speed Particle Phd Filter

Particle Filter Based on Iterated Importance Density Function and Parallel Resampling

A Fine Resampling Algorithm for General Particle Filters

Research on SSUPF Algorithm Based on Particle Swarm Optimization

Particle Filter with Fine Resampling for Bearings-Only Tracking

Federated unscented particle filtering algorithm for SINS/CNS/GPS system

Research On Particle Filter Based On Spherical Unscented Transformation

An adaptive unscented particle filter for a nonlinear fractional-order system with unknown fractional-order and unknown parameters

Distributed multi‐station target tracking based on unscented particle filter and Dempster‐Shafer theory

Improved design of proposal distribution for particle filter