Research on a Particle Filtering Multi-Target Tracking Algorithm for Distributed Systems

The growth of unmanned aerial vehicle applications in the low-altitude economy demand advanced multi-target tracking systems. Unlike traditional approaches that assume independent measurements, distributed systems generate coupled measurements containing additional target relationship information. T...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Sensors (Basel, Switzerland) Jg. 25; H. 11; S. 3495
Hauptverfasser: Han, Bing, Ge, Zilong, Su, Zhigang, Hao, Jingtang
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Switzerland MDPI AG 31.05.2025
MDPI
Schlagworte:
Accuracy
Algorithms
Analysis
Cost control
coupled measurements
data fusion preprocessing
Distributed processing (Computers)
Drone aircraft
Efficiency
Innovations
Kalman filtering
Linear programming
Methods
multi-objective tracking
Noise
optimal particle weights
Optimization
Sensors
Surveillance
Tracking systems
Unmanned aerial vehicles
China
coupled measurements
data fusion preprocessing
multi-objective tracking
optimal particle weights
ISSN:1424-8220, 1424-8220
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The growth of unmanned aerial vehicle applications in the low-altitude economy demand advanced multi-target tracking systems. Unlike traditional approaches that assume independent measurements, distributed systems generate coupled measurements containing additional target relationship information. This paper proposes a novel distributed particle filtering algorithm through introducing the coupled measurement into the conventional particle filtering method. In the proposed method, we fuse direct and coupled measurements via optimization and then build a cost function to optimize the particle weights. Comparative evaluations across motion models, noise levels, and the number of targets demonstrate the outperforming performance of the proposed method compared to conventional particle filtering and the unscented Kalman filtering algorithm, with more than 7% accuracy improvement over baselines. The results prove particular robustness to measurement noise and the increasing number of targets.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:1424-8220
1424-8220
DOI:10.3390/s25113495