Distributed Multi-Sensor Control for Multi-Target Tracking

被引：1

作者：

Blair, Aidan ^{[1
]}

Gostar, Amirali Khodadadian ^{[1
]}

Tennakoon, Ruwan ^{[1
]}

Bab-Hadiashar, Alireza ^{[1
]}

Li, Xiaodong ^{[1
]}

Palmer, Jennifer ^{[1
]}

Hoseinnezhad, Reza ^{[1
]}

机构：

[1] RMIT Univ, Melbourne, Vic, Australia

来源：

2022 11TH INTERNATIONAL CONFERENCE ON CONTROL, AUTOMATION AND INFORMATION SCIENCES (ICCAIS) | 2022年

基金：

澳大利亚研究理事会;

关键词：

distributed sensor network; sensor control; co-ordinate descent; random finite sets; multi-target tracking; MULTI-BERNOULLI FILTER; SENSOR MANAGEMENT;

D O I：

10.1109/ICCAIS56082.2022.9990364

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

This paper proposes a new sensor control algorithm for multi-target tracking applications within distributed sensor networks. In multi-target tracking applications, most sensor control algorithms are designed for centralized sensor networks, where there is a central processing node that is computationally inefficient. This paper first provides a conceptual and mathematical overview of the multi-sensor multi-target tracking framework, using random finite set (RFS) filters and sensor fusion. We will also provide an overview of the existing sensor control methods. We then explore coordinate descent-based sensor control and introduce a fully distributed algorithm utilizing coordinate descent and an information-theoretic objective function. This method is tested on synthetic data and compared to alternative methods. The results show that the proposed method outperforms equivalent independent multi-sensor control methods.

引用

页码：231 / 239

页数：9

共 34 条

[31] The Cardinality Balanced Multi-Target Multi-Bernoulli Filter and Its Implementations [J].