Scene-adaptive radar tracking with deep reinforcement learning

Multi-target tracking with radars is a highly challenging problem due to detection artifacts, sensor noise, and interference sources. The traditional signal processing chain is, therefore, a complex combination of various algorithms with several tunable tracking-parameters. Usually, these are initia...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Machine learning with applications Ročník 8; s. 100284
Hlavní autori: Stephan, Michael, Servadei, Lorenzo, Arjona-Medina, José, Santra, Avik, Wille, Robert, Fischer, Georg
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier Ltd 15.06.2022
Elsevier
Predmet:
Radar tracking
Reinforcement learning
Scene adaptation
60G35
Radar tracking
68T10
68T05
Scene adaptation
Reinforcement learning
ISSN:2666-8270, 2666-8270
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Multi-target tracking with radars is a highly challenging problem due to detection artifacts, sensor noise, and interference sources. The traditional signal processing chain is, therefore, a complex combination of various algorithms with several tunable tracking-parameters. Usually, these are initially set by engineers and are independent of the scene tracked. For this reason, they are often non-optimal and generate poorly performing tracking. In this context, scene-adaptive radar processing refers to algorithms that can sense, understand and learn information related to detected targets as well as the environment and adapt its tracking-parameters to optimize the desired goal. In this paper, we propose a Deep Reinforcement Learning framework that guides the scene-adaptive choice of radar tracking-parameters towards an improved performance on multi-target tracking. •Variance-aware reward to encourage exploration and enhance tracking performance.•Transfer learning with an auxiliary task improves three-phase learning.•Scene-adaptive hyperparameter adjustment improves radar tracking performance.
ISSN:2666-8270
2666-8270
DOI:10.1016/j.mlwa.2022.100284