Multi‐task tracking and classification with an adaptive radar

Abstract

This study investigates the use of adaptive radar waveforms in performing a multi-task operation involving the tracking and classification of a single target. The work applies a previously reported hierarchical fully adaptive radar framework and considers a situation in which the target tracking function is performed using a transmit waveform having adaptive pulse repetition frequency, pulse width, bandwidth, and coherent processing interval parameters, while the classification function can be performed using only a fixed radar waveform. Two alternate strategies for selecting the next waveform are then described. The first strategy is focussed on maintaining fixed measurement performance goals in terms of target range and velocity and classification entropy, while the second approach varies these performance goals depending on the target state. Both approaches are examined for simulated datasets at W-band representing a moving human target, as well as for laboratory measurements of this scenario. The results for both simulation and experiment show the benefits of including adaptive waveforms in performing multiple radar tasks.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Memory‐based deep reinforcement learning for cognitive radar target tracking waveform resource management

“…The cognitive ability of the receiver is just as crucial as waveform diversity in cognitive transmission, which can prevent interference energy from aliasing on the target echo and improve the radar anti-jamming and anti-clutter capabilities [9][10][11]. At the system level, cognitive intelligence is reflected in CR resource management, which conducts multi-level scheduling of radar tasks and system resources based on perception information to realise efficient allocation of transmitting resources [2,[12][13][14][15]. To deal with the external dynamic environment, the Observation-Orientation-Decision-Action (OODA) loop is implemented by perceiving, processing, scheduling, and decision-making regarding multiple transmitting resources.…”

A joint model and data driven track segment association algorithm for manoeuvring target tracking

“…When a target track is interrupted by mistake, the military sensor, such as radar or sonar, will renumber the new‐born track and a single target may generate multiple track identities (IDs) during surveillance. A large number of redundant track IDs not only increases the computational complexity and the storage requirements, but more importantly, misleads the tactical centre into forming a wrong threat assessment [5].…”