Dynamic Programming Based Track-Before-Detect Methods for Airborne Radar System
Feilong Ku,Wei Yi,Ming Wen,Jinghe Wang,Suqi Li,Lingjiang Kong
DOI: https://doi.org/10.1109/radar.2019.8835536
2019-01-01
Abstract:Benefited from the flexibility of the aircraft platform, airborne radar shows prominent advantages such as good visibility and excellent maneuverability. In military applications, it is of great significance for airborne radar to detect and track low signal-to-noise ratio (SNR) targets, while it is intractable to traditional methods. Dynamic programming based Track-Before-Detect (DP-TBD) approach can achieve superior performance for dim targets by jointly processing multiple frames of raw data, which relies on a finite and discrete state space. For DP-TBD in ground-based radar, the state space is demarcated by the boundary of scanning area and quantized according to the criterion of discrete measurements. However, the motion of aircraft causes the ego-motion of scanning area for airborne radar system. Consequently, the demarcation and discretization strategy of ground-based radar cannot be directly applied to airborne radar because the criteria of discretization do not match each other in different frames. In this paper, an intuitive solution is presented firstly, it demarcates the state space to merely cover all scanning areas and quantizes the space by the criterion of measurements adopted in the first frame. After a thorough analysis of the previous algorithm, another method is proposed, which creates a "mapped" state space changes with scanning area to make the ground-based strategy working on airborne radar. Simulation examples demonstrate the implementation processes of the algorithms. Comparison between two algorithms and advantage in detection is also illustrated.