A Low-Complexity Radar Detector Outperforming OS-CFAR for Indoor Drone Obstacle Avoidance

As radar sensors are being miniaturized,there is a growing interest for using them in indoor sensing applications such as drone obstacle avoidance. In those novel scenarios,radars must perform well dense scenes with large number of neighboring scatterers. Central to performance the detection algorithm used separate targets from background noise and clutter. Traditionally,most systems use conventional CFAR detectors but their degrades scenarios many reflectors. Inspired by advances non-linear target detection,we propose high-performance,yet low-complexity detector we experimentally validate our on dataset acquired mounted drone. We show that proposed drastically outperforms OS-CFAR (standard automotive systems) specific task navigation more than 19% higher probability given false alarm. also benchmark against recently multi-target an improvement 16% compared CHA-CFAR,with even larger improvements both OR-CFAR TS-LNCFAR particular scenario. To best knowledge,this work improves state art high-performance yet critical applications.