Statistical Radar Cross Section Characterization for Indoor Factory Targets

TL;DR

This paper statistically analyzes the radar cross section (RCS) of various indoor factory targets, including drones, humans, and robots, using different motion states and distribution fitting to understand their radar signatures.

Contribution

It provides the first comprehensive statistical RCS characterization of indoor factory targets considering various motions and fits multiple distributions to the data.

Findings

Lognormal distribution fits all target RCS data

RCS varies significantly with target motion and orientation

Different targets exhibit distinct RCS statistical behaviors

Abstract

In this work, we statistically analyze the radar cross section (RCS) of different test targets present in an indoor factory (InF) scenario specified by 3rd Generation Partnership Project considering bistatic configuration. The test targets that we consider are drones, humans, quadruped robot and a robotic arm. We consider two drones of different sizes and five human subjects for RCS characterization. For the drones, we measure the RCS when they are are flying over a given point and while they are rotating over the same point. For human subjects, we measure the RCS while standing still, sitting still and walking. For quadruped robot and robotic arm, we consider a continuous random motion emulating different tasks which they are supposed to perfom in typical InF scenario. We employ different distributions, such as Normal, Lognormal, Gamma, Rician, Weibull, Rayleigh and Exponential to fit the measurement data. From the statistical analysis, we gather that Lognormal distribution can fit all the considered targets in the InF scenario.