Automatic vehicles classification approaches for WiFi-based Passive Forward Scatter Radar

Abstract

This paper shows the feasibility of a passive forward scatter radar (PFSR) based on WiFi transmissions for automatic classification of surface vehicles. To this purpose, proper automatic classification schemes are employed, able to exploit the forward scatter target signatures in the time domain. The considered approaches have been extensively tested against experimental data sets. The reported results prove that the exploited geometry yields quite stable and diverse signatures for the considered targets despite they belong to the same cars category. This results in a remarkable classification capability for the conceived sensor, thus showing the practical applicability of the WiFi-based PFSR system for surface traffic monitoring.

Stand-alone WiFi-based Passive Forward Scatter Radar sensor for vehicles classification

“…The proposed approaches are extensively tested against experimental datasets in order to verify their practical feasibility. This paper is a companion to another paper submitted to this conference [1]. Specifically, with the proposed approaches we complement and extend the results in [1] by providing an effective solution for a realistic implementation of the conceived sensor.…”

“…In a companion paper submitted to this conference [1], we have shown that a WiFi-based passive radar system can be successfully exploited for surface vehicles automatic classification by adopting a forward scatter radar geometry, i.e. an extreme bistatic configuration where the bistatic angle is close to 180°.…”

“…In fact, thanks to the "forward scattering" mechanism that is invoked to model the transmitted energy scattered by the target [2], [8], quite stable profiles are observed in the received signals when a given vehicle repeatedly crosses the transmitter (Tx) -receiver (Rx) baseline. These profiles were shown to be characteristic signatures of the specific target so that they could be easily exploited by a proper automatic classification stage [1]. However, the procedure adopted in [1] to extract the vehicles signature from the received signals yields some intrinsic limitations.…”

DVB-T-Based Passive Forward Scatter Radar: Inherent Limitations and Enabling Solutions

“…DNR=30 dB. The curves are scaled to the expected background level for each case, evaluated using (21), so that the result is representative of the expected signal-to-disturbance ratio. Apparently, progressively smaller values of 𝜂 have to be preferred as the target moves far from the baseline.…”

“…The studies reported in the open literature addressing passive forward scatter radar (PFSR) have investigated the above aspects when parasitically exploiting different illuminators of opportunity operating from Very High Frequency (VHF) to K bands [11], there including Global Navigation Satellite System (GNSS) transmitters [9][10] [13][17] [18], Global System for Mobile Communications (GSM) and Long-Term Evolution (LTE) base stations [12] [20], radio and television broadcast transmitters [14][15] [22], and WiFi access points [16] [19] [21]. Many of these studies focused on the capability of a PFSR to detect and track ground-based or aerial targets, whereas some of them have investigated the potential for improved automatic target classification [18]- [21].…”