Road Surface Recognition at mm-Wavelengths Using a Polarimetric Radar

Vassilev, Vessen, 1969 (author): Chalmers tekniska högskola,Chalmers University of Technology

(creator_code:org_t)

2022
2022
English.
In: IEEE Transactions on Intelligent Transportation Systems. - 1524-9050 .- 1558-0016. ; 23:7, s. 6985 -6990

Related links:: https://research.cha... (primary) (free); show more...; https://research.cha...; https://doi.org/10.1...; show less...

Abstract Subject headings

We demonstrate detection of ice formations on a road surface using a polarimetric radar operating at 87.5-92.5 GHz. The radar measures the scattering parameters of the surface at horizontal and vertical polarizations and their cross-polarization components. We demonstrate detection of ice for radar beam directed at up to 45⁰ angle of incidence with respect to the surface which allows for road surface characterization in front of a vehicle. The method used is based on a statistical approach where the 2-port scattering parameters are measured multiple times and used to calculate an average scatter coherence matrix representing the surface. The coherence matrix is then decomposed to eigenvalues/vectors, which are used to estimate polarimetric attributes such as target entropy(degree of randomness) and polarimetric pedestal (degree of depolarization). Through measurements of dry, ice-covered and wet road surfaces, we show that both entropy and depolarization are increased with respect to dry surface when a thin ice layer is formed, while their value decrease for the case of wet surface. It is also shown that these polarimetric attributes are not sensitive to surface roughness in dry conditions, minimizing the probability of false alarm due to road surface wear.

About the subject

ENGINEERING AND TECHNOLOGY: ENGINEERING AND ...; and Civil Engineerin ...; and Infrastructure E ...

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.