Due to recent advances in miniaturisation and cost reductions of radar systems, an increasing amount of vehicles is equipped with automotive radar systems to enable advanced driver assistance systems (ADAS) which improve driving comfort and road safety. Examples of these systems include automatic braking as well as adaptive cruise control. To advance the capabilities of these systems and to take another step towards enabling autonomous driving, the next generation of radar sensors aims to enhance the situational awareness of the driver and the vehicle by, amongst others, classifying objects in the surrounding area of the car. Arguably, the most important objects to correctly classify in this process are vulnerable road users (VRUs).

With the largely completed transition of automotive radar systems from the previously used 24 GHz band to the new 77 GHz band, the performance of automotive radar systems has improved further due to larger available bandwidths and reduced wavelengths. The reduced wavelengths allow for an increased angular resolution with the same aperture size while the increased bandwidth results in enhanced range resolution. However, all currently commercially available automotive radar systems are single-polarised and therefore do not consider an important source of information, namely the polarisation state of the backscattered radiation, which can benefit classification of objects greatly.

To gain better insight in the benefits of this factor, this thesis is concerned with investigating the effectiveness of polarisation information for classification of objects in the automotive scenario. In particular, two types of objects are considered, namely road surfaces and their conditions, as well as vulnerable road users.

Chapter 2 considers the scattering of electromagnetic waves from road surfaces. Measurements on multiple different road surface materials are performed and their electrical properties are determined. These measurement results are subsequently used to simulate scattering from road surfaces at mm-wave frequencies using numerical methods. Also, a new method for computing the range-Doppler signature of road surfaces in dynamic scenarios is proposed.

In chapter 3, to accompany the numerical results from the previous chapter, a method is developed to determine statistical polarimetric radar cross section models of road surfaces at mm-wave frequencies. Measurements are performed and the radar cross section models of several types of surfaces and road surface conditions are determined. These models are subsequently used to simulate polarimetric backscattering from road surfaces as well as to determine the optimal sensing wave polarisation to be used for single-polarised radar systems.

Chapter 4 considers the classification of road surfaces and road surface conditions using H, α, and A features. For the first time, these features are adapted for road surface classification purposes in the automotive scenario and it is shown that these features can benefit surface classification.

Chapter 5 shifts the attention from classification of road surface conditions to classification of vulnerable road users. In order to do this, a novel automotive polarimetric MIMO radar system with corresponding signal processing algorithms has been developed and is presented in this chapter. This radar is subsequently used for measurements of moving VRUs and several features are considered to classify them. It is demonstrated that polarimetric information is indeed able to increase classification perfomance of VRUs compared to single polarised radar.

This thesis demonstrates that polarimetric radar is useful in the automotive scenario besides the fields in which it has already been established such as synthetic aperture radar and meteorological radar applications. The methods proposed within this dissertation can be used to enhance classification accuracy of road surface condition which may lead to improvements in road safety by for example enhancing the performance of anti-lock breaking systems. Also, the methods proposed in this thesis to use polarimetric information for classifying vulnerable road users are shown to improve classification performance of this type of objects which could help enable autonomous driving in the future. @en

Incoherent backscattering of mm-waves from natural rough surfaces is considered. A novel method is proposed to determine the statistical properties of surface scattering from range profile measurements. The method is based on modeling the road surface as a grid of uncorrelated scattering elements, described by normalized scattering matrices. Using this model, expressions are derived to estimate the mean value and covariance matrix of surface scattering from measurement data. This procedure is then applied to measurement data of four road surface types, namely: 1) dry asphalt, 2) wet asphalt, 3) asphalt covered by basalt gravel, and 4) old asphalt. Using the derived statistical normalized radar cross-section models, two novel applications are proposed. First, a procedure for synthesizing/simulating surface clutter is proposed. This procedure is subsequently used to simulate received power from surfaces comprising patches of one or multiple road surface conditions. Excellent agreement between simulation and measurement results is demonstrated. Second, a method for determining the optimal polarization of the electromagnetic sensing waves used in a single-polarized radar system is proposed. This method is based on factorizing the antenna polarization vector into two bounded parameters, allowing for numerical evaluation of the minima and maxima for targets with a specified scattering matrix. This method is further extended to work with statistical descriptions of scattering matrices by means of Monte Carlo simulations.

@en

A novel approach based on the entropy-alpha-anisotropy decomposition, also known as the $Hlpha A$ decomposition, for the recognition of road surface conditions using automotive radar is presented. To apply the $Hlpha A$ decomposition to automotive radar data, a dedicated signal processing pipeline has been developed. To investigate its effectiveness, fully polarimetric measurements of surface scattering were performed in lab conditions as well as outdoors on actual road surface material under various conditions. A high-level analysis using the Euclidean distances between cluster centroids and the standard deviations of the $H$ , $lpha $ , and $A$ features is performed, and it is shown that the proposed pipeline can provide an opportunity for classification of road surfaces, leading to enhanced road safety. Finally, the effect of neglecting the cross-polar components of the fully polarimetric measurements is considered. It is shown that in this case, the $A$ feature cannot be used anymore. Despite this, the $H$ and $lpha $ features can still be used and several road surface conditions can still be distinguished from each other at the cost of less separation between the classes, thus leading to a trade-off between classification accuracy and radar system cost/complexity.@en

The convergence of polarimetric scattering parameters and H, α and A features of road surfaces under various conditions is analysed. It is shown that the number of radar measurements used for surface classification can be traded off with accuracy of the estimation of the mean value and covariance of S-parameters and H, α and A features. Furthermore, it is shown that the H, α and A features converge at the same rate, independent of antenna orientation angles or considered road surface conditions.@en

A method for extracting fully polarimetric statistical properties of road surface radar cross sections is presented. This method is subsequently applied to extract radar cross section information from an asphalt road surface. Furthermore, an approach is introduced to synthesise the scattered signal of road surface returns as measured by a radar. The extracted statistical properties of asphalt are subsequently used in this synthesis procedure, and excellent agreement with the experimental measurement results is demonstrated.@en

A method to compute road surface signatures in the automotive scenario is presented. This method is subsequently applied to a rough asphalt surface without undulations. It is shown that, due to the observation geometry, the road surface signature experiences significant spreading along the Doppler-axis, resulting in a distinct range-Doppler map that could be used for classification purposes. This was also confirmed with an experimental recording from a car driving around the university campus. The variation of Doppler-shift and incident angle within a range bin and over a coherent processing interval is also evaluated.@en