Gerd Wanielik

Polarimetric road condition classification and data visualisation

The automobile industry has been continuously involved in efforts to improve driving conditions and comfort. Therefore Daimler Benz is interested in finding new techniques to monitor the road traffic conditions in order to make the information available for other electronic on-board systems. For this purpose a full polarimetric, incoherent microwave radar system for Mueller-matrix measurements was designed. It works at 76 GHz, employing 5 Impatt oscillators and 7 Shottky detectors. Special software tools for classification and visualisation of the acquired data has been developed. The paper introduces this system and describes its functionality as well as the concept of the data processing software. Measurement results are shown.

Millimeter-wave imaging of traffic scenarios

Methods for the automatic interpretation of complex road traffic scenarios are currently developed on the basis of real measured data. The measurements are performed using a fully polarimetric instrumentation radar at 76 GHz. Results of those measurements will be presented, especially with respect to automatic image interpretation. The features of the measurement system are very similar to a prospective automotive radar with the exception of possible Doppler range and physical dimensions. In parallel to the measurement and algorithmic activities, a radar sensor for automotive applications using monolithic integrated millimeter-wave circuits (MMIC) is under development. Real aperture imaging is achieved with a focal plane multibeam array. The most important design criteria are low cost and compactness. System aspects of this radar sensor will be discussed including results of the millimeter-wave circuits.

Object detection in traffic scenes by a colour video and radar data fusion approach

Object detection is one of the key functions in autonomous driving. For this purpose different sensor types-such as laser or millimeter-wave (MMW) radar-are in use but most systems are solely based on vision (Thomanek et al., 1994). In contrast, this paper presents a data fusion approach for joint radar video object detection.

Superposition of polarized waves at layered media: theoretical modeling and measurement

The detection of ice layers on road surfaces is a crucial requirement for a system that is designed to warn vehicle drivers of hazardous road conditions. In the millimeter wave regime at 76 GHz the dielectric constant of ice and conventional road surface materials (i.e. asphalt, concrete) is found to be nearly similar. Thus, if the layer of ice is very thin and thus is of the same shape of roughness as the underlying road surface it cannot be securely detected using conventional algorithmic approaches. The method introduced in this paper extents and applies the theoretical work of Pancharatnam on the superposition of polarized waves. The projection of the Stokes vectors onto the Poincare sphere traces a circle due to the variation of the thickness of the ice layer. The paper presents a method that utilizes the concept of wave superposition to detect this trace even if it is corrupted by stochastic variation due to rough surface scattering. Measurement results taken under real traffic conditions prove the validity of the proposed algorithms. Classification results are presented and the results discussed.