Phenomenological Modeling of Millimeter-Wave Automotive Radar

Automotive radars are an increasingly essential element in advanced driving assistance systems (ADAS) in order to make autonomous driving feasible. The automotive market segment is highly sensitive to price. This imposes additional constraints on the development of robust and reliable vehicular sensors. In order to ensure a high quality in testing and yet minimize the time-to-market, virtual and semi-virtual testing have gained more currency lately. It allows safe testing while including adverse conditions and enables low-cost evaluation of new experimental radar architectures. In this paper, we introduce and validate a phenomenological radar sensor simulation at millimeter wave that allows for virtual performance analysis of an automotive radar system in realistic environmental conditions. In particular, we extract effects from adverse weather conditions from field measurements and translate their characteristics to stimuli-generating models. Finally, we directly evaluate our radar simulation via state-of-the-art noise-waveform-based automotive radar with real interference signals.