Multipath Detection and Reflection Surface Estimation in Automotive Scenarios

Automotive radar plays an immense role for future driver assistance functions and autonomous driving. Complex wave propagation environments have a significant ef-fect on the radar's received signal which may lead to possible performance degradation. Especially multipath propagation from specular surfaces may lead to ghost targets and possible system false alarms. Mitigation methods can be applied on sample, detection or object level. The approach taken within this paper is an extended signal model targeting to detect multipath situations using several range-Doppler cells using samples. After the proposed algorithm detected a multipath situation, the reflection point as well as the specular surface can be estimated. Examples showing the performance on real world data are provided.