Classifying Guardrail System Radar Signatures using Full Physics
Simulation for 77 GHz Automotive Radar
Automotive radar is one of the key sensor technologies for active safety
and comfort advanced driver assistance systems(ADAS). Vehicles equipped
with radar sensors can determine the range, velocity and angle of
arrival of multiple targets simultaneously in a highly dynamic
environment. At 77 GHz, road infrastructure and buildings are an ever
present source of clutter that can affect crucial target detection.
Guardrails present a unique clutter challenge due to their ubiquity,
proximity to ego vehicle and extremely large radar cross section(RCS).
Due to their large RCS, guardrails can mask the existence of soft
targets such as pedestrians in their vicinity. Therefore, it is crucial
for sensor perception algorithms to identify and filter out the effects
of guardrails. This paper presents a full-physics, simulation-based
study of several full-scale road traffic scenes with different guardrail
arrangements. By studying the Range-Doppler(RD) plots of each of the
scenes at 77 GHz, we demonstrate the distinctly different radar
signatures of guardrails in four key road settings that normally occur
in driving. Using the results from this study, we characterize both the
range and velocity behavior of various guardrail sections. Results from
this study can be used to train perception algorithms to accurately
identify and filter out guardrail systems in different driving scenarios
and thus potentially prevent future accidents.