Automotive RADAR Sensor Modeling with Multi-Scale Electromagnetic Field Solvers

RADAR Sensors are going to be an integral part of autonomous vehicles. One of the main objectives of these sensors in autonomous vehicles is to get the Doppler range profile for surrounding traffic. In this paper, we use a similar RADAR for ground speed sensing in the off-highway scenario. There are several challenges in integrating the RADAR sensor with vehicles such as sensor position from ground, location on vehicle, electromagnetic interference with other electronic devices, enclosure design etc. Ground conditions and properties are also critical in the off-highway scenario for speed sensing. We propose to use the physics based electromagnetic field solvers to understand and mitigate some of these challenges and speed up the design. Electromagnetic field solvers tend to scale poorly with distance of propagation, especially in 3D modeling. The computational cost increases dramatically, when wave propagation over multiple wavelengths is desired with FEM type solvers due to constraints on spatial resolution. The time scales are also extremely wide, considering the wave oscillation time period (~ ps) and vehicle speed (~s). These challenges can be overcome by switching to the multi-scale solvers paradigm. In this paper, we discuss how these solvers work and how they can be used to model the ground speed sensing application in the off-highway scenario. We present critical aspects, which solvers should have to capture the relevant physics. We further discuss about effective modeling of ground conditions which stay true to underlying physics in the multi-scale solver paradigm. Initial results are presented which demonstrate the applicability of these methods.