Misspecified Cramér-Rao bound for Terahertz automotive radar range estimation

Obtaining detailed information regarding the surrounding environment of a host vehicle is crucial for the operation of autonomous driving and advanced driver assistance systems (ADAS). This necessitates radar systems equipped with high-resolution capabilities in range, Doppler, azimuth, and elevation. Terahertz (THz) frequency band radar systems offer the potential for nearly-optical resolution while being compact and lightweight. Nevertheless, the frequency-selective range-dependent attenuation of the THz propagation channel poses challenges to implementing automotive radars operating within this frequency band. The computational complexity of estimating the parameters of the THz channel adds to the challenge. Hence, automotive radar systems often adopt a simplified flat channel model for THz radar range estimation.

This study delves into the implications of this assumption on THz radar range performance estimation through the misspecified Crame'r-Rao bound (MCRB). Simulations are conducted to assess the ability of the proposed approach in evaluating the impact of frequency-selective misspecification on radar range estimation accuracy and detection range.