Misspecified Cramér-Rao Bound for Multipath Model in MIMO Radar

Automotive radars need to provide reliable information on the host vehicle's surroundings in all operating conditions, including dense urban scenarios. Urban scenarios are characterised by the presence of multiple man-made flat surfaces. The multipath phenomenon occurs when radar operates near smooth flat surfaces. It can significantly degrade the radar performance in terms of the probability of false-alarm and direction-of-arrival (DOA) estimation accuracy. In many practical scenarios, the direct path and multipath DOAs are unresolvable, inducing bias in target DOA estimation. The conventional Cramér-Rao bound (CRB) is unable to predict this misspecification phenomenon and, therefore, is not useful. This work investigates the multipath effect on the multiple-input multiple-output (MIMO) radar DOA estimation performance by evaluating the misspecified CRB (MCRB). The ability of the derived MCRB to evaluate the multipath effect on the MIMO radar DOA estimation performance is investigated via simulations of the typical automotive scenarios.