Small, inexpensive radars can be effectively used as a part of a pre-crash warning system. To facilitate this, an engineering tool has been developed to predict sensor performance as a function of sensor location, update rate, transmitted power, antenna beam widths and a variety of other system characteristics. A performance model for a single radar has been developed from a statistical analysis of measurements of fundamental properties including propagation through concealing fascia panels, scattering from target vehicles, and system noise levels. Some assumptions about the random scattered signal permit a statistically optimal detection algorithm to be defined. The parameters of that model are related to the radar's design parameters, which permit the expected radar performance (i.e., probability of detection and false alarm rate) to be estimated as a function of those parameters.
Abstract