Bit-Limited Sub-Nyquist Pulse-Doppler Radar

Pulse-Doppler (PD) radars, which prefer to operate with large bandwidth signals to attain superior range resolution, are widely used due to their superior performance in target detection and parameter estimation. However, the increasing data flow of such radar systems induced by large bandwidth presents significant challenges to signal processing and hardware implementation, particularly concerning analog-to-digital converters (ADCs), which are tasked with converting high-bandwidth inputs into digital representations at or above the Nyquist rate. In this work, we introduce a hybrid analog-digital processing architecture, which implements sub-Nyquist sampling as well as low-bit quantization, specifically tailored for PD radar applications. We refer to the designed architecture as the bit-limited sub-Nyquist pulse-Doppler radar (BiLiPD) receiver, designed with cost-efficiency in mind, utilizing low-rate and low-resolution ADCs to reduce both cost and power consumption. Specifically, the received radar echoes are acquired under the sub-Nyquist sampling framework, and then, quantized with low-resolution ADCs. The building modules of the proposed BiLiPD receiver, including the analog preprocessing, the ADCs, and the digital processing, are jointly designed to mitigate the challenges posed by sub-Nyquist sampling and low-bit quantization. We incorporate structural constraints into the design problem, formulating it under the task-based quantization framework and employing gradient descent methods for its solution. Our simulation results illustrate that the proposed BiLiPD receiver operating under low-rate low-bit constraints is capable of accurately recovering target parameters, with its target recovery performance approaching that of classic PD processing operating with unlimited resolution ADCs while notably outperforming that employing task-ignorant low-bit quantization.