Adaptive Microwave Photonic Multi-Coset Sampling With Optimized Sampling Pattern

An adaptive microwave photonic multi-coset sampling (MCS) method is proposed to enable accurate and efficient sub-Nyquist acquisition of multiband radio frequency (RF) signals. The method integrates an activated Nyquist band sensor and a sampling pattern optimization module to dynamically optimize the sampling pattern based on prior spectral knowledge. The activated Nyquist band sensor can efficiently estimate the prior spectral support. A fine microwave frequency-sweeping technique is introduced to characterize both the sampling pattern and the measurement matrix, enabling precise calibration and control of the relative time delays in the sampling channels. A four-channel, 1-GSa/s microwave-photonic MCS system is implemented and experimentally validated. Multiband signals spanning four and eight distinct Nyquist bands across 0.01–40 GHz with a cumulative bandwidth of 2 GHz are adaptively sampled and successfully reconstructed. The proposed method demonstrates the ability to achieve the theoretical minimum sampling rate while maintaining a well-conditioned measurement matrix through optimized sampling pattern design. The experimental results further reveal that the system is robust to spectral aliasing and capable of resolving prealiasing components under limited sampling channels. This work represents, to the best of our knowledge, the first hardware demonstration of an adaptive microwave photonic multi-coset sampler with full-band sensing and sampling pattern control. Owing to its simple architecture, low sampling rate, and adaptability, the proposed system offers a promising solution for next-generation multiband RF front ends, particularly in applications such as multiband radar, wideband spectrum sensing, and cognitive radio.