Our team develops silicon photonics technologies in communications, computing, and sensing. In communications, we improve bandwidth and yield with wavelength division multiplexing and demultiplexing. In computing, we enhance parallel optical processing through broadband convolutional computing, demonstrating flexible multiply–accumulate operations. In sensing, we create a compact, low-cost silicon photonic fiber-optic gyroscope with tactical-grade performance.

We developed silicon photonic WDMs in O-band, achieving crosstalk below –30 dB and wavelength shift within ±1 nm. Combined with broadband light sources, these devices enable optical convolutional computing, demonstrating high-accuracy inference in MNIST digit recognition. We also realized a silicon photonic FOG that achieves bias instability below 0.1 deg/hr, offering low cost and compact size.

WDMs and demultiplexers support high-speed communications and data center interconnects, improving bandwidth efficiency. Broadband light–enabled convolution accelerates low-power edge AI tasks such as image recognition and sensing. The miniaturized silicon photonic gyroscope provides precise navigation for UAVs, autonomous vehicles, and satellites, with strong potential for commercialization.