Search for: All records

Artificial intelligence and neuromorphic computing driven by neural networks has enabled many applications. Software implementations of neural networks on electronic platforms are limited in speed and energy efficiency. Neuromorphic photonics aims to build processors in which optical hardware mimic neural networks in the brain. 

Neuromorphic photonics exploit optical device physics for neuron models, and optical interconnects for distributed, parallel, and analog processing for high-bandwidth, low-latency and low switching energy applications in artificial intelligence and neuromorphic computing. 

Active control of interference is necessary with increased cell density, more complicated environmental reflections, and coexistence of multiple networks for next-generation wireless communications. The existing radio receiver architectures for spatial interference cancellation (SpICa) are limited by the spatial nulls created by a phased-antenna array (PAA) and cannot cover wide modulated bandwidths (BWs). We propose a discrete-time-delay-compensating technique for canceling spatial interferences with wide modulated BWs to reduce the dynamic range requirement for the data converter. Integral to the proposed circuit is a switched-capacitor-based multiply-and-accumulate processor that incorporates a reconfigurable phase interpolator and time interleaver for precise digitally tunable delays and multiplication of the input signal to an orthogonal matrix. The digital time interleaver enables 5-ps resolution with a reconfigurable range up to 15 ns. The measured results demonstrate greater than 35-dB SpICa over 80-MHz modulated BWs in the 65-nm CMOS with 52 mW of power consumption.