MZM Array-based Optical Compressive Sampling for 6G Networks

Abstract

The 6G communication network development presents an immense challenge in managing data rates that are in theorder of Terabits/sec. As it stands, our communications are based on the Digital Signal Processors that are required in dealing with data compression and data processing. These electronic chips have, however, reached the physical end where they can no longer match the incredible speed of light-based data without dissipating enormous quantities of electricity and heating up extremely. Moreover, converting light signals to electrical signals to be processed, called the O-E-O bottleneck, can cause a considerable delay or latency, which is not acceptable with real-time 6G applications such as autonomous driving and remote surgery. Our study suggests a novel approach termed All-Optical Compressive Sampling to address this. We do not digitize the light signal, but do the compression within the optical hardware with an array of Parallel Mach-Zehnder Modulators. The physics of photonic interference can be used to cut back the amount of data by splitting the high-speed laser signal into a number of parallel paths and using specific analog weights in the form of DC bias voltages to cut the data when it remains a light signal. This is a high-dimensionality data stream mapping process that can be effectively reduced into a lower-dimensionality compressed stream on-the-fly, without any slow electronic intervention that is required. The proposed architecture is tested in MATLAB using numerical simulation. The reported BER, Q-factor and compression performance is the result of simulations, and experimental validation with photonic hardware is not considered in this paper. The findings suggest that MZM array is capable of doing hardware-level dimensionality reduction that can dramatically reduce the required rate of sampling at the receiver. The proposed framework does not include repeated O-E-O processing stages is expected to decrease processing latency and significantly lower power usage than conventional digital signal processing systems. This architecture removes the electrical domain and pushes the computational load to the optical domain of operation, offering the capacity of 6G infrastructure high-capacity front-haul with high efficiency and scalability.