Micro-ring resonator thermal stabilization system based on neural network and tuning method

Abstract

The invention relates to a micro-ring resonator thermal stabilization system based on a neural network and a tuning method thereof, the micro-ring resonator thermal stabilization system comprises a micro-ring resonator, a proportional to absolute temperature (PTAT) temperature sensor, a neural network processor and a micro heater, the micro-ring resonator is connected with the PTAT temperature sensor, the neural network processor is connected with the micro-ring resonator, the micro-ring resonator is connected with the micro-ring resonator, and the micro-ring resonator is connected with the micro-ring resonator. The PTAT temperature sensor is connected with the neural network processor, the neural network processor is connected with the micro-heater through the feedback loop, and the micro-heater is arranged below a coupler of the micro-ring resonator; the PTAT temperature sensor measures the temperature of the micro-ring resonator in real time, the neural network processor uses a gradient descent method to optimize a loss function value representing thermal tuning error precision by learning a prior relation sample between the temperature of the micro-ring resonator and the voltage of the heater, and therefore the voltage of the heater at the real-time temperature is obtained. The heater applies this heating voltage to regulate the microring temperature, maintaining the microring resonator thermally stable.

Description

technical field [0001] The invention belongs to the communication technology, in particular to a thermal stabilization system and tuning method of a microring resonator based on a neural network. Background technique [0002] With the development of optical fiber communication technology, optical communication networks need to continuously improve work performance and reduce operating costs. The core technology lies in the miniaturization, integration and scale of optical waveguide devices. New optical waveguide devices with multiple functions, such as basic units that can simultaneously realize optical filters, delay lines, buffers, and various all-optical signal processing, realize powerful photonics on a substrate by integrating this unit on a large scale "System on a Chip". The microring meets the above two requirements, and its micronano-scale size is very suitable for large-scale monolithic tight integration, and can realize filters, delay lines, buffers, lasers, rout...

AI Technical Summary

Problems solved by technology

[0005] The invention measures the power through the on-chip photodetector at the download end of the microring resonator, which provides a local measurement of the average power. However, integrating the photodetector at the download end will greatly increase the device footprint, and the invention is not suitable for When the optical power changes, the feedback control algorithm used is the PID algorithm. Its main disadvantage is that the set point current depends on the on-chip power. Once the optical power of the microring input signal changes, it needs to be recalibrated, and the locking speed is s level

Images