High precision wavelength estimation method for integrated optics

TL;DR

This paper introduces a simple, integrated optical wavelength measurement method using a tunable micro ring resonator and neural networks, achieving high precision in a compact system.

Contribution

It presents a novel neural network-based calibration and estimation approach for integrated optics wavelength measurement, significantly improving precision and system simplicity.

Findings

Achieves 8 pm wavelength precision in 1549-1553 nm range

Improves precision by a factor of five over simple lookup methods

Enables simplified, integrated optical wavelength monitoring

Abstract

A novel and simple approach to optical wavelength measurement is presented in this paper. The working principle is demonstrated using a tunable waveguide micro ring resonator and single photodiode. The initial calibration is done with a set of known wavelengths and resonator tunings. The combined spectral sensitivity function of the resonator and photodiode at each tuning voltage was modeled by a neural network. For determining the unknown wavelengths, the resonator was tuned with a set of heating voltages and the corresponding photodiode signals are collected. The unknown wavelength was estimated, based on the collected photodiode signals, the calibrated neural networks, and an optimization algorithm. The wavelength estimate method provides a high spectral precision of about 8 pm (5*10^(-6) at 1550 nm) in the wavelength range between 1549 nm to 1553 nm. A higher precision of 5 pm (3*10^(-6)) is achieved in the range between 1550.3 nm to 1550.8 nm, which is a factor of five improved compared to a simple lookup of data. The importance of our approach is that it strongly simplifies the optical system and enables optical integration. The approach is also of general importance, because it may be applicable to all wavelength monitoring devices which show an adjustable wavelength response.