# High-Precision and Low-Complexity Silicon Waveguide-Integrated Temperature Sensor System

## TL;DR

This paper presents a new, low-cost silicon-based temperature sensor with high precision using a waveguide and machine learning.

## Contribution

A novel silicon waveguide-integrated temperature sensor with high resolution and low complexity using CNN for demodulation.

## Key findings

- The sensor achieves a best resolution of 0.00287 °C (2.87 mK) in the temperature range of −20 °C to 100 °C.
- The use of a compressed taper structure improves speckle data collection efficiency and reduces system complexity and cost.
- Experimental results align well with theoretical values, confirming the sensor's reliability and resolution.

## Abstract

This paper proposes a high-sensitivity-integrated temperature sensor with low complexity based on a silicon waveguide. The waveguide layout is optimized through the finite-difference time-domain (FDTD) simulations, and a compressed taper structure improves the efficiency of speckle data collection while reducing the system complexity and cost. To achieve precise temperature demodulation, this paper employed a convolutional neural network (CNN) for nonlinear fitting. Experimental results demonstrate the sensor’s ability to perform temperature measurement in the range of −20 °C to 100 °C, with a best resolution of 0.00287 °C (2.87 mK). The resolution and reliability of the measurements are validated by comparison with the theoretical values. This study introduces a novel approach to silicon waveguide-based temperature sensing.