Temperature Compensation for Semiconductor Gas Sensors Based on Whale Optimization Algorithm–Least-Squares Support Vector Machine

A whale optimization algorithm–optimized least-squares support vector machine (WOA-LSSVM) temperature compensation model is proposed to compensate for the temperature drift of the output signal of semiconductor gas sensors in practical applications. The whale optimization algorithm is used to optimi...

Full description

Saved in:
Bibliographic Details
Published in:Sensors and materials Vol. 35; no. 8; p. 2933
Main Authors: Xu, Yunde, Cheng, Zhenzhen, He, Guofeng, Liang, Chengwu
Format: Journal Article
Language:English
Published: Tokyo MYU Scientific Publishing Division 24.08.2023
Subjects:
Algorithms
Artificial neural networks
Back propagation networks
Drift
Error reduction
Gas sensors
Kernel functions
Least squares
Mathematical models
Neural networks
Optimization algorithms
Parameters
Particle swarm optimization
Regularization
Sensors
Support vector machines
Temperature compensation
ISSN:0914-4935, 2435-0869
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A whale optimization algorithm–optimized least-squares support vector machine (WOA-LSSVM) temperature compensation model is proposed to compensate for the temperature drift of the output signal of semiconductor gas sensors in practical applications. The whale optimization algorithm is used to optimize the selection of the regularization parameter γ and the kernel function parameter σ2 in the LSSVM model, and the temperature is corrected by predicting the output of the sensor through the parameter-optimized LSSVM model. Experimental results show that after the temperature compensation of a semiconductor gas sensor by the WOA-LSSVM model, the temperature coefficient of the sensor sensitivity is reduced from 4.21 × 10−3 before compensation to 2.001 × 10−5 after compensation, and the relative error is reduced from 17.68 to 0.08%. The prediction results of the WOA-LSSVM model are compared with those of the least-squares support vector machine, particle swarm optimization–least-squares support vector machine (PSO-LSSVM), and whale optimization algorithm–back-propagation neural network (WOA-BPNN) models. The WOA-LSSVM model had the highest compensation accuracy and can effectively improve the robustness of semiconductor gas sensors to temperature drift.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0914-4935
2435-0869
DOI:10.18494/SAM4417