.png){kind=icon}
.png){kind=icon}
Return
Multiple-Sensor System Detection via Feature Combination Selection and Low-Redundancy Optimization
.png){kind=icon}
PRE
AI
DOI：10.1109/TIM.2024.3476565.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In this article, a multisensor system for rapidly detecting bacterial infection is developed, including a negative pressure odor information acquisition module with multiple types of sensors and the corresponding sensor conditioning circuits. To capture both transient and steady-state information from the sensor response curve, multiple features are extracted, including maximum response value, maximum first-order derivative, corresponding to the maximum response value, and so on. In addition, an iterative feature combination selection process is employed to choose features from the initial feature set that contribute significantly to classification accuracy. Due to the redundant information and noise introduced by the sensors' broad-spectrum response characteristics and hardware circuit interference, a low feature redundancy feature selection algorithm combined with the grouping characteristics of sensors is proposed. The bacterial culture experiment is conducted based on the designed system. Compared with the existing algorithms, the designed algorithm demonstrates superior classification performance with fewer features on the bacterial culture dataset.
Keywords:
Sensors
Microorganisms
Feature extraction
Sensor phenomena and characterization
Sensor arrays
Sensor systems
Gas detectors
Circuits
Optimization
Valves
Bacterial infection
feature combination selection
low-redundancy optimization
multisensor system
Journal
IF:
5.9
Papers:
1.9W
Citations:
5.8W