Enhancement of Array Optimization Algorithm via Information Theory for a Novel Multisensor Detection System
Junhui Qian, Yuanyuan Lu, Mengchen Lu, Ziyu Liu, Peng Xu
Abstract
This paper develops a low-cost dual-input multi-sensor odor detection system relying on sensor detection technology and the changing characteristics of the odor itself. The finite element analysis method is used to investigate the gas chamber, and the uniformity and stability of the flow field distribution under the wide range of flow velocity are verified through simulation results and actual verification experiments. To cope with the interference brought by redundant information, improve the system’s detection accuracy, and reduce the number of sensors, we propose a sensor array optimization algorithm based on information theory. Specifically, according to the dynamic changes of the information contained in the sensor array during the optimization process, we develop a modified mutual information optimization (MMIO) framework. Experiment results show that the proposed MMIO can obtain significant system performance compared to the existing methods.