Reliable E-Nose System using the Improved Optimization Technique based ANN

Since from last decade, there is a growing interest in a system that detects the pollutant gases and other environmental information is called Electronic Nose (E-Nose) networks. The gases such as methanol, Liquid Petroleum Gases, ammonia, etc. are harmful for human beings; therefore, such frailness required detecting automatically as well as safety alarm promoted in a specific field. The critical challenges of the E-nose system are efficient to detect with minimum error and overhead. In this paper, we targeted to design the optimized machine learning-based algorithm to detect and alert the pollutant gases, Humidity, O2 Level, and Air Temperature in the real-time datasets. We initiated E-nose design using Artificial Neural Network (ANN). Using essential ANN leads to poor accuracy and error rates, as they failed to select the best solutions during the training process. Thus, we next use the Particle Swarm Optimization (PSO) based ANN called ANN-PSO to improve the accuracy rate and error performances for E-Nose systems. Finally, the proposed Improved Optimization Technique based ANN (IOT-ANN) machine learning model designed and evaluated in current this research work. The IoT-ANN it is based on a bio-inspired algorithm to achieve reliable training during the E-Nose prediction.