ADFIST: Adaptive Dynamic Fuzzy Inference System Tree Driven by Optimized Knowledge Base for Indoor Air Quality...

Area of Science:

• Environmental Science
• Public Health
• Sensor Technology

Background:

• Indoor air pollution is a major cause of increased mortality and morbidity.
• Specific pollutants such as PM10, PM2.5, CO2, CO, tVOC, and NO2 pose significant health risks.
• These risks include respiratory diseases, central nervous system dysfunction, cardiovascular disease, and cancer.

Purpose of the Study:

• To assess indoor air quality by measuring key pollutant concentrations.
• To develop and implement an advanced model for processing and predicting indoor air quality data.
• To evaluate the model's performance against existing studies and benchmark datasets.

Main Methods:

• Utilized an Internet of Things (IoT)-based sensor system for real-time pollutant measurements in a rural Indian site.
• Implemented an Adaptive Dynamic Fuzzy Inference System Tree for data processing.
• Employed global and local search algorithms for knowledge base design and model tuning, respectively.

Main Results:

• Achieved normalized root mean square errors (NRMSE) for pollutants: PM10 (0.6679), PM2.5 (0.6218), CO2 (0.1077), CO (0.2585), tVOC (0.0667), and NO2 (0.0635).
• The proposed model demonstrated enhanced prediction performance.
• Validation against existing literature and benchmark datasets confirmed the approach's efficacy.

Conclusions:

• The developed IoT-based system and Adaptive Dynamic Fuzzy Inference System Tree provide an effective method for indoor air quality assessment.
• Accurate monitoring of pollutants like PM2.5 and CO2 is crucial for mitigating public health risks.
• The model's performance indicates its potential for real-world application in improving indoor environments.