Multi-sensor Air Quality Detection And Analysis System Based On Zynq

With the rapid development of social production methods,more work and recreational and sports activities can be carried out indoors,which leads to people spending much more time indoors than outdoors.Compared with the ordinary people,the elderly,the weak,the sick,the disabled,the young,the pregnant and other frail people spend more time indoors.Therefore,the indoor environment has a greater impact on them.It is particularly important to accurately and timely detect the concentration of pollutant gases in the indoor environment,and to comprehensively classify the indoor air quality level by integrating various indicators.To address the above issue,this paper designs a multi-sensor air quality detection and analysis system based on Zynq.The system is mainly divided into two parts,the detection terminal and the Android App.The detection terminal takes use of numerous electrochemical gas sensors to collect some parameters of indoor air,such as formaldehyde,ammonia,TVOC,temperature and humidity,and alcohol gas.The intelligent algorithm built in the terminal can effectively suppress the interference of alcohol gas on formaldehyde detection,while also synthesizing the five indicators of formaldehyde,ammonia,TVOC and temperature and humidity to judge the current indoor air quality level.When the data processing is completed,the detection terminal sends the data to the App via bluetooth.The Android App realize the display and storage of the data.The system designed in this paper uses the KNN-based formaldehyde anti-interference algorithm and the BP neural network-based air quality classification algorithm to process the data.In the formaldehyde anti-interference algorithm,the formaldehyde sensor data,alcohol sensor data and PPM-400 st formaldehyde detector data are combined to form a complete algorithm model.This algorithm can effectively improve the detection accuracy of the common formaldehyde sensor while suppressing the interference of alcohol gas.For the large number of repetitive multiplication and addition operations in the BP neural network,a special acceleration module is designed in this paper.Combined with the characteristics of Zynq SOC,the designed acceleration module is implemented in the PL part in the form of IP cores.Under the condition of satisfying the computational accuracy and time complexity,an additional computational control module is designed in this paper to reduce the power consumption of the system.Finally,the simulation and verification of each RTL module are carried out by building a verification platform,and the overall function of the system is tested by building a simulated experimental environment.The results show that in the system designed in this paper,the functions and calculation accuracy of each module can meet the design requirements.Therefore,the formaldehyde anti-interference algorithm can effectively suppress the interference factors in formaldehyde detection,and the air quality classification algorithm can accurately classify.