SmartAIR: Smart energy efficient framework for large network of air quality monitoring systems

Abstract

Live street-level air quality monitoring is important application of sensor networks. Such application reveals human exposure to hazardous air pollutants. It assists general public, army troops, environment agencies and the Government in decision-making every day. Live data visualization and data fusion plays crucial role in presenting pollution updates effectively for end-users. We propose efficient interactive, live data visualization in our application. Our application efficiently renders pollution data fast in under 10ms. Users will be instantly aware of pollution levels in their desired location. Continuous data-logging at data centers from large-scale of sensor networks poses major challenges. We use policy based network management technique to reduce unwanted data-logging requests. We implement novel policies in identifying and rejecting numerous unwanted requests at data centers. Each data-logging involves computationally expensive database operations and with our policy specification we were able to cut down expensive operations significantly $(\geq 83\%$ reduction, especially in denser regions like traffic congested roads). Finally, we implement Lazy load scheme to make our application more energy efficient. With this scheme we save data and battery in end-users device over longer periods of time. We conducted several real-life trials and we observed negligible mobile data consumption $(\leq 1MB$ for 1 - hour). Similarly, we observed negligible power consumption $(\leq 4 \%$ in 1- hour run) in end-users device. Our implementation of novel policies and schemes provide real-life benefits to data centers and end-users. Our end-users experience better, faster and lively pollution updates. Our data centers experience relatively lesser network load and less computation overheads on scaling up.