| dc.description.abstract | Rapid urbanization and industrialization has resulted in a sustained degradationof environmental quality parameters. To keep track on contamination level, it getsto be essential to develop a realistic model for monitoring.Traditional methods for air pollution measurement are expensive and have a spatialconstraint. With these limitations, air pollution monitoring in broader area isnot feasible. Wireless sensor network (WSN) creates an opportunity to monitorthe pollution level by creating low cost WSN. It provides information about differentpollution related issues and pollution data of various spatial regions. Themain aim of this project is to develop a WSN with low cost multi-sensor node,to measure distinctive air toxin and pollution parameters are transferred to sinkwith the help of radio modules. To maintain data accuracy, calibration of each sensor is required which is performed by comparing sensor's sensing data with reference data. With the help of topology control protocol and diffusion mechanism,it becomes possible to route data from different location to centralized hubor vice-versa. Data gathering and data aggregation protocol is applied for datacollection and effective data monitoring. Data analysis plays an important rolein recognition of influenced parameters and sensor's calibration. However, influencedparameters are also useful in derivation of particulate matter (PM) concentrationwith support vector machine (SVM) or artificial neural network (ANN)model. We performed an experiment on sensor and calibration by applying curvefitting technique to reduce the error between sensing data and reference data. Forcarbon monoxide and nitrogen dioxide, behavior of sensor is varies. Experimentswere performed on sensors and estimated error values of different parametersare : carbon monoxide - 39.142% ,nitrogen dioxide - 14.5849% ,ozone - 6.5881%,temperature - 2.5356%, humidity - 12.5125% and PM - 7.48%. While transferringdata, process takes longer duration due to limited features of radio module. Ourfuture endeavor is to provide a facility to access the pollution related informationof different places by developing graphical user interface (GUI). Being an ad-hocnetwork, maintenance of network connectivity and topology is a major challenge.Network must be fault-tolerant and self-healing to avoid any failure in connectivityand to reduce errors in pollution related data. It is also useful for detectionof faulty nodes (which continuously sense false data) and re-route the network ifrequired. | |
