Executive Summary : | A scalable wireless sensor network for continuous real-time water quality monitoring is envisaged. An ever-increasing population and associated farming/industrial activities critically rely on the continuous availability of freshwater. This has put enormous environmental stress on the scant water resources (2.6% of total water) available in the world. Recharging these reservoirs are monsoon-dependent, and failures are often attributed to climate change and decades-old crop patterns in India. Harmful contaminations are released into our water sources on a regular basis. Monitoring water resources has become essential to prevent outbreaks of diseases or chronic ailments, develop settlements etc. Conventional approaches depend on fixed water monitoring stations at sparse locations and can be expensive for large-scale deployment. These typically need manual data collection. To address the need for high spatial and temporal sampling investigators propose sensors that are (1) Low Cost: Easy and cheaper to make, (2) Environmentally friendly: Easy to dispose of and (3) Real-time: Continuous real-time direct sample-to-answer measurement. The complementary detection platforms based on Colorimetric, Electrochemical, Electrical and Optical methods will be used for target analytes such as (1) Acidityor pH (2) Dissolved CO2 or pCO2 (3) Dissolved oxygen or pO2 (4) nitrates (5) ammonia and (6) heavy metals (e.g. Pb, As, Cd) and (7) E-Coli. Land-based and immersed/floating sensors will be employed. In both cases, data will be collected at the cloud using wireless modules. The power requirement of these sensors and associated networks will be met by integrating multi-modal energy harvesting methodologies, a combination of solar photovoltaics, small hydro and wind energy harvesting systems. For the data communication from the sensors, the network architecture will be based on Low Power WAN (LPWAN) which offers higher battery life. The main challenge in implementing the transceiver is in the reduction of power, without compromising on bit error rate, system latency and sensitivity. For this purpose, they will implement a wake-up receiver. Furthermore, compact and high-efficiency transparent antenna topologies will allow the co-existence of solar panels and antennas. In large-scale deployments, rationalization of the location of the monitoring station and water quality parameters can capture more variability. While important savings on the amount transmitted data can be achieved using distributed compression, the collected data are further subjected to time series analysis, learning and prediction of relevant parameters. For the latter aspects, machine learning techniques appear very promising and accurate. Notice that since the sensor operating environment is covered by water bodies, accurate physical system modeling, new error control strategies and access protocols can enable scalability to large networks. |