A Basic Study of Data Collection Ratio of Drifting Sensor Networks for Monitoring Waterways

In this paper, we model the data collection ratio of drifting sensor networks that are designed for observing waterways such as sewer pipes and rivers. A drifting sensor network is a system to collect data observed in a waterway using sensor nodes that are drifting on a waterway and access points placed along the waterway. It needs a high cost to collect data in long waterways such as rivers and sewer pipes by human hands. In a drifting sensor network system, small sensor nodes thrown into a waterway measure data such as temperature and water level, and take pictures while drifting and transmit the collected data to fixed access points (APs). Such systems will be helpful to save labor cost and time to inspect waterways. The amount of data collected by a drifting sensor network is affected by the nodes' movement that depends on the water flow, the number of nodes, the probability of node crashes, etc. Therefore, it is important to know the suitable values of parameters such as the number of nodes and battery size. We modeled the behavior of a drifting sensor network including node mobility, communication between nodes and APs, energy consumption, and crash of nodes. We derived the expected value of the data collection ratio (the ratio of the amount of collected data against the distance of the observation area) analytically. The results give us a guideline to choose important parameters, the number of nodes and battery size, and are useful to design more sophisticated systems including communication between sensor nodes for sharing backup data to improve the reliability.