Development of Reduction Method of Management Cost in Mud-slide Detection Sensor Network

点击翻译按钮获取中文摘要

Against mud-slide disasters of slope caused by rainfall, the sensor network system that measures the pore-water pressure and rainfall at each position of the slope has been studied. However, the system has a problem that the management cost is high. The cost is brought by (a) maintaining sensor nodes to exchanging battery and to repair, and (b) analyzing sensor data to manually determine the numeric parameters to predict mud-slides. To reduce these costs, we develop the following methods: (1) optimization method of the sampling rate by trading-off between sensor data quality and battery lifetime, (2) categorization method of sensor nodes based on similarity of sensor data to reduce maintenance times and target sensor nodes for analysis, and (3) mud-slide prediction method with automatic calculation of prediction parameters. We also conducted the experiment to verify the effectiveness of the developed methods with actual data.