Monitoring systems in landslide areas are important elements of effective Early Warning structures. Data acquisition and data retrieval allow the detection of movement processes and thus are essential to generate warnings in time. Apart from precise measurement, the reliability of data is fundamental, because outliers can trigger false alarms and lead to the loss of acceptance of such systems. For the monitoring of mass movements, it is important to know if there is movement, how fast it is and how trustworthy is the information. ln the context of the joint project “A Sensorbased Landslide Early Warning System (SLEWS)”, this thesis deals with the development of a prototypic monitoring system for different types of landslides and the investigation of its suitability, but also its limitations. The developed system is based on a modern Wireless Sensor Network (WSN) for data transmission.
lt is characterized by a self-organizing (ad-hoc) structure, with bi-directional communication in real-time and multi-hop data transfer. For the detection of surface deformations in landslide areas, small low-cost micro sensors, also called Micro-Electro-Mechanical-Systems (MEMS), and Position Sensors from the automobile industries, different industrial applications and from other measurement technologies were chosen. ln laboratory tests the accuracy and resolution of the sensors integrated in the WSN environment were investigated. Furthermore, field tests in landslide areas were also performed to prove system stability under real conditions. For the improvement of data quality but also the reduction of errors and outliers, the concept of Multi-Sensor Data Fusion was applied. For this purpose, a model for Multi-Sensor Data Fusion from the Joint Directors of Laboratories (JDL) was adapted on landslide monitoring.