Study Of MSDF Application In Network Monitor System For Mountainous Urban Drainage Pipeline

The city drainage pipeline network needs comprehensive monitoring to keep its operation safety. The pipeline network should be monitored especially when it is located in the mountainous city because the terrain is complex, geological structure is special and natural geological disasters such as landslides, avalanches occur frequently. The natural geological disasters may break drainage pipeline network and cause sewage leaks. Consequently, it is essential to construct a monitor system to monitor these geological disasters. Due to the complex types of monitor sensors, massive data collected from the field should be fused using multi-sensor data fusion (MSDF). MSDF is the hot researching focus both domestic and international. At present, most of geological disasters monitoring systems were neither reducing transmission data nor conduct multi-sensor data fusion to improve monitoring accuracy.Owing to scale of monitoring drainage pipeline network in huge and medium sized mountain cities, a MSDF model was designed for mountainous city drainage monitor system based on studying MSDF technology and its classes and models. The model is consisted of two layers: one is conducting multi-sensor data fusion on similar data while the other is on heterogeneous data. When conducting multi-sensor data fusion on similar data, the data should be pre-processed by filtering, treating outlier and predicting states with the Kalman filter estimate to reduce unnecessary network traffic. And then, the data collected from the same type of sensors distributed in different locations were fused by modified centralized adaptive weighting Kalman filter. The subjects of the whole monitoring area were estimated comprehensively, transmission traffic was reduced in the wireless sensor network, and accuracy of the monitoring system was increased. For the heterogeneous sensor data fusion, a weight based D-S evidence theory algorithm was designed to do comprehensive analysis on monitoring subjects for the different types of sensors. The experiment results verified the effectiveness of the fuzzy comprehensive evaluation and weight-based algorithm for D-S evidence theory. Through the data fusion operation by implementation of the feature distance-based D-S evidence theory algorithm on conflicting data for monitoring subjects, the monitoring accuracy can be improved, the influence of conflict data on final monitoring decision may be prevented, and the reliability of the final decision of monitor system can also be improved. The thesis designed the MSDF model, same types and heterogeneous of sensors data fusion algorithm. The final experiment results verify that the MSDF reach the goals of reducing transmission traffic and improving the accuracy of monitor system for the mountainous city drainage pipeline network monitoring system.