Abstract: Multi-sensor data fusion is a kind of data processing method. It refers to fusing data from a large number of sensors for various and multi-level treatment, creating a sort of information more meaningfully, while this sort of information is almost impossible to get for any single sensor. Researching of WSN-based data fusion in water quality, it puts forward a multi-sensor data fusion model of two levels, suitable for water quality monitoring, which combines self-adaptive weighted theory and fuzzy comprehensive evaluation theory together. Processing the same type of data collected from monitored area, it uses the self-adaptive weighted theory, while it uses the fuzzy comprehensive evaluation theory for different types of data. The experimental result, made by lake water sample to Chaohu Basin, shows that using the model can effectively reduce the amount of data traffic, reduce the rate of deviation, improve the credibility of monitoring result.

Key words: Wireless Sensor Network（WSN）, water quality monitoring, data fusion, self-adaptive weighted theory, fuzzy comprehensive evaluation theory

摘要： 多传感器数据融合是一种数据处理方法，可以对来自多个传感器的数据进行多方面、多层次的处理，从而产生更有意义的信息，而这种信息是单一传感器难以获得的。通过在水质监测应用背景下研究基于WSN的数据融合，提出了一种针对水质监测的两级数据融合模型：处理传感器所采集到的监测数据时，对于同类型的数据采用自适应加权理论进行第一级数据融合；对于不同类型的数据采用模糊综合评价理论进行第二级数据融合。对巢湖流域水样监测的实验结果表明，这种采用自适应加权理论和模糊综合评价理论相结合的数据融合模型，能够有效降低监测网络的数据传输量，降低监测数据的误差，提高水质状态监测的可信度。

关键词: 无线传感器网络（WSN）, 水质监测, 数据融合, 自适应加权理论, 模糊综合评价理论