传感数据世系的有效压缩传输与查询方法

Efficient compression propagating and querying method for sensor data lineage

为了有效地分析传感器网络应用中产生数据异常的原因并形成追溯链,该文提出一种基于感知压缩和列存储理论的传感数据世系压缩传输、存储与查询(Compressed propagating storing and querying of sensor data lineage,CPSQSDL)方法。论文分析被感知事件的传感数据世系之间蕴含有时间、空间相关关系,设计了一种适合传感数据世系压缩感知的随机投影观测矩阵,保证CPSQSDL方法具有近似k-term的最优恢复误差。论文对压缩世系进行了形式化定义,给出了近似世系查询算法,形式化证明了压缩世系的恢复误差边界,并在真实数据集上通过实验验证了此方法的有效性。

In order to analyze the reason that generates the abnormal data in sensor network applications effectively and to construct the tracing chain,we propose a transmission and storage methodbased on compressed sensing and column stored theory for the sensor data lineage,called CPSQSDL（Compressed propagating,storing and querying of sensor data lineage）,in this paper.We analyze the temporal and spatial correlation among the sensor data lineages of events,and find a suitable randomized projection observation matrix to ensure that k-term optimal reconstruction error.We describe the formal definition of compressed sensor lineage and design an algorithm for querying approximate lineage and formal proof of its error boundary.Experiments on the real data set prove the effectiveness of the proposed method.