A Multi Path Abnormal Data Filtering Algorithm for Space-Air-Ground Integrated Networks

Space-Air-Ground Integrated Networks (SAGINs) are important infrastructures of 6G. With the support of SAGINs, data collection in unattended areas becomes possible. However, data measured by various devices in unattended areas is often unreliable. Original collected data are usually accompanied with noise, error, or inconsistent value. To ensure the accuracy and efficiency of urban computing, real-time and light-weight outlier filtering becomes a prerequisite of urban applications. In practical SAGIN applications, software defined wireless networking provides the capable of managing the whole wireless network globally, while edge computing enables the network edge of SAGIN hold the computation capacity. Based on the combination of these two technologies, a real-time multi path abnormal data filtering algorithm is proposed. In the proposed algorithm, two abnormal data filters are utilized, a hierarchical distributed scheme is performed in the edge server to acquire low overhead and real-time requirements. The performance of the proposed outlier filtering algorithm is validated in the simulations.