Grid-Based Index of Spatiotemporal Data for AIS Trajectories

With the rapid growth of global maritime business, effectively managing and utilizing Automatic Identification System (AIS) data has become a key technical challenge. Current data management methods, such as Quad-tree and R-tree, as well as spatiotemporal indexing techniques like GeoMesa, struggle to handle the high retrieval complexity and computational demands of large-scale, dynamic AIS trajectory data. To address this, we propose a grid-based indexing approach for AIS trajectory spatiotemporal data. Firstly, we propose a multi-scale adaptive hierarchical grid-based spatiotemporal identification code generation method called GeoSOT-ATD (GeoSOT-AIS Trajectory Data). This method profiles AIS trajectory data to generate a one-dimensional spatiotemporal identification code. GeoSOT-ATD employs an adaptive hierarchical iterative strategy to ensure efficient indexing performance while controlling the number of grid cells. Next, based on this one-dimensional spatiotemporal identification code, we introduce two indexing methods: direct indexing and split indexing. The direct indexing method ensures the uniqueness of the primary key for each row and uses two column families to store spatiotemporal and non-spatiotemporal attributes, respectively. The split indexing method divides the spatiotemporal identification code into a three-part structure and refines the filtering range of column family conditions to improve retrieval efficiency. We compare the spatiotemporal range query performance of GeoMesa and GeoSOT-ATD under spatiotemporal iso-selectivity conditions. The results show that the spatiotemporal indexing method proposed in this paper improves retrieval performance by approximately six times compared to GeoMesa under uneven spatiotemporal distribution.