Linear Structure Index for Network-Constrained Moving Objects.

A dimension-reduced and linearly ordered (DRLOP) index structure for network-constrained moving objects is proposed to address the challenges posed by the rapid increase in the volume, diversity, and intensity of spatio-temporal data. The critical metadata rectangles are projected into reduced-dimensional phase points, and then arranged into a linear order structure to speed up retrieval through filtering and binary search. Using a dataset generated by a moving object generator from the underlying road network in Texas, our experimental results demonstrated the superior efficiency of the proposed index compared with the MON-Tree in terms of query and index creation. DRLOP exhibits the advantages of simple index structure, small storage space consumption, and high efficiency of spatio-temporal query operation, which is especially suitable for retrievals of large-scale historical trajectory data.