On Comparing the Similarity and Dissimilarity Between Two Distinct Vehicular Trajectories

In this paper, we study the problem of comparing similarity and dissimilarity between two distinct vehicular trajectories by proposing an adjacency-based metric. This approach has a broad application in building truthfulness vehicles evaluating paths hazardous materials transportation. Given sequences $A$ $B$ Point Interests (POIs) visited road/transportation network, is to delete some nodes from obtain $A'$ $B'$ such that number adjacencies maximized. duplicated are allowed, e.g., represent road edges. We prove NP-hard when noisy POIs deleted all remaining must be involved adjacency. On other hand, adjacency involvement constraint dropped, variation as hard Exact Set Cover hence cannot approximated within constant factor unless P=NP. At last, design practical based on local search show algorithm very effective simulation data.