GLOSS: Tensor-Based Anomaly Detection in Spatiotemporal Urban Traffic Data

Abstract: Anomaly detection in spatiotemporal data is a challenging problem encountered in a variety of applications including hyperspectral imaging, video surveillance and urban traffic monitoring. In the case of urban traffic data, anomalies refer to unusual events such as traffic congestion and unexpected crowd gatherings. Detecting these anomalies is challenging due to the dependence of anomaly definition on time and space. In this paper, we introduce an unsupervised tensor-based anomaly detection method for spatiotemporal urban traffic data. The proposed method assumes that the anomalies are sparse and temporally continuous, {i.e.}, anomalies appear as spatially contiguous groups of locations that show anomalous values consistently for a short duration of time. Furthermore, a manifold embedding approach is adopted to preserve the local geometric structure of the data across each mode. The proposed framework, Graph Regularized Low-rank plus Temporally Smooth Sparse decomposition (GLOSS), is formulated as an optimization problem and solved using alternating method of multipliers (ADMM). The resulting algorithm is shown to converge and be robust against missing data and noise. The proposed framework is evaluated on both synthetic and real spatiotemporal urban traffic data and compared with baseline methods.