Semantic-aware anomaly detection in real time parking data

In this work, we introduce and experimentally evaluate a novel approach for real time anomaly detection in smart car parking applications. We attach semantics on top of raw real time parking data collected from sensors of parking lots. We use knowledge from historical data to detect anomalies on real time data. Attaching semantics on top of raw data helps reduce the learning time by a factor of 3.1x and also provides the error checker a distinct context to look into potential problems.