Differential Deviation Based Abnormal Driving Behavior Detection

According to a report on road safety, more than half of fatal crashes are originated from aggressive drivers. The rear-end collision is the most frequent type of collision in the USA and 87% of them occur due to distracted driving behavior. These facts show that drivers are at increasing risk from nearby other abnormal drivers. The detection of such abnormal driving behavior is important otherwise it may jeopardize the safety of other innocent drivers. However, the driving behavior is a combination of both internal (e.g., skills) and external (e.g., road type, traffic conditions) factors which make the abnormal driving behavior detection largely subjective. On the other hand, this subjectivity could be handled through the analysis of driving behavior deviation among drivers. In this paper, we tackle this problem and propose a differential deviation based abnormal driving behavior detection method that is not only mining the individual driving behavior but also compares it with others to infer the differential deviation among drivers. The inferred differential deviation is then used to detect abnormal driving behavior. We demonstrate the benefits of the proposed method through extensive simulations conducted on simulated traffic data. Our preliminary result has shown that the proposed method can identify all driving behavior anomalies with about 85 % accuracy under different settings.