Nonnegative Matrix Factorization based privacy preservation in vehicular communication

Vehicular Ad hoc Networks (VANETs) provide enjoyable driving experience through traffic information collected by moving vehicles on the road. When a vehicle discovers any events such as car accident, traffic congestion, hazardous road condition, etc., it shares such information with other vehicles through Road Side Units (RSUs). Inevitably, its location information also needs to be included in the message for specific and accurate information. However, drivers may not be comfortable sending their locations to others because there is a chance that the privacy of the driver can be compromised. Thus, they might choose not to report discovered events, which can cause highly degrading performance of the network. In this paper, we propose a Nonnegative Matrix Factorization (NMF) based privacy preservation scheme to perturb the source location data without cryptography while it can still calculate the location of the event occurred. The proposed scheme utilizes the intrinsic property of NMF to distort the data for protecting driver´s location privacy. It then clusters the drivers in accordance with their locations, the relative distances and directions, as well as the timestamps. By doing so, events can be identified based on the clusters while driver´s private information is preserved.