Private Drivers Identification based on users' routine

This paper presents Private Secure Routine (PSR) as a paradigm with two main objectives: i) identify drivers depending on their habits/routine and ii) keep private drivers' data. We implemented PSR exploiting the secure Multi-Party Computation (MPC) technique against a honest-but-curious attacker model. Moreover, we evaluated PSR by establishing its accuracy in combination with other existing research works based on machine learning techniques. Evaluation of PSR is performed on different test-beds, considering single-owner and two-owners identification.