TPSense: A Framework for Event-Reports Trustworthiness Evaluation in Privacy-Preserving Vehicular Crowdsensing Systems

Abstract

Vehicles with abundant sensors and sophisticated communication capabilities have contributed to the emergency of vehicular crowdsensing systems. Vehicular crowdsensing is becoming a popular paradigm to collect a variety of traffic event-reports in intelligent transportation research. However, event-reports trustworthiness and drivers’ privacy are under the threats of the openness of sensing paradigms. This paper proposes TPSense, a lightweight fog-assisted vehicular crowdsensing framework, which guarantees data trustworthiness and users’ privacy. Firstly, we convert the data trustworthiness evaluation problem into a maximum likelihood estimation one, and solve it through expectation maximization algorithm. Secondly, blind signature technology is employed to generate a pseudonym to replace the vehicle’s real identity for the sake of drivers’ privacy protection. Our framework is assessed through simulations on both synthetic and real-world mobility traces. Results have shown that TPSense outshines existing schemes in event-reports trustworthiness evaluation and the reliability of vehicles.