Incentivizing Private Data Sharing in Vehicular Networks: A Game-Theoretic Approach

TL;DR

This paper proposes a game-theoretic, decentralized architecture for vehicular networks that incentivizes data sharing while enhancing privacy through secure computation, aiming to optimize data utility and participant participation.

Contribution

It introduces a novel decentralized network architecture combined with a game-theoretic incentive model for privacy-preserving vehicular data sharing.

Findings

The model balances privacy concerns with monetary incentives for vehicle owners.

Optimal payment and sharing frequency are derived from real vehicular trace data.

Decentralization reduces risks of data reconstruction by adversaries.

Abstract

In the context of evolving smart cities and autonomous transportation systems, Vehicular Ad-hoc Networks (VANETs) and the Internet of Vehicles (IoV) are growing in significance. Vehicles are becoming more than just a means of transportation; they are collecting, processing, and transmitting massive amounts of data to make driving safer and more convenient. However, this advancement ushers in complex issues concerning the centralized structure of traditional vehicular networks and the privacy and security concerns around vehicular data. This paper offers a novel, game-theoretic network architecture to address these challenges. Our approach decentralizes data collection through distributed servers across the network, aggregating vehicular data into spatio-temporal maps via secure multi-party computation (SMPC). This strategy effectively reduces the chances of adversaries reconstructing a vehicle's complete path, increasing privacy. We also introduce an economic model grounded in game theory that incentivizes vehicle owners to participate in the network, balancing the owners' privacy concerns with the monetary benefits of data sharing. This model aims to maximize the data consumer's utility from the gathered sensor data by determining the most suitable payment to participating vehicles, the frequency in which these vehicles share their data, and the total number of servers in the network. We explore the interdependencies among these parameters and present our findings accordingly. To define meaningful utility and loss functions for our study, we utilize a real dataset of vehicular movement traces.