Crowdsourced Traffic Event Detection and Source Reputation Assessment Using Smart Contracts

Area of Science:

• Computer Science
• Transportation Engineering
• Data Science

Background:

• Real-time traffic event data is vital for transportation safety and efficiency.
• Crowdsensing offers broader coverage than traditional infrastructure but requires data trustworthiness assessment.
• Detecting data manipulation, deception, and spam is critical for reliable crowdsourced traffic information.

Purpose of the Study:

• To design and demonstrate a system for road traffic event detection and source reputation assessment.
• To address the challenge of unreliable data sources in crowdsensing applications.
• To develop a decentralized system resilient to institutional data manipulation.

Main Methods:

• A system integrating road traffic event detection and source reputation assessment was developed.
• The system was designed for decentralized operation using smart contracts on a blockchain platform.
• Evaluation involved synthetic datasets and real user data from a mobile application within a driving simulator.

Main Results:

• The proposed system accurately detects various manipulative and misreporting behaviors in crowdsourced traffic data.
• The system demonstrates rapid convergence to a final trust score.
• Effectiveness was validated even within the resource constraints of a blockchain virtual machine.

Conclusions:

• The developed system enhances the trustworthiness of crowdsourced traffic data for improved transportation management.
• Decentralized implementation on blockchain increases resilience and eliminates single points of failure.
• The approach is effective in identifying and mitigating unreliable data sources in real-world traffic scenarios.