Trustworthy Intersection Management by Crowd-sourced Digital Environment

Abstract

Nowadays, the automotive industry are developing smart functionalities for vehicles, and these developments are the prior steps in the process of making autonomous vehicles. Certainly, the big future technology improvement in the automotive industry would be the autonomous cars. The time when these vehicles are on the road, they will be communicating and cooperating for faster and safer travel. To adapt to this improvement, the traffic systems should become capable of cooperating with these smart vehicles. This project considers one of the essential elements of traffic systems, which is road intersection management. A smarter intersection management system should reduce the waiting and passing time of vehicles. Autonomous Intersection Management (AIM) is an advanced management system developed by researchers at the University of Texas for road intersections. It includes a scheduler and simulation manager to detect collisions of vehicles. However, AIM has several drawbacks like centralization(notably single point of failure) or security. In this thesis, we have worked to achieve a decentralized secure solution. In this project, using the Design Science Research Methodology, we have designed a blockchain based Autonomous Intersection Management. Since the blockchain technology is a potential platform to implement a secure and distributed system, we have made our solution based on this technology. With the help of a popular blockchain framework called Ethereum, we implemented a proof of concept and tested with an intersection simulation tool AIM 4.0. To validate the functioning of our solution, we have conducted various experimentation based evaluation on this simulation. As a result, we have demonstrated the feasibility of blockchain based AIM thereby implementing a distributed design to address the problem. Also it is shown that the replication from the blockchain technology could improve the availability of the AIM in case of system failures. As the autonomous cars are involving human lives, we have tried to create a safer solution. Through this, our intersection management can be robust against the vehicle’s faking their positions. Assuming the autonomous cars have advanced sensors for their environmental awareness, we have tried to incorporate this sensors’ information from the crowd vehicles to validate the locations of the other vehicles.