Secure blockchain-based traffic load balancing using edge computing and reinforcement learning

Abstract

Congestion represents one of the major problems in constantly growing cities, and the expansion and modernization of the traffic system are a major priority to all government infrastructures. In effect, more efficient traffic represents a more efficient economy. Solving the issue of congestion by increasing the number of roads is not always the most cost-effective solution as it represents massive changes in city infrastructures that have been present for decades. Urban planning shapes the environment around us but fails to address specifically the future traffic clogging. Our research tackles the problem of traffic congestion by proposing a system for vehicle detection, identification, and count, allied with reinforcement learning for traffic congestion anticipation and prediction. The need for a real-time and efficient system led us to push the research and development onto an Edge Computing platform using IoT and secure transactions using Hyperledger Fabric blockchain. Blockchain intervenes as a security protocol in the proposed system. The native form of Internet of Things does not include security protocols which are important to a large scale implementation. Ensuring the security of transactions and permanent access control is the main reason why blockchain is rooted in our system architecture. Our project aims to reduce the traffic load on roads experiencing significant congestion, and improve overall city traffic system without costly investment into new communication infrastructures and city planning.