Study on Incident Detection and Management of Vehicular OBD Data on Blockchain

Abstract

A majority of the research on accident detection systems involves increasing the precision at which it can be detected. This approach proposes a test-bed for a vehicular incident detection system that uses real time data collected from OBD -II port of automobiles and later passed on to Edge devices within vehicles to detect an accident or other rash driving behavior patterns. Machine learning algorithms are developed and deployed onto edge devices like Jetsonnano for real time pattern recognition. Once an incident like accident or rash driving is triggered, the data is cached and written onto Inter Planetary File Systems (IPFS) a distributed file storage system and put onto Ethereum Blockchain for later analysis and documentation by various stakeholders like police, RTO, law, forensic team etc. Blockchain acts as an immutable ledger providing proof of all incidents. Further using the obtained data, dashboards of the incident can be generated for further visualization and understandability to the concerned stakeholder