Integrating Controller Area Network (CAN) with Cloud-Based Data Storage Solutions for Improved Vehicle Diagnostics using AI

Abstract

Machine learning and AI algorithms are gaining popular use and acceptance in virtually all sectors. The leading drivers of this evolution are computing and data storage advancements, hence an increasingly interconnected world. While this is applauded, massive information tracking and diagnostics are less embedded in vehicle engineering.

Cars are increasingly well-connected and with the advent of car sharing and autonomous vehicles, all of them will need to be autonomous and fully safe. Cars sold in this last decade are equipped with multiple embedded electronic control devices, nowadays increasingly interconnected via Controller Area Networks (CAN). At the same time, all new vehicles are connected, not only cellularly but also via storage systems that are attached to clouds. These systems have evolved to be platform agnostic, can accept a high arrival rate of data, and become low-latency storage systems in clusters spread over several continents. They deliver geo-localized near real-time storage that can be used to monitor one car within a several-kilometer radius.

Such systems can store in real-time (limited minutes latency) data that collectively spans 40 million kilometers covering the 20,000 cars being sold every day. These cloud-attached storage systems are not only capable of storing vast amounts of data in real-time but they are also used as data platforms whereby AI can constantly monitor vast sets of cars