AWS Integrated Machine Learning Architecture for Advanced Driver Assist System (ADAS)

A cloud-based infrastructure for an advanced Embedded System that can help detect the state of a vehicle driver from a sensor suite's input, and the subsequent pre-processing of data, along with Machine Learning, both integrated as well as on-cloud, to determine the state of the driver is presented. The Embedded System is a useful extension to existing Advanced Driver Assist System(s) (ADAS), and with Machine Learning as well as a cloud infrastructure, provides a robust and fault-tolerant system. The Embedded System designed is also integrable with traditional vehicle communication system(s), namely the CAN (Controller Area Network) Bus, this makes it easily integrable to existing vehicle ecosystem(s) and also makes it scalable based on need. The integration with cloud also helps the System tackle existing challenges presented with integration of Embedded Systems onto the cloud, thus the system presented in robust, scalable, fault tolerant and realized both locally and on-cloud.