Implementation of Distributed AI in an Autonomous Driving Application

Abstract

Vehicle driving is an art to be performed with maximum attention. A small distraction or error in the driving practice may lead to severe problem to the people and the vehicle. The autonomous driving systems are implemented partially in few applications to rectify such human errors through an Artificial Intelligence (AI) algorithm. The AI algorithms require certain peripheral units like camera and sensors for their operation and are very effective and fast compared to the manual process. The computational complexity of autonomous driving systems are very high than the other applications where it requires continuous monitoring and instantaneous processing. Therefore it requires a huge amount of memory space and heavy processors. To address such limitations, the recent year applications are implemented with a cloud communication system for processing the collected data in a remote place. However, security and communication concerns present in such models have led this proposed work to implement a distributed AI architecture for an autonomous driving system.