Forward Lane Integrity Watchdog System

Abstract

Today there exist a lot of smart vehicles which can change lane on their own, using their sensors to detect the vehicles around them and using various neural or non-neural algorithms to detect the lane on the road. But these are inherently limited to well-structured road environment and struggle with unstructured road or damaged road. This paper aims to propose a new system, based on cloud and deep-learning neutral networks to process images from each region to train a neural network to be highly efficient in that particular region. We use “Collective wisdom” of people along with data analysis to improve the accuracy of the model.