Vision driven lane changing system of self driving car
Abstract
Accidents due to driver error persists despite road rules and regulations being updated regularly, mostly because of reckless driving and non-measured actions while lane changing. Therefore, the inclusion of a lane changing system in self-driving cars can greatly reduce these fatal tragedies resulting in safer streets along with an easy mode of transportation for people of all ages and conditions. The system we designed uses image processing which consists of creation of region of interest, perspective transformation, threshold operations, canny edge detections, histogram and various calculations for lane detection and lane maintenance. It further recognizes object detection models to detect vehicles, stop signs and traffic signals using the Haar cascade Deep learning method to make decisions and maneuver our vehicle accordingly. The output of the analysis is deployed on a protype level build with the help of Raspberry Pi 3B+ having integrated connection with Arduino and motor driver system.