Vision Based Intelligent Vehicle Steering Control Using Single Camera for Automated Highway System

摘要

Generally, stereoscope vision is used for guiding autonomous intelligent road vehicles, as image depth can be calculated easily. However, when one of the camera fails, it will advantages to have suitable guidance algorithm that can detect the lane marking using single camera. The primary objective of this work is to develop and implement control algorithms for identifying and guiding the intelligent road vehicle in the assigned lane using image-processing techniques, using single camera. It deals with dividing the video being taken by the camera, into several number of frames. Then, the obtained frames are processed using Image acquisition techniques in Matlab. It detects the lane to be followed through image processing, calculates the angle of movement required for the robot to stay in the assigned path and commands the steering system with appropriate control algorithms. The identification of lane is done using color detection and boundary marking techniques. The image initially undergoes Inverse Perspective Mapping (IPM) for viewing the 3-D space in a 2-dimensional array. Once the IPM of the acquired image undergoes lane detection algorithm, the angle of any curve in the lane is detected using angle detection module. The robot is also made to move at the center of the path using Fuzzy Logic algorithms, where it takes distance from the left detected edge as the input and outputs the angle to be moved by the robot to stay in the center. The developed algorithm has been tested using P3-DX Pioneer mobile in laboratory condition. The experimental result obtained show that the algorithm used by the vehicle to follow the center of the lane, with any given angle or position of initiation, has worked all the times under given experimental conditions, with a maximum error of 2 cm.