Steering angle estimation for autonomous vehicle navigation using hough and Euclidean transform

The recent trend of proliferation of road accidents and fatalities, calls for a methodology that can appease the burden on the drivers during certain crucial moments, requiring fast response and decision making. This paper expatiates a novel and effective procedure for determining the steering angle required to keep the vehicle in the middle of the lane. The procedure used is divided in multiple stages. The first stage includes determination of the road´s midpoint using Robert edge detection and morphological operations. This midpoint is then used to determine the angle by which a vehicle must steer to maintain its position at the center of the road. It is done using Euclidean distance transform and trigonometric equations. The input data suffers non-linearity, irregular intensity, non-uniform contrast difference which makes edge extraction ineffective. To tackle such situations, Hough transform and virtual road edge was used to determine the road and lane edges. The algorithm was tested rigorously using video samples, acquired from camera mounted on hood of a car, characterizing various road conditions. The results underscored the robustness and effectiveness of the same.