Paper Title
Speed Discipline for Autonomous Vehicles in Indian Driving Conditions

Abstract
Popularity of autonomous smart vehicles is increasing due to their ability to improve road safety and transportation efficiency. One of the key challenges in autonomous driving is predicting the speed of surrounding vehicles to ensure safe and efficient maneuvering. In this work, a speed prediction approach is implemented during real time vehicle scenarios on Indian roads. For this purpose, the (You only look once)YOLO v7 object detection algorithm with Intersection over union(IOU) is used. Our approach involves training a deep neural network on a dataset of annotated video frames consisting of labeled bounding boxes around the vehicles in front and corresponding speed measurements. The trained model is then used to predict the real time speed of the vehicle using a camera feed as input. Performance evaluation of the model is done on a real-time driving experience based data set and high accuracy in speed prediction is achieved. Keywords - YOLOv7, Error Analysis, Linear Regression, IOU