Downloads: 2 | Views: 81 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1

Research Paper | Computer Science & Engineering | India | Volume 12 Issue 5, May 2023

Training and Testing Autonomous Driving Software using Udacity Car Simulator

Attla Vamshi Krishna Reddy | Guda Varun Sai | Sarvesh Vishwakarma | Annam Mani Sai | Dantuluri Abhishek

Abstract: Deep Learning has been shown to be a successful method for autonomous driving, allowing cars to identify and adapt to their surroundings in real time. Many advances in deep learning-based approaches for autonomous driving have been developed in recent years, including object identification, semantic segmentation, and path planning. Object detection allows the vehicle to distinguish items in its immediate surroundings, such as pedestrians, automobiles, and traffic signs. Semantic segmentation allows the vehicle to distinguish assorted items in the area as well as their borders, offering a more thorough awareness of the surroundings. Path planning is the process of determining a safe and efficient route for a vehicle based on data from object detection and semantic segmentation. Deep learning's capacity to learn from massive volumes of data is one of its primary advantages for autonomous driving, allowing the automobile to adapt to changing events and conditions. Yet, there are still obstacles to overcome, such as assuring the safety and dependability of self-driving vehicles and resolving ethical issues about decision-making.

Keywords: Autonomous driving, Deep learning, Convolutional neural networks (CNNs), Reinforcement learning, Decision-making, Perception, Accuracy

Edition: Volume 12 Issue 5, May 2023,

Pages: 2206 - 2210