Path Finding of Auto Driving Car Using Deep Learning

Chih Yung Tseng, Purdue University

Abstract

Self-driving cars can provide a sustainable, safe, convenient and congestion free transportation system. By using artificial intelligence, especially in machine learning, it can approach the achievement that we want. However, being able to infallibly recognize objects is just one of several challenges which artificial intelligence must solve. One of the solutions is to apply deep learning and computer vision. Convolutional Neural Networks’ (CNN) deep architecture classification approaches have gained popularity recently for its ability to learn middle and high level image representations. In previous experiments, there were many similar examples using different CNN models to train the robot for graphic recognition and obstacle avoidance. However, there is still much room for improvement in the department of image recognition, especially pertaining to its accuracy In this project, CNN has been applied as a training tool to process image classification and object avoidance on remote robotic cars built with the Nvidia Jetson Nano developer kit. The kit was programmed using the wireless programming environment, Jupyter notebook. In addition, two different CNN models have been applied to analyze the output result performance. The main purpose is to train the robot to identify objects and improve its accuracy. The recognition and accuracy rate under different conditions can be observed by comparing the two models with different graphic inputs conditions. This project adopts the pre-train model for real time demonstrations and can be executed in a cloudless environment (without networks involved). As a result, the robot can achieve a high accuracy rate in both CNN models output result performance. Moreover, the pre train model can execute in local service to accomplish cloudless.

Degree

M.Sc.

Advisors

Wang, Purdue University.

Subject Area

Artificial intelligence|Computer science

COinS