Abstract
Autonomous Diving is one of the hot topics that has many application domains in the industry, such as car manufacturing, robotics, etc. In essence, it is a combination of technologies, including sensors, cameras and AI, to enable a vehicle to travel to a destination without the intervention of a human operator. This project demonstrates how it can be implemented by using Autonomous path finding algorithms and direct Cozmo - an Anki corporation developed, Science, Technology, Engineering and Mathematics (STEM) education robot to travel a path. Cozmo comes as a package with Cozmo Software Development Kit (SDK), Code Lab, and sufficient documentation. We will use Convolutional Neural Networks (CNN) to classify the images collected by Cozmo, and direct it to proceed towards the right direction. A major goal of this project is to demystify Autonomous Driving technology for K-12 students, by decomposing and visualizing the Autonomous Driving process into multiple, easy-to-understand, customizable components, and developing easy-to-follow projects for k-12 students. For this goal to be achieved, developing a web portal in the Flask framework illustrating the theoretical idea and also a demonstration is the way to go. We hope through the web portal the student will have a better understanding of Artificial Intelligence (AI), model and also can get a view of what is really happening in Cozmo’s head while driving. We hope that through this project, we can provide a general methodology on developing interactive projects to introduce AI technologies (such as Autonomous Driving) to K-12 CS courses.