Truman Yang
Embedded Systems
Collecting ping pong balls after playing requires a great deal of time and effort. In this project, an AI-based ping pong ball collector will be developed to minimize labor and human intervention. NVIDIA Jetson Nano Developer Kit will be utilized to run multiple neural networks in parallel for needed applications such as object recognition and obstacle avoidance.
(1) Design and implement an autonomous robot that collects ping pong balls autonomously. (2) Literature review will be conducted. (3) Implement two to three existing neural network models/machine learning techniques to Jetson Nano AI platform and uncover the best approach for the project. Further, make improvement on the best approach you determined.
(1) Evaluate object recognition and obstacle avoidance models in terms of their accuracy, efficiency, and complexity.
(2) Build a good size robot so that it can store around 20 balls.
(1) Software development in Python.
(2) Make improvements on existing design and implantation.
(3) Idea generation technique with Scamper
(1) Programming on the Jetson Nano development Kits
(2) Design, implement, and document the software as specified above.
Build the ping pong ball collector.
Design and implement the 2-3 object recognition models.
Design and implement 2-3 obstacle avoidance models.
Evaluate the performance of different models and solutions.
COE318: Software Systems
TY06: Autonomous Ping Pong ball collector | Truman Yang | Wednesday August 31st 2022 at 09:26 AM