Robotic Zippin’ Airhockey

Our project 

We were tasked by the UBC Engineering Physics Project Lab to develop a platform for a next-generation AI and Automation course. Over the last two years, we designed and built an automated air hockey table that allows students to pit intelligent bot players they designed against human opponents. 

Our inspiration 

We were excited by the multidisciplinary nature of this project, which involved mechanical design, physics simulations and applied robotics. 

Our biggest challenge 

The biggest challenge was finding a way to control the air hockey mallet to move across the table. We had developed a model of the robot physics but were unable to account for all of the real-world discrepancies and spent several months tuning our model. 

What excited us most 

We were excited to work as a team and solve complex robotics problems together. We were never satisfied with “good enough” and worked tirelessly to make our robotic air hockey table better, faster and smarter. We spent more time on this project than on all of our other courses combined because of how much fun we had working together.  

The most interesting/surprising thing we learned 

We took a deep dive into control theory over the last several months and learned how to develop a feedforward control model that integrates the mechanics of our robot with error-correcting aspects that allow us to apply the model to the real world. 

Our project’s future 

The project lab intends to keep the project running. The aim is to develop the robot to the point where Engineering Physics could feasibly run an AI-focused course around it. Students would develop autonomous software agents for the table which could compete against human opponents as well as each other. There is definitely still some work that needs to be done before it is at that level of reliability, and we are excited to see what the next team or teams come up with.