Chasing dreams in the cloud

Last weekend, the University of Pennsylvania hosted its semi-annual PennApps: the world's original college hackathon.

 

Hackathons are 2436 hour long events where students code, build, and/or wire a project of their choosing from scratch. Sleep, hygiene, and nutrition are put on the back-burner during the arduous process of creating projects worthy of the praise of tech giants such as Google and Amazon. Every engineering student dreams of winning big at a hackathon—especially one as prestigious as PennApps.

However, the Winter 2018 PennApps grand prize winners had much deeper motivation. Liam Dugan, Zhengyi Luo, and Austin Small built their project, CloudChaser, on the concepts of determination, family, and destiny. After the hackathon, Zhengyi and Austin were able to strengthen their own unique blends of academia and industry. Liam successfully demonstrated the world of engineering to his younger brother, providing the support he himself lacked at a younger age.

 

Their 3D-printed demo robot, Chase, is no Terminator—it's only about the size of a Kleenex box. But despite

Chase's unassumingsize, it is able to get some serious brain power from the cloud, thanks to some clever engineering from the team. While most robotics efforts today include an on-board processor that handles computation, CloudChaser passes off its number crunching to the cloud. This enables the development of smaller, yet more intelligent, robots.

 

Cycling between different laboratories at Penn to complete their project, the team successfully 3D-printed, wired, and coded CloudChaser within 36 hours. Although the team knew they were up against harsh competition, they were satisfied in knowing that they had emerged from the hackathon as more complete engineers. However, last minute bugs scared the CloudChaser team. Even Liam's little brother admitted the situation looked bleak. But their win confirmed the significance of what the team was hoping to develop: a solution to a huge bottleneck mobile robotics.

 

 

Zhengyi, whose research in computer vision directly coincides with CloudChaser, is particularly excited to finally discover a calling towards his future research topic.

"I was trying to figure out my life. One day, Liam says to check his project out. And I said to myself that 'This is it.'"

The CloudChaser team offered some advice to future hackathon-goers:

“Every single step of the way there WILL be something that goes wrong.”

These are their big takeaways from PennApps:

1. Always test/familiarize yourself with your libraries before hacking.
2. Have applicable potential in your project.
3. Just have a minimum viable product—but make it cool.

The CloudChaser team took home a big win, and it has no plans on stopping now. Since CloudChaser presents a solution to one of the drawbacks facing mobile robotics today, the team plans on publishing a paper with their findings. Furthermore, the CloudChasers are excited to envision the potential of allowing computationally intensive image processing and machine learning algorithms to be used in low resource robotics.


The team hopes that CloudChaser is the first for a future of smaller, smarter robots.