The Embedded Machine Learning Club won second place in the 2023 F1Tenth Autonomous Vehicle Race down in San Antonio in May! This competition challenges students to build autonomous vehicles, at a tenth of the scale of a regular vehicle. The vehicles are expected to sense their surroundings and drive automatically within a race track. Think of it like a tiny Tesla!
Cassidy Petrykowski, CEO of the Embedded Machine Learning Club, learned about the F1Tenth competition about a year ago and began to fundraise the necessary money that the club would need to buy the race cars. During that year of fundraising, Cassidy spent time researching with NC State faculty and gaining the necessary technical knowledge she needed to lead the club in this endeavor.
In the Spring 2023 semester, club members applied to be on the F1Tenth team within the Embedded Machine Learning Club. Shortly after, they began building their vehicles for the San Antonio F1Tenth Autonomous Vehicle Race that took place on May 9th. The cars are built to “memorize” a race track and follow commands built into its system by the EML team to navigate the course correctly.
Additionally, the car is programmed to sense walls, lines, or boundaries of a never-before-seen track and navigate its way through, autonomously while club members cheer on their F1Tenth cars from the sidelines.
The team of 14 students flew to San Antonio, bringing their autonomous vehicles, named Derek and Darius. The first day consisted of setting up the small-scale race track where the cars could practice, and eventually compete. Each car had to complete a qualifying round before the day of the race. Derek quickly and confidently passed the qualifying round. Darrius also passed, but the team realized they needed to make some adjustments to his vehicle before the big race.
The next day the team members returned to the track with Derek and Darius for some pre-race testing and practice rounds. They met with the other competing teams, consisting of students from the University of Pennsylvania, Carnegie Mellon, and more.
The San Antonio Race
Darrius was up first, facing a car that also traveled from NC State! The Embedded Machine Learning team was still anxious to win and progress their vehicle, Darrius. This race required the cars to drive autonomously and complete as many laps as possible in a set amount of time. Despite some errors during the race, Darrius completed the most laps and progressed to the next round!
Darrius’s success was short-lived when it competed against UPENN in the next round. Unfortunately, Darrius lost and was dismissed from the rest of the competition. But Derek’s journey would live on. In Derek’s first round, he won by completing 10 laps in 5 minutes, more laps than his opponent.
The EML club nervously awaited its next round of competition with Derek going up against Carnegie Melon’s car. Our NC State EML team won their race against CMU and advanced to the final round! That’s when UPENN was up against Derek, who had previously defeated Darrius. Sadly, the team lost to UPENN again, but they secured their second-place win!
The club took us along with them via Instagram. They kept us updated on the competition and posted videos of the trials, victories, and errors. Students cheered on the club members from Raleigh in the Instagram direct messages.
Many Instagram users responded to the victories saying they were “so proud of the pack.” Others responded with hand-clapping emojis, hearts, and other praises of the team’s journey.
Cassidy told us that the club is a space for students to start at zero and grow from there. She wants people who have questions to feel comfortable asking them, and to eventually grow in their knowledge enough to one day be able to answer those questions for newcomers.
Hear more from Cassidy about getting involved in the F1Tenth competition and what her hopes for the Embedded Machine Learning Club are in the future!
IEEE Intelligent Vehicles Symposium
After a successful race in San Antonio, Daniel Yanke, an electrical and computer engineering student, traveled to Anchorage, Alaska on June 4-7 to represent the Embedded Machine Learning Club at the IEEE Intelligent Vehicles Symposium.
At the symposium, the F1Tenth IV2023 Championships were held where seven F1Tenth vehicles competed against each other. Some of the notable teams up against NC State at the race were Nagoya University from Japan and Clemson University!
The competing cars raced two sets of 10 laps each. The total time from all 20 laps was used to determine the winner. Each race consisted of two cars on the track at a time. The races were broken up into brackets to determine the teams who would compete against each other and advance to the next round.
The Embedded Machine Learning Club’s F1Tenth car, Derek, raced against each competitor’s car in Alaska. Derek made it to the final round and won the championship title against Clemson University with a per-lap average time of only 0.05 seconds faster over the 20 laps!
Daniel said “it was close, and every second counted right up to the end! It went so fast. I was there for hours every day testing and tuning various changes and upgrades to the algorithm, but it feels like I was only there for an hour or two. Coming in first was amazing.”