Cutting-edge vehicles, speed, and competition are nothing new at the Indianapolis Motor Speedway. But on October 23, 2021, the home of the Indy 500 will experience a first: Fast cars without drivers. With some help from the official edge computing sponsor, ADLINK, universities around the world are designing race cars for the Indy Autonomous Challenge (IAC) capable of competing on the course – and winning – in the first-ever, high-speed autonomous land race using full-sized vehicles.
Competitors are working to achieve the goal of crossing the finish line in the 20-lap race in 25 minutes or less, which requires an average speed of 120 mph – and there is substantial motivation to finish in the top three. Solving this complex problem and crossing the finish line first will earn the winning team $1 million. Second- and third-place finishers will receive $250K and $50K, respectively.
Meet the University of Hawaii AI Racing Tech Team
One of the teams competing in the ground-breaking event is the University of Hawaii (UH) AI Racing Tech Team under the mentorship of engineering leader Gary Passon. Leading the Autonomous Vehicle Tech program on the Maui campus, Passon retired to Hawaii from a successful career in high-tech but brought with him his passion for auto racing. His unique blend of technology, edge computing, and racing expertise is guiding his fifteen-person team of undergrads, grads, research assts and industry mentor as they optimize their vehicle perception, path and control algorithms.
However, Passon’s team, composed of students from the UH Maui Autonomous Vehicle Technology class, have some experience of their own. They developed an evKart for the 2020 Purdue evGrandPrix. This international autonomous electric go-kart race challenged students to build their own autonomous software and integrate with state-of-the-art hardware and system, and race them at Indianapolis.
The UH AI Racing Tech Team designed an evKart that used vision cameras, Lidars, GPS, and inertial measuring unit (IMU) devices to control the vehicle, with all computing processing power and navigation onboard. The team used NVIDIA Jetson modules, providing speed and power efficiency to the embedded vision processing and control. The team also chose the ADLINK ROScube, which runs on ROS 2, is designed to help developers build robotics applications at the edge.
The UH AI Racing Tech Team’s evKart design qualified them for the evGrandPrix competition, which, unfortunately, had to be canceled due to the COVID-19 pandemic. However, the team put all that they learned to prepare for the Purdue evGrandPrix to good use.
Speeding Ahead with Edge Computing
The UH AI Racing Tech Team began working in the Spring of 2020 to qualify for the Indy Autonomous Challenge. But instead of an evKart, this time, the team worked on a design of an autonomous software stack for a full-sized Dallara AV-21 racecar. And they have an influx of fresh talent and new perspective with University of California San Diego Contextual Robotics Institute autonomous racing team led by Jack Silderman Ph.D. also joining the UH AI Racing Tech Team.
This vehicle is powered by the ADLINK AVA-3501series robot controller, specifically designed for vehicle use, but shares a common platform and similar configuration to ROScube, making the transition from the team’s evGrandPrix project to the IAC easier and more familiar.
ADLINK developed a Docker image built around its robotic control system. This IAC racecar software support enables IAC university teams to race using Open Robotics ROS 2 with Autoware.Auto autonomous driving packages, Eclipse Zeno V2X and Eclipse Cyclone DDS with Eclipse iceoryx zero-copy built-in. UH designs are focused on Open Source products.
Dallara AV-21 Racecar 
ADLINK AVA-3501 Series Robot Controller
Out of the box, the Dallara AV-21 race platform with the AVA-3501 series robot controller can navigate the car around the Indianapolis Motor Speedway. However, it’s up to the UH AI Racing Tech Team and teams from other universities to design a racing algorithm and system of sensor packages that can negotiate their vehicles on the racetrack – and innovate so that their entry will achieve the best time.
Furthermore, the designs the teams are creating for the competition may also solve some persistent challenges related to autonomous vehicle commercialization. Their projects may provide insights and solutions to issues such as avoiding sudden obstacles at high speeds while maintaining control. The IAC is also helping to advance autonomous vehicles and edge computing in general by fostering the next generation of engineers, technologists, and machine learning experts who are developing skills that can change the world for the better.
The Indy Autonomous Challenge Teams to Watch
To qualify for the October 2021 race, teams tested their software virtually in May 2021 in a simulation powered by Ansys. Among the entrants, 15 teams qualified for the race in Indianapolis, taking the competition – excitement – to the next level.
The University of Hawaii AI Racing Tech Team will be there! Follow the team on Twitter @AIRacingTech.
For information and updates on the race, visit the Indy Autonomous Challenge website
I never hesitate to remark on your article because it has had such a profound impact on me. You’re doing a fantastic job. Keep up the fantastic work!