The Consumer Electronic Show (CES) is a mecca for high-tech products, traditionally focused on consumer electronics, computers and gaming. More recently, robotics, cars and autonomy have an increased presence with autonomous cars offering ride-hailing trips on the famed Las Vegas strip and Tesla cars transporting attendees in the Loop (engineered and operated by the Elon Musk owned Boring Company) at the Las Vegas Convention Center.
The Las Vegas Motor Speedway hosts racecar events like NASCAR and Indy 500. The main track is an asphalt surface, 1.5 miles long with 20-degree banked turns and 9-degree banked front and backstretch. It seats 123,000 people.
Fusing the CES theme of autonomy with racecars is exciting. IAC (Indy Autonomous Challenge) will run an autonomous racecar competition at the Las Vegas Motor Speedway on January 9, 2025 during the CES 2025 exhibition. This is the fourth time IAC has held a race at CES, having made history with the world’s first head-to-head autonomous racing at CES 2022. At CES 2025, IAC will include a first-of-its-kind multicar autonomous race with three or more racecars battling to cross the finish line first.
IAC recently held a thrilling and action packed event at the Indianapolis Motor Speedway with 9 teams competing against each other in individual speed and 2-car passing competitions. The speed event produced a world record speed of 184 miles/hour (Cavalier Racing, University of Virginia), the fastest ever on a closed track, while the passing competition clocked in at a 170 miles/hour speed (PoliMOVE-MSU, a collaboration between Politecnico di Milano, Italy, Michigan State University and University of Alabama). For reference, the fastest speeds recorded in human driver racing events like Formula 1 (F1), Indy 500 or NASCAR is ~240 miles/hour.
The CES 2025 event ups the ante. First, in addition to the 9 teams that competed in Indianapolis (2 from Italy, 1 from Germany, 1 from South Korea and 6 from the United States), a tenth team (Caltech) has joined the roster. Apart from the individual speed trials, there will be elimination rounds with 2 cars competing in the passing competition. The winners will then compete in 3 and 4 car passing competitions. It is likely that some more records will be set like breaking the Indianapolis IAC speed record of 184 mph and approaching the 240 mph benchmark set by human driven racecars on a track.
Racecars and DARPA
The IAC was recently awarded a DARPA contract for test and evaluation of the AI and software for autonomous racecars. As part of the TIAMAT program (Transfer from Imprecise and Abstract Models to Autonomous Technologies), the goal as stated by DARPA program manager, Dr, Alvaro Velasquez is “ to learn and and transfer autonomy across diverse, low-fidelity simulations and leveraging their shared semantics (e.g., rules of engagement). This leads to a more rapid transfer of autonomy from simulation to reality – perhaps even as early as the same-day versus weeks/months with traditional approaches”.
Paul Mitchell is the CEO of the IAC. His goal starting with the CES 2025 event is to build an operational module spanning perception, localization, vehicle dynamics & control and path planning that leverages and fuses the best elements from the various competing teams. “After three years and 20,000 miles of high-speed autonomous racing, we have observed that different university teams’ AI drivers can be more skilled in certain areas than others, such as sensor fusion or path planning. IAC is launching the OpenDriver platform, supported by DARPA, to rapidly identify, test, and then combine the best skills of different AI drivers from university research teams and other performers worldwide. The result will be Physical AI capabilities that can be used to solve real-world commercial and national security challenges.”
What Does it Take To Win??
Based on discussions with Paul Mitchell and the leaders of the winning teams – Professor Madhur Bell (Cavalier Racing – University of Virginia, winner of the IAC individual speed event), and Professor Sergio Savaresi (PoliMOVE – Politecnico di Milano, Italy, winner of the IAC 2-car passing competition), it takes perfection in 5 dimensions to succeed and win in closed track autonomous car racing:
- Localization: knowing the car position relative to the track at all times is critical for making decisions on negotiating turns and maximizing speed on straight sections. GPS sensors, cameras and LiDAR provide this capability.
- Perception: situational awareness of the vehicle surroundings including obstacles and other racecars is especially important in multi-car races. The ability to discriminate relevant sensor data (LiDAR, radar, camera) and process it rapidly allows for low control latency.
- Physical Modeling of Vehicle Dynamics: interestingly, this is emphasized as one of the key differentiators between the various teams. Physical models of vehicle dynamics under different conditions of track conditions and configurations, weather, wind, ambient and tire temperature are used along with sensor data (pressure and temperature sensors) to adjust algorithms for vehicle control and push the limit in terms of “safe” speed. Refining these models in terms of fidelity and accuracy is part of the Sim-to-Real revolution in physical AI implementations.
- Path Planning & Vehicle Control: items 1-3 provide inputs for path planning and control of the vehicle and decisions on speed, direction, steering and braking.
- Strategy and Tactics: Trial laps conducted prior to the final event allow teams to tune their algorithms for various events. Passing tactics especially in multi-car races can also be prepared depending on the specific opponent and their behavior and capabilities. Generally, teams prepare different software versions (conservative to aggressive) to account for different temperature, wind speed, weather and competitors.
The Winners At The Indianapolis Event Share Their Thoughts
Professor Sergio Savaresi, leader of the PoliMOVE team (Politecnico di Milano, Italy) feels that the 3 and 4 car passing competition at CES 2025 is exciting and challenging, and will require significant scenario planning. “Moving the head-to-head competition from 2 to 3 (or 4) cars will significantly increase the complexity of the planning module of the autonomous car. The passing competition between PoliMOVE and Unimore at the IAC 2024 Indianapolis event allowed us to collect a very large amount (almost 30 minutes) of perception data of the opponent car, at very high speeds. Analyzing this data will help us plan for CES 2025”.
According to Professor Madhur Bell, leader of Cavalier Racing (University of Virginia), localizing the vehicle is critical, as is path planning (staying as close as possible to the race-line) and using accurate vehicle dynamics models, especially to negotiate left hand turns on the oval portions of the track. Perception is not as critical in the speed competition but very important for the multi-opponent passing phase of the event. “Adding just one or two extra autonomous cars makes the challenge exponentially harder. We must not only focus on the car directly ahead but assess complex racing dynamics in real-time – like when to hang back or seize an opportunity if two cars are locked in their own overtake. The fast, precise overtakes we achieved at the IAC event in September 2024 gives us the the foundation to make progress on our simulations and autonomy stack. The intelligence, collision avoidance and decision making portions of the stack are critical for the winning team”.
The Consumer Electronics Show is an exciting event – certainly for new and yet to arrive hardware products and increasingly to display AI integration to enable Autonomy of Things ( AoT™) and real world autonomy applications. The IAC event at CES 2025 is poised to increase the excitement factor on January 9. 2025.
Read the full article here
