Race without drivers: A look into A2RL, where cars are taught how to think and act like racers

Picture this – a race where cars zoom around the track, but there’s something different – there are no drivers behind the wheel. Instead, it’s computers racing against each other. Think of it like a high-speed video game but in real life. This is autonomous car racing.

Welcome to the future of racing, where the track is a battleground for algorithms and coding skills. Engineers and programmers are the real stars here, crafting the lines of code that drive these cars. It’s a whole new kind of competition, where winning depends on how well humans can teach machines to race.

Abu Dhabi’s Yas Marina Circuit is set to witness the world’s biggest autonomous racing on the 27th of April 2024. Named Abu Dhabi Autonomous Racing League (A2RL), this race was conceptualised and created by ASPIRE. This motorsport mixes mechanical autonomy with artificial intelligence (AI) to create a racing series in which autonomous vehicles will compete against one another.

With fully autonomous vehicles taking centre stage, A2RL promises not just exhilarating competition but also serves as a crucible for advancing autonomous technology and reshaping the future of transportation. A2RL features eight teams comprised of university entrants and research institutions, each equipped with a fully autonomous Dallara Super Formula SF23 car.

Just as the concept, the cars, and the race are different. Let’s deep dive into this with Stephane Timpano, CEO of ASPIRE.

How autonomous is the car

The number one question that arises in anyone’s mind while hearing about fully autonomous racing is, “How autonomous is the race/car”. To this, Timpano said, “Sometimes people ask if the cars are remotely controlled or similar. The truth is that once they’re competing on the track, there is absolutely no human input other than the stop instruction on safety grounds either by the team or race control. They are completely autonomous.”

The teams will undoubtedly keep an eye on the vehicles and systems, but they will not be involved in any other way. Long before any racing happens, scientists, engineers, and software developers labour to comprehend the automobile and its systems and create algorithms for racing under challenging circumstances.

The autonomous racing cars are equipped with a sophisticated array of sensors, cameras, and LIDARs (Light Detection and Ranging) to perceive their environment. The seven cameras and three LIDARs provide a 360-degree view, enabling the cars to navigate the racetrack and interact with objects in real-time. The data from these sensors is fused together, allowing the car’s AI system to make informed decisions during races.

While A2RL cars operate autonomously during races, human involvement remains integral throughout the development process. A team of developers work to understand the intricacies of the vehicles and refine algorithms for optimal performance. The human element manifests in the preparation and fine-tuning of the cars, with teams striving to outperform one another through innovative approaches and technological prowess.

Why the Super Formula SF23

The choice of the Super Formula SF23 car as the base vehicle for modification to race autonomously was strategic. SF23 is currently the fastest open-wheel race car in the world after Formula One. A ‘spec’ racing series was deemed appropriate to allow teams to focus primarily on autonomous system development rather than vehicle development.

The Super Formula SF23, originally designed for human drivers, offers exceptional performance and incorporates sustainable technologies. The modifications applied to the SF23 cars for A2RL include the integration of sensors, actuators, and control modules. While the cars possess a basic level of autonomy, the teams have the freedom to adapt and refine their software, allowing them to approach racing and risk-taking in their unique ways.

Technical challenges

This race comes with additional challenges as compared to regular F1 races. One of the core pillars of A2RL lies in conquering the myriad of technical challenges inherent in developing autonomous racing cars. With vehicles hurtling around the track, teams face the daunting task of coding intuitive human actions into machine algorithms. Managing tyre wear, executing strategic overtakes, and navigating complex racing scenarios all require meticulous programming and precise execution.

“The challenges that teams face cannot be understated. We’ve seen the resources, effort, and time that have gone into developing current automotive systems, now imagine the vehicle is travelling at 270km/h and competing against other autonomous cars. It’s a new frontier, a combination of top-level Formula racing and leading-edge autonomous systems,” said Timpano.

Teaching a machine to think like a human (especially a racer) is no easy task to achieve. Instead of human reflexes, it is computer programmes making split-second decisions. This is achieved with next-generation AI and ML technologies that learn as they go. As there are no drivers in this car, there is also no actual cockpit. The ‘cockpit’ consists of a stack of sophisticated sensors, computers and technology that will allow it to drive itself.

Race track: Testing ground for automotive technologies

The importance of tackling all these challenges and having this event is immense. Autonomous racing is not only a sport. It is a platform for research and development in the field of autonomous vehicle technology.

Motorsport has long served as a testing ground for innovative automotive technologies, with advancements trickling down to road vehicles (like active aerodynamics, carbon fibre, active suspension etc).

A2RL aims to utilise motorsport to advance road safety technology and foster greater consumer trust in AI and autonomous tech. By stress-testing transformative technologies on the racetrack, the series identifies key challenges and areas for improvement, accelerating the development of autonomous systems. As the series progresses, the technology will become more advanced, offering insights and capabilities that can revolutionise vehicles and infrastructure, ultimately leading to safer and more efficient road travel.

“At the end of the day, we are talking about a future where commuting and road travel is a stress-free experience and available to everyone – all while making it far safer. In terms of safety, autonomous systems are already outperforming human drivers in trials, so the potential benefit of fully matured technology and its widespread use is vast,” said Timpano.

Source: www.edgemiddleeast.com