Each year, the renowned Goodwood House plays host to an array of motoring excellence for the esteemed Goodwood Festival of Speed, where the racing industry’s finest gather to compete. This race, considered deceptive in its complexity, challenges participants to manoeuvre the fastest car-driver pair through an intricate 1.8km course that is only utilised for racing purposes three days a year. It houses a mix of vehicles and drivers of various age and experience – a rally of racing royalty and newcomers, vintage sports cars, new-age supercars, and everything in between.
Despite its everyday use as a driveway, this course is anything but simple during the three-day festival. The track, often slippery and narrow, features added difficulties like the formidable Flint Wall and the deceiving Molecomb corner, which have seen renowned racers fail among the hay bales.
Given these complexities, it’s perfectly clear why the Indy Autonomous Challenge, backed by tyre manufacturer Bridgestone, included their robotic racing car candidate at the 2024 Festival of Speed. The Goodwood hill climb, deceptive and demanding even to the globe’s most accomplished racers, proposes a unique challenge for an autonomous car that navigates via Lidar scanners and cameras.
The autonomous car was received with curiosity at the festival, loaded with conventionalists and hosting famous figures such as Mika Hakkinen, Max Verstappen, and Richard “The King” Petty from NASCAR.
The Indy Autonomous Challenge (IAC) aims to identify the most proficient and agile autonomous racing car using a Dallara-built frame and a Honda-made engine from an Indy Lights vehicle. This series serves as the precursor to the famous Indycar racing series, which boasts the renowned Indianapolis 500 as its highlight.
Each vehicle in the competition is run by a single university or a consortium of them and is fitted with the latest in autonomous driving sensor and software technology.
For those who hold traditional motor racing dear, fear not – the purpose of the International Autonomous Challenge (IAC) is not to replace human drivers with artificially intelligent ones. The ultimate aim is to utilise the highly competitive setting of motorsport as a platform for accelerating the development of autonomous driving technology on a broader scale.
Marc Ferlet, IAC’s Director of Marketing, outlined a niche in the autonomous and self-driving sectors, particularly in the area of high-speed autonomy testing and development. He noticed that major corporations rooted mostly on America’s west coast were mainly concentrating on suburban and urban applications of autonomous ‘robo-taxis’.
The IAC’s innovative thought was to utilise a robot for tasks humans struggle with, such as high-speed, Autobahn-like driving, instead of activities we are fairly proficient in. Furthermore, it’s generally less challenging to teach a robot how to manoeuvre on extensive multilane highways or racetracks than on cluttered, disordered city streets.
It is a deeply engrained and widely accurate notion that racing induces improvements in car development and models. However, Ferlet stresses that the IAC is not seeking to advance racing. Instead, their original plan involved a single race in October 2021. Following the success of that event, corporate entities and academic institutions expressed interest in using the platform for research purposes and objectives.
Ferlet seemed amused at the suggestion of pitting an AI against a human driver at Goodwood, to compete against some of the finest racers ever. Nonetheless, it is clear that the IAC’s objective is to advance technology for everyday driving, rather than to revolutionise the motorsport industry.
Ferlet emphasises that the idea was not to suggest we would replace racing drivers, a concept no one desires. Rather, it might be intriguing to observe a car operating autonomously, conducting safe speeds of high magnitude. The car is presented at diverse locations, such as Goodwood and the Indianapolis Motor Speedway, where spectators can see the software safely driving the vehicle at high speeds of up to 270km/h without incidents. This increases the acceptability of computers driving at 130km/h on major roads.
But there’s more to the Goodwood trip than just exhibitions. The track, as mentioned, is complex and the house is notorious for poor mobile phone and GPS signal, so the IAC vehicle’s software had to lean heavily on its cameras, in-built sensors, and particularly its Lidar.
Provided by Luminar, the IAC vehicle is equipped with four Lidar sensors, one per corner, which is double when compared to the amount of sensors in conventional cars. According to Ferlet, the vehicle’s software leaned more toward using the Lidar over the cameras while navigating the Goodwood driveway, registering a maximum speed of 66.96 seconds, hitting a top speed of 162.8km/h. To put this into perspective, the fastest time of the weekend clocked 43.98 seconds, achieved by racer Romain Dumas in the powerful 2,000hp electric Ford Supervan. This indicates that AI has much to improve before matching Dumas’s speed.
It is essential to note that, for the IAC team, tyres play a significant role in the drives at Goodwood. The IAC receives substantial backing from Bridgestone and the cars have been put through their paces at Bridgestone’s extensive testing facilities near Rome. Not uncommonly, the Goodwood event presented a rigorous challenge with shifting grip up the hill owing to considerable rainfall. Experienced human racers can instinctively gauge tyre grip, but this is a task a computer still struggles with. An added complication was that the IAC car was running on cold tyres, since the AI, as of now, cannot perform a burnout to heat up the tyres at the start line.
As Ferlet expressed, understanding grip is second nature for a race car driver, while it remains a complex issue for autonomous vehicles. A thorough understanding and modelling of tyre performance and track behaviour is required. Insights such as this could be vital for future robotically driven cars which is one of the key reasons Bridgestone support the IAC. As put by Sara Correa, Bridgestone’s chief marketing officer, motorsports are integral to Bridgestone’s essence and will continue to be. The lessons learned from such ventures will be applied towards developing sustainable mobility solutions for society and clients.
However, despite arguing that creating a robotic racing series is not the aim and that the IAC serves for autonomous vehicle technology experiments, Ferlet conceded that the software managing the vehicle can be calibrated to exhibit varying levels of “aggression” in its driving style. Parameters need to be set for the car, and decisions need to be made on the reaction to any occurrence – whether it should halt or proceed. The teams also must determine the car’s yielding point in multi-car events. The personalities of teams, and possibly even their nationalities, are reflected by these decisions. And, as Ferlet remarks, racing team egos are comparable to that of top racing drivers.
All at once, the prospect of a competition involving robotic cars doesn’t seem too terrible. You could instruct one vehicle to mimic the driving style of Alain Prost, another to replicate Ayrton Senna’s flair, and one more to emulate Nigel Mansell’s technique. We’d have ourselves a thrilling revival of the 1980s racing scene.