This year’s Goodwood Festival of Speed will see the first-ever fully autonomous hill climb when Roborace’s Robocar takes on the 1.16-mile track in July.
Guided only by automated systems, Robocar will be tasked with navigating hay bales, flint walls and forests on the Goodwood estate, using a variety of sensors that give it 360-degree machine vision around the car.
“Roborace plays an important role in the future of mobility, challenging public perceptions and providing a platform to advance new technologies, said Charles Gordon-Lennox, the Duke of Richmond and founder of the Festival of Speed. “This makes them the perfect partner to undertake this significant feat.”
The vehicle weighs 1,350kg and is powered by four 135kW electric motors used to drive each wheel, for a combined 500-plus hp. An NVIDIA DRIVE PX 2 computer processes Robocar’s data, which includes inputs from the LiDAR, radar, GPS, ultrasonic and camera sensors.
Lucas di Grassi, CEO of Roborace, said: “We are excited that the Duke of Richmond has invited us to make history at Goodwood as we attempt the first ever fully – and truly – autonomous uphill climb using only artificial intelligence.”
Roborace provides the car with an API (application programming interface) as a platform for teams who then add their AI driver algorithm to the vehicle. Development of the automated driving system for the Robocar at Goodwood is led by automotive technology company ARRIVAL.
“The Goodwood hill climb presents a real challenge for level 4/level 5 autonomous driving systems,” said Sergey Malygin, chief of software and AI at ARRIVAL. “It is a narrow track with complex geometry. Turns and hills with a great deal of tree coverage mean you can’t rely on GPS/RTK signal for localisation. Use of all advanced sensors, including LiDARs and cameras with deep learning based computer vision methods are needed to perform well at this course.”