Planning for driverless car 'races' begins
ByAutoSport
Published 07/06/2017 at 17:13 GMT
Roborace's second driverless racing car is due to be introduced this month and the initiative will now begin to work out how the first competition between its cars will work.
Eurosport
Image credit: Eurosport
The Robocar made its public debut in support of Formula E's Paris ePrix last month, completing an autonomous lap of the circuit on race day.
Previously Roborace had only run its development mule, christened DevBot and based on an LMP3 car and with a cockpit so it could use a human driver while testing the software, at Formula E events.
Roborace now has three versions of 'DevBot', and chief technology officer Bryn Balcombe told Autosport the second Roborace was due some time in June.
Balcombe said that gave his company "a lot of opportunities to try different things".
Asked when a first 'race' could be expected, Balcombe replied: "It's really defining what Roborace is as a racing competition.
"If we stick to the traditional format of 10 cars just following each other around, there's very little differentiation you can have - it becomes lap time differentiation.
"We're looking at formats for the short-, medium- and long-term.
"It's not been defined yet but it includes road-relevant features, like traffic on the track.
"If you're talking about autonomous driving competition you need to push the boundary of the autonomous driver - not powertrain or chassis or aero or tyres.
"It's the ability to perceive, think and act."
Balcombe said "it's quite tight" between getting the cars back from Formula E's July 29-30 Montreal season finale and the start of the next season.
That suggests the 2017/18 curtain-raiser, set to take place in November in Hong Kong, might be too soon for Roborace to hold its first competition.
Roborace's priority for the remainder of the current Formula E season is to demonstrate to prospective entrants that the infrastructure, as well as the technology, is ready.
"The rest of the season for us is incremental developments and developing the platform to invite the teams to join," he explained.
"The team write the AI drivers, they take the risk, they define the performance level. The hardware is defined by us.
"We're gathering data so we can share that with all the teams so they can use that for training, and we're building a simulator environment in the background.
"So the objective is for us to show there's a working platform waiting for teams to arrive."
The Robocar's lap in Paris was at a very slow speed, but Balcombe insisted its pace will improve.
"The first thing is the car understands its environment in order to work out what action to take," he said.
"Speed is an action - it's making the decision, 'can the car proceed at a speed safely?'
"What we're proving is it knows where it is at all times, it knows where other objects are.
"Just like a human driver it can then start to push the limits."
Scan me
Related Topics
Share this article
Advertisement
Advertisement