Abu Dhabi to host $3 million competition to design autonomous drone system

Unmanned vehicles will have to search a group of ships to identify a 'rogue actor'

Abu Dhabi will host a competition with a $3 million prize fund that challenges engineers to design a system of unmanned vehicles to autonomously search a group of ships to identify a “rogue actor”.

The Mohamed bin Zayed International Robotics Challenge will stage a scenario involving a group of vessels in the ocean off the coast of the emirate in June 2023.

At least one of the boats involved will be a “bad actor”.

The system, which will use unmanned drones and boats, will have to identify which one is the rogue vessel.

Quote
What we will get out of this won’t be anything that’s ready to sell or produce
Dr Arthur Morrish, Aspire

The mission must be completed autonomously, with no human help.

The challenge, which has never been done before but is theoretically possible, is expected to be a spectator sport and will be televised live.

Open to international universities, research institutions, companies and individual designers from all over the world, it will involve a collaboration of unmanned aerial vehicles and unmanned surface vehicles, to perform complex navigation and manipulation without using satellite navigation or GPS.

The competition is organised by Aspire, the dedicated technology programme pillar of the Advanced Technology Research Council.

“There are a number of things that happen in the world to any country that has a maritime coastline,” Dr Arthur Morrish, chief executive of Aspire, told The National.

“So you have to worry about illegal fishing, piracy, smuggling, human trafficking and these are security issues. But they are not kind of navy security issues. They are coastguard security issues that impact people’s everyday lives.

“What we wanted to do was set up a challenge that would do what robots do best, which is the dull, the dirty and the dangerous associated with that specific problem.”

Competitors can use between five and 20 drones to survey a defined area, which will have fewer than 10 vessels. At least one of them will be rogue.

“So your job as a competitor is to design a fully autonomous system that will find that needle in the haystack, that bad actor that’s hiding amongst all the others,” Dr Morrish said.

“And you only have limited intelligence.

“So first of all you have to find it. And finding it means that your team is going to get that police report, that intel report, and you are going to give it to your system and from then on it’s on its own.

"You are not allowed to look at the feed it sends back to tell it whether it’s right or wrong. That’s the system’s job.”

There are several stages to the competition, which will culminate in the challenge in more than 18 months’ time.

White papers with ideas are being accepted already until the end of the year.

Anyone who can persuade the challenge they are capable will move to the simulation phase. No more than five teams will graduate from that to compete in the final in June 2023.

Dr Morrish admits that the task is complex, and there is a chance no one will win.

The total prize purse is $3,250,000 with the team winning the first prize taking home $2m.

The team placed second will receive $500,000 and the team finishing third will win $250,000.

A special cash award of $500,000 will be awarded to all teams progressing to the demonstration phase with the amount to be split among the teams.

“What we will get out of this won’t be anything that’s ready to sell or produce. If it was I would be surprised," he said.

“What I expect to happen is that this is the first step. We will look at what worked and what didn’t work and either Aspire or our customers will want to go and work with some of the competitors. The people we think will want this have very specific needs.

“I won’t be happy but I guess I will be OK if no one wins it this time, as I guess it was a significant challenge. And we will try it again.”

Abu Dhabi uses drones to lay seeds in mangrove regeneration plan

Updated: November 11th 2021, 12:10 PM
EDITOR'S PICKS
NEWSLETTERS