430pm November 11, 2o10, School of IT Building, University of Sydney.
In his final public spiel as Director of the Australian Centre for Field Robotics, Hugh Durrant-Whyte connected his personal motivations and values with his ambitious goals to create a successful research centre. ACFRS is the darling of Sydnovate, the awkwardly-named commercial arm of the University, which hosted Durrant-Whyte’s talk. Durrant-Whyte leaves ACFR at the end of the year to become CEO of National ICT Australia (NICTA).
Durrant-Whyte begins in a candid way by claiming that ‘We’re academics to have fun; and make money along the way’. His approach is casual self-confidence, making light of his sizable ambitions and achievements. He resists wearing a suit, and takes pride in not being able to drive (something that became a problem at least once in lacking an intuitive understanding of how vehicles drive).
He says he was first attracted to robots through science fiction. Using a taste for sci-fi to explain the personal motivations of roboticists is almost too easy, but this genre can’t be underestimated in its capacity to help generate gadget-building subjectivity. The fact that many of these sci fi narratives are dystopian warnings about the perils of future robots can’t compete with the ‘cool factor’ of the shiny fictional hardware it’s own autonomous will to become actual.
In his early work in the UK in the early 1990s, Durrant-Whyte attempted to solve problems of navigation and control. Once he moved to Australia, this problem would continue to drive his work on a much larger scale in loading ships and digging up ore in mines. The main advances since the early work, though, were in knowing how to frame engineering solutions that would be attractive to industries with problems robot systems might solve. And he saw Australia, with so much space, far from anywhere, as offering many of these attractive problems. It also has the advantage of more open airspace for UAVs (Unmanned Aerial Vehicles) and prettier oceans for marine robotics.
His advice:
1. Define your expertise. Now and your future vision.
2. Build the team. Motivation to have an impact. Be generous.
3. Try to be the best. Uniqueness; value to industry. Relevance to Australia.
4. Start small, and get bigger. At one point say, ‘I’ll do that report but it will cost $40k’
The big break for the ACFR came at a point of crisis on the waterfront, when Chris Corrigan was taking on the unions, says Durrant-Whyte. Captains of industry had long perceived the waterfront was holding the nation to ransom. I’ve wondered the extent to which port automation was a union-busting initiative, and how much it was an improvement in process.
By 2000 ACFR was able to demonstrate a multiple-million dollar laser-guided straddle carrier, which can pick up 3 shipping containers at one time. Durrant Whyte admits that the demo came close to running over Corrigan. This work resulted in a commercial spin-off and the development of an automated port in Brisbane in 2003. They had bridged ‘the chasm between research and application’ with 36 automated straddle carriers in Brisbane; run remotely from Sydney, now signed off from University.
The second example Durrant-Whyte gave was mine automation. ACFR’s early approaches proposed individual robots, but after more research they realized that the mining companies (and Rio Tinto) wanted to automate the entire process from discovering the minerals through getting them out of the ground and carting them away. The whole process, says Durrant-Whyte, centres on information, and systems that allow operators to have a complete picture of the state of the mine: ‘just like a video game’.
ACFR boasts alliances with other major players: US Army, US Air Force, BAE Systems (an innovative UAV project that crashed on its first flight); a spin-off company Marathon, secured a $57m contract with the US Marines for a system of robot-controlled targets, based on the Segue.
His strongest message is always to own your IP. It allows you to continue with the research (companies want to own it, but really only want the exclusive rights to use it). Publish after patenting. Write down what you’re going to do (avoid requirements creep). Manage expectations. Universities are not a charities for cheap research. Limit what you promise and stage your projects.
Mining is going to dominate robotics in Australia and world wide. There will be more flying robotics for the environment (such as helicopters that can spot weeds and blast them. Look for social robotics and aged care.
Robots are cool and fun.
Leave a Reply