Driverless vehicles: The future of car ownership is Pay-As-You-Go – IoT Now – How to run an IoT enabled business
Once the preserve of science fiction, commercially available autonomous vehicles (AVs) will soon be a reality. AVs are currently being tested on UK roads by major vehicle manufacturers including Nissan, Volvo and Ford, and 2019 is being cited as a realistic year for launch in some limited geographies.
From forecourt to scrapyard, a new car in the UK lasts an average of 13.9 years, which is why if you got one today, it might very well be the last car you buy.
Professor Andrew Blake and Doctor Subramanian Ramamoorthy bolster FiveAI’s academic roster, bringing fully autonomous vehicles closer to reality
Professor Andrew Blake is the Director of the Alan Turing Institute, the UK’s national research institute in Data Science, and an Honorary Professor in information engineering at the University of Cambridge. He is one of the world’s leading researchers in computer vision having completed a PhD in Artificial Intelligence at the University of Edinburgh in 1983, he later moved to Microsoft Research in Cambridge to found the Computer Vision Group which developed the algorithms for image processing and 3D vision underpinning several Microsoft technologies, including Kinect.
Dr Subramanian ”Ram” Ramamoorthy is a Reader in Robotics in the School of Informatics at the University of Edinburgh, affiliated with the Institute of Perception, Action and Behaviour. He leads the Robust Autonomy and Decisions research group, whose focus is on achieving interactive intelligence in autonomous robots capable of working with humans and other robots. He is an Executive Committee Member of the Edinburgh Centre for Robotics, one of UK’s leading centres for robotics research.
These latest appointments follow FiveAI’s announcement in October that Professor Philip Torr, head of the University of Oxford’s Torr Vision group, a global, state-of-the-art research team comprising 25 post-doctoral and PhD students doing pioneering work in the field of machine learning for computer vision would take up the role of FiveAI’s Chief Scientific Advisor.
“Fully autonomous urban vehicles need the industrialisation of emerging science from the fields of computer vision and machine learning to meet the clear safety goals, particularly in their ability to recognise objects, their states, motions and localities to the highest possible levels of accuracy.” said Professor Andrew Blake “Across the world, there is now a race to build an intelligent pipeline of technologies in fields where UK academics have been consistent pioneers and where we have some of our very best people engaged. I am excited, alongside Phil and Ram, to help FiveAI, a British company, to succeed in leveraging our undoubted fundamental research leadership into winning that race and so create a global leader as this market explodes.”
“Even once fully autonomous vehicles can accurately perceive the world around them, they must plan their paths and interact with other scene actors cooperatively and safely. In an urban scene, that has become one of the most challenging research problems to solve since the numbers of actors and behaviours can be vast and the cost of making an error so huge,” added Subramanian ”Ram” Ramamoorthy of the University of Edinburgh. ”Not surprisingly early autonomous vehicle programs focused on reactive collision avoidance resulting in classic ‘frozen robot’ and other unusual behaviors which arguably made our roads less safe. But we now know how to use context-sensitive observations and learnt behaviours to predict what actors will do, update those predictions at high frequency and so build systems that predict and cooperate with other actors just like safe human drivers. Developing, integrating and commercialising this novel science into systems will help differentiate the FiveAI solution and I’m delighted to be working alongside FiveAI’s world-class team.”
Stan Boland, co-founder and CEO at FiveAI, said: “Safe fully autonomous driving in the urban environment remains the ultimate unsolved challenge. Achieving the perception accuracy and motion prediction performance required in the kind of cities we have here in Europe means solving some of the most complex problems in artificial intelligence and computer vision, and that requires the very best minds in their respective fields. Andrew, Ram and Phil are among the world’s leading authorities in visual perception, decision making and motion planning so we are honoured to have them aboard at FiveAI.”
The UK’s digital and culture minister says the CES tech show’s chief was wrong to claim the government is doing too little to support its start-ups at the Las Vegas event. Matt Hancock was responding to criticism that his team’s efforts were a “source of embarrassment” when compared to France and other countries’.
Driving smarter: solving the challenges of autonomous vehicles with AI – IoT Now – How to run an IoT enabled business
There is no longer any doubt that autonomous vehicles (AVs) will soon be a reality on our roads. However, there are multiple challenges which must first be overcome. Initial AVs will be restricted to operating on certain “known” roads, plus these early vehicles won’t be truly autonomous and will require human control and oversight at times.
It’s a sobering fact that even in this technological age, we still see 1.2 million deaths on our roads each year. Furthermore, studies show that nearly 95% of road accidents are actually the fault of a human driver.
Serial tech entrepreneur Stan Boland has sold three tech firms for a total of over $1bn, but is now abandoning his tried and tested semi-conductors and turning his hand to driverless cars. “Our previous companies have all involved microchips, but this one won’t have any chips,” says Stan Boland of his latest venture, FiveAI.
Professor Torr leads the University of Oxford’s Torr Vision group, a global, state-of-the-art research team comprising 25 post-doctoral and PhD students doing pioneering work in the field of machine learning for computer vision.
FiveAI, funded in July 2016 by Amadeus Capital Partners, Kindred Partners and Notion Capital, has established its goal to deliver the safest solution for Level 5 autonomy in driverless cars and other autonomous vehicles. Level 5 autonomy describes vehicles operating safely in complex urban environments without any driver involvement whatsoever. Current commercially available systems are classed as Level 2.
After achieving his PhD in the field of computer vision at the University of Oxford, Professor Torr’s career includes pioneering work in computer vision at Microsoft Research both in Redmond, WA and Cambridge, UK before returning to academia at the University of Oxford. His work covers fields such as object recognition, 3D reconstruction, tracking, and scene understanding and attracts the highest honors in the field of computer vision, including the biennial ICCV Marr Prize. Professor Torr is regularly awarded top paper status at the most prestigious computer vision conferences including IEEE CVPR, ECCV, ICCV and BMVC, as well as receiving honorary mentions at NIPS machine learning conferences.
“We’ve made tremendous advances over the last few years in our field of computer-vision,” said ProfessorTorr. “We have developed advanced techniques that deliver superior performance to human abilities on many perception tasks that are crucial for autonomous driving. I’m very excited to be working with such an accomplished team of engineers and business leaders that can build on our research to make autonomous vehicles a reality on a global scale.”
Commenting on Professor Torr’s appointment, Dr Stephen Allpress, co-founder and Chief Technology Officer of FiveAI, said: “We are thrilled to be working with Professor Torr and it’s a huge advantage that we can access a deep technology base to help power and guide our technology platform. Cutting-edge perception capabilities are a crucial element of realising Level 5 autonomy, meaning that we need to go well beyond the relatively simple task of reliably recognising other vehicles on a highway. Instead we must accurately perceive all possible dynamic agents in the scene, including pedestrians and cyclists, with a vastly superior precision-recall performance and at normal road speeds in different lighting and weather conditions.
“Professor Torr’s group is recognised as a global authority in the vital science of computer vision and its expertise will be vital as we work to achieve this ambitious goal. His contribution and supervision of FiveAI’s work in computer vision will be a key ingredient to ensure we can deliver the safest autonomous vehicle technology in the world.”