Jaguar Land Rover to lead academic collaboration on driverless car development involving 10 UK universities and TRL. Jonathan Newell investigates.
With few technological barriers remaining in delivering individual elements of vehicle autonomy and so much to be gained financially in the automotive industry, vehicle technology appears to be following a path of the inexorable development of sensors, connectivity, control systems and on-board computing.
However, despite the progress being made in individual technologies, there remains an enormous challenge in plumbing them all together into a single coherent driverless car that can be used in mixed traffic and will be accepted by the consumer base.
For autonomous vehicles to operate correctly and safely, the whole system needs to be seamlessly integrated and robust. This is where the main challenges to true autonomy lie. Each element of the sensing and control systems must operate together, along with external systems, which could be other vehicles, the road infrastructure or wider “cloud-based” services.
Autonomous control system must also integrate with the most important element of all – the driver. The driver cannot be forgotten, because autonomy will progress in stages and each stage involves varying degrees of “driver necessity”.
Indeed, a certain degree of driverless control has already arrived, with the general availability of such features as self-parking, adaptive cruise control, lane departure warning and autonomous emergency braking.
As such features extend and become increasingly co-dependent and integrated, the degree of autonomy will increase, but the driver will remain present to take control when necessary as well as when desired.
Car manufacturers, predominantly in the UK, Germany and Japan, are further sharpening the cutting edge of driverless cars. It is in their interests to develop technology that is marketable now, even though money is being spent on long-term research for autonomous concept vehicles.
One manufacturer taking a pragmatic approach is Toyota with its recent release of the “Highway Teammate” prototype vehicle. This modified Lexus GS that has been designed to operate in hands-off mode from the on slip-road to the off slip-road. By limiting autonomous control to motorway driving only, Toyota expects to have its Highway Teammate marketable by the end of the decade.
The on-board systems of the Highway Teammate car are relatively simple, achieving lane and distance control using established lane recognition technology, radar and LiDAR sensors, blind spot cameras and adaptive cruise control, as well as advanced mapping and navigation systems.
However, even with such limited autonomy, the car is still dependent on communication with the infrastructure to be able to recognise road conditions, such as variable message sign instructions, temporary speed restrictions and lane closures. It is also dependent on the quality and clarity of road markings.
Tests are currently taking place on the Shuto Expressway in Tokyo which is geared up for such pre-requisites, but it is unclear how the car will be able to operate outside such controlled test conditions.
In Germany, Daimler is active in the promotion of dialogue with research institutes and universities on the social implications of autonomous driving and has so far invested around EUR1.5 million, supporting over 20 scientists looking into the social effects of autonomous driving.
Daimler is also putting resources into developing the answers to legal and ethical questions, with the steering committee on this topic being led by top members of Daimler’s management board at the company, including Professor Thomas Weber, the man in charge of Group Research and Development at Mercedes-Benz Cars.
UK Collaboration with Academia
The UK is not lagging behind in defining the direction for the future development of driverless vehicles and, rather than letting the lead be taken by the automotive industry, intends to put itself firmly in control of how the technology develops.
According to the UK Government’s Business Secretary, Sajid Javid, only research programmes can help to keep the country at the forefront of technology development in vehicle autonomy. “The UK Government has no intention of being a passenger in innovation, so is pioneering autonomous car technology in partnership with industry,” he said.
The partnership getting the Government so excited is an £11 million autonomous vehicle research programme being undertaken by Jaguar Land Rover in partnership with the Engineering and Physical Sciences Research Council (EPSRC). This was announced in early October.
The project has engaged 10 UK universities and the Transport Research Laboratory (TRL) to work on five projects that were selected as the most appropriate for integrated driverless car development.
Of these five projects, just two are predominantly technology driven, examining
(a) sensor technology for terrain mapping and road user classification; and
(b) the autonomous car as part of a wider distributed system of connectivity and control incorporating cloud computing.
Commenting on these collaborative projects, Jaguar Land Rover’s Director of Research and Technology, Dr Wolfgang Epple said, “To realise the future potential for fully autonomous vehicles, we need to give drivers, pedestrians and other road users the confidence that a car driving around with little or no human input is a safe, viable and rewarding experience. These collaborative projects will bring some of the UK’s leading academics together with our autonomous driving team to address the fundamental real-world challenges that are part of our journey towards autonomous driving.”
Autonomy within a distributed system
One key project within the collaboration will examine what needs to be addressed before driverless cars can be allowed on the roads without jeopardising the safety of other road users, including cyclists and pedestrians.
To study these technologies, TRL will work with the University of Surrey, Warwick University and Imperial College London on a project to understand how distributed control systems and cloud computing can be integrated with vehicles. The project aims to design and validate a novel Secure Cloud-based Distributed Control (SCDC) framework for connected and autonomous cars.
TRL’s Chief Scientist, Alan Stevens says, “The project will explore how increasingly automated and connected vehicles can operate safely and securely when connected to each other and, via the road infrastructure, to cloud-based resources. Ultimately the aim is to develop a secure framework that will enable the implementation of safe and robust semi-autonomous functions on future cars in the short term, and fully autonomous cars in the long term.”
The Five “Towards Autonomy – Smart and Connected Control” (TASCC) projects
|Pervasive low-TeraHz and Video Sensing for Car Autonomy and Driver Assistance||University of Birmingham, Heriot-Watt University, University of Edinburgh||To map terrain and identify hazards such as potholes as well as to detect and classify other road users such as pedestrians, cyclists and other vehicles using a combination of low-THz sensors and advanced video cameras.|
|Human Interaction: Designing Autonomy in Vehicles||University of Southampton, University of Cambridge||To study a wide range of drivers with different experience in simulators and in vehicles to examine how they react to vehicle autonomy and how best to design the driver-automation interaction.|
|Driver-Cognition-Oriented Optimal Control Authority Shifting for Adaptive Automated Driving||Cranfield University, UCL||To conduct a comprehensive study of driver attention and cognitive control characteristics when interacting with the vehicle controller to achieve a safe engagement and smooth and swift control-authority shift between the driver and the vehicle controller during adaptive automated driving.|
|Secure Cloud-based Distributed Control (SCDC) Systems for Connected Autonomous Cars||University of Surrey, Imperial College London, University of Warwick, Transport Research Laboratory (TRL)||To examine autonomous vehicles as intelligent agents that learn from their environments and exploit various sources of information to become increasingly autonomous using modern wireless technologies and the power of cloud computing.|
|The Cooperative Car||University of Warwick||The project will develop intelligent driver systems and cooperation and behaviour modelling techniques that learn about drivers, enabling vehicles to cooperate with each other and with urban transport infrastructures.|