From smart to mindful cars: taking up the self-driving challenge
With about 90 % of car accidents being due to human errors, a world of driverless cars sounds like the Holy Grail of road safety. But to become more efficient than humans, cars need to understand each other and make quick and sound decisions based on the information gathered. The I-GAME project is aiming to put such interoperability at the forefront of European research efforts. If there is one thing in the world that reminds us of how time flies, it’s the look of children when we mention things like typing machines, faxes or a world without the internet. Soon enough, we may have to add yet another concept to this list: that of cars relying on a driver. The idea may still sound like science fiction, but cars that drive themselves are actually just around the corner. Technologies such as intelligent parking assist systems, dynamic cruise control or lane departure warning systems are becoming standard in new cars. Major high-tech companies are experimenting with automated driving in real-life conditions, and some car manufacturers are already rolling out software updates that see their cars drive themselves on highways or pick you up at the entrance of your favourite shopping centre. What’s still missing, however, is a guarantee that a world full of automated cars is realistic from the perspective of safety and traffic fluidity. The cars of the future will need to be able to ‘talk’ and to understand each other in order to avoid crashes and optimise decisions — all this in situations that can sometimes see hundreds of cars manoeuvre at the same location. The objective of the EU-funded I-GAME (Interoperable GCDC AutoMation Experience) team is to enable this interoperability by developing generic, resilient and fault-tolerant technologies and testing them in its ‘Grand cooperative driving challenge’ (GCDC) which is due to end in June 2016. This challenge will see teams compete against each other in designing and implementing the most effective cooperative vehicle system possible. The research*eu results magazine had the chance to discuss the project status and the potential of self-driving cars with Bastiaan Krosse, Jeroen Ploeg and Almie van Asten, respectively programme manager automated driving, senior scientist cooperative driving and project coordinator of I-GAME. What are, according to you, the main benefits of automated driving? Bastiaan Krosse: The main benefits with automated driving are more comfort, more safety, more efficiency and more accessibility to a larger community, for example the elderly. The benefits in terms of safety have already been observed with commercially available technologies such as adaptive cruise control and lane assist. Then, when it comes to efficiency, automated driving translates into more efficient use of fuel or batteries, roads and even parking spaces. Safety indeed seems to improve with each automation technology. Do you think automation as imagined by I-GAME bring us closer to zero fatalities on European roads? BK: I-GAME is about cooperative automation, which means interaction and cooperation between vehicles. Implementing such cooperation — which implies that each vehicle will ‘know’ the intentions of others on the road — will definitely enhance safety. European statistics show that about 90 % of traffic accidents are caused by human errors. The combination of both V2X and automation will potentially have a positive impact on safety. What are the main barriers to a wide-scale adoption of automated driving and how do you plan to overcome these? BK: Before cooperative automation can be considered for deployment on European roads, we first need to achieve interoperability between OEMs in terms of communications means and channels. The interactions and message sets must be aligned at European scale. Another significant hurdle is the robustness of the vehicle systems under all operational conditions. I-GAME challenges vehicles in complex scenarios with high requirements for the performance and robustness of the system as a whole. But a widely-accepted testing and validation process for automated vehicles, and eventually their certification, is yet to be set up. One of the objectives of the project is to develop a safety assessment procedure for the vehicles that enter the I-GAME challenge. This procedure can be a first step towards a validation process of cooperative automated vehicles in real-world situations. Where do you stand with the development of the I-GAME architecture? Jeroen Ploeg: The I-GAME architecture is characterised by the four main components. First, we have the overall system architecture which was set up in June 2014 and provides the main framework for further developments in the project. Then, there is the interaction protocol consisting of a wireless message sequence and in-vehicle controllers to execute our two scenarios. The protocol for the first scenario (cooperative and automated merging) had to be redeveloped because of its lack of accuracy in practice, so we came up with an improved version which will be tested over the coming months. The interaction protocol for the second scenario (cooperative intersection), on the other hand, has been thoroughly designed on a simulation level and will be implemented on benchmark vehicles in the coming months. The third component of the I-GAME architecture is the in-vehicle software architecture. It has now been fully determined, although adaptations may still be required as a consequence of scenario 2 interaction protocol implementation. Final piece of the puzzle, the communication architecture is based on the current ETSI C-ITS standards on Cooperative Intelligent Transport Systems. We want to use the currently available commercial Vehicle-to-Vehicle communication equipment to support multi-vendor solutions. In order to do this, we utilise currently available message sets as much as possible. But because current messages sets are specified mostly for road and vehicle condition awareness, and not for automated driving interactions, we have been designing extensions or new messages to support the complex interactions needed in GCDC scenarios. Interoperability is necessary, but it can also be dangerous under the threat of hackers. What are your plans in this regard? JP: Although it is acknowledged that interoperability poses a threat to security, the latter topic is not included in I-GAME for reasons of focus. However, ongoing developments in this field are closely monitored. A follow-up project is currently being set up, focusing on truck platooning with a significant involvement of several truck OEMs, which explicitly addresses security among other things by investigating the impact of security measures on automation performance due to, for instance, increased latency. Big companies like Google or potentially Apple are working on their own automated driving solutions. Is this compatible with the open approach of I-GAME? Almie van Asten: External companies are supplying the highly automated vehicles used within the I-GAME project, and the communication part will also be multi-vendor. Enabling communication between multi-vendor vehicles and having them working together requires compatibility of interaction protocols and message sets, but not of specific hardware. Therefore any party who wants to use their automated vehicle supported by V2X and is willing to use the protocols and message sets (on a software level) provided by the project is welcome to join. Actually we challenge companies like Google and Apple to compete and cooperate in the 2016 GCDC. Speaking of the upcoming GCDC, why did you choose to involve external stakeholders in the form of a driving challenge? Are you happy with stakeholders’ responses so far? AvA: To get wide acceptance of the developed interaction protocols and message set, we want maximum input and multi-vendor implementations in different hardware configurations. This challenge will accelerate the developments of the protocols by evaluating them in real-life conditions. As such, it is a first step towards the standardisation and implementation of these protocols and message sets for cooperative automated driving. For further information, please visit: I-GAME http://www.gcdc.net/i-game
Countries
Netherlands
