When HG Wells and other writers were defining a literary genre that came to be called science fiction, their words often sat alongside artists' visualisations of tomorrow's worlds. These sometimes included personal transport systems. Small unmanned vehicles would be available on demand to carry individual travellers between destinations. No need to drive yourself, the robot vehicle would be there.
A century on, and the dream persists, so much so that the EU's Sixth Framework Programme for research has backed a project that goes under the label CyberCars2.
http://cordis.europa.eu/ictresults/index.cfm?section=news&tpl=article&BrowsingType=Features&ID=90845
The idea behind this and related projects is to remove drivers from the transport system. Michel Parent, who directs the R&D programme into automated transportation at the French National Institute for Research in Computer Science and Control (INRIA), puts it like this: "the more you can automate vehicles and make them work on the existing infrastructure – in other words roads – then personal, rapid transit becomes feasible."
The results of this research appear in papers with titles bearing the usual technical jargon. Take, for example, 'Using Ontology-based Traffic Models for more efficient Decision Making of Autonomous Vehicles'.
http://www-c.inria.fr/cybercars2/publications/papers/7-1048151529.pdf/download
Fortunately, the paper is easier to fathom than the title. Among other things, it deals with the transport equivalent of a challenge that faces anyone who wants to work on autonomous robots. Do you 'plumb in' a complete map of the world and then let the robot work out where it is in that world view? Or do you leave it to the robot to sense where it is and build up its own memory of its environment?
The idea of an 'abstract world model' has plenty of advantages. For example, you don't have to keep the 'map' up to date, the system can adapt itself when there are road works and a route is closed. The disadvantage is that the system has to bristle with sensors that can build up the world view. The computational burden is also much heavier. As the paper puts it, "A disadvantage of the approach is the requirement for an additional processing step to generate the abstract traffic model, which is not a trivial task.'"
This paper is part of an interesting library of papers from the research programme, including some with more human friendly titles.
http://www-c.inria.fr/cybercars2/publications/papers/articles-published-by-members-of-the-consortium/
Programmes like CyberCars2 may not deliver those science fiction dreams, but by giving researchers ambitious and publicly exciting goals they do provide a test bed for less ambitious technologies that might be more immediately valuable. For example, fully autonomous vehicles would have to be able to communicate between one another, an idea that could also enhance safety and reduce congestion in existing vehicles.
CyberCars2 has given us a better understanding of how vehicles can communicate with one another, an essential prerequisite for cooperation between any vehicle, manned or otherwise. The programme has also begun to lay down, and test, the standards and protocols that are essential for any new communications technology.
With the bits and pieces slowly coming together, it may be not be quite 100 years before those century-old science fiction dreams become a reality. Better still, some of the developments might even begin to appear in the next generations of vehicles.
Categories: Commentary ,
Comments
All comments
You need to be registered with the IET to leave a comment. Please log in or register as a new user.