Getting out of a jam

Being stuck in a traffic jam is sadly a part of modern city life, and it’s not a green scene. Exhaust fumes from idling engines, and all that wasted petrol combusting internally mean it would be far better to avoid these snarl ups altogether. Just as they reinvented the wheel with built-in road condition detectors the University of Portsmouth has come to the rescue again, with their smart system called CADRE: ‘Congestion Avoidance Dynamic Routing Engine’. This system uses vehicles that are carrying a super-smart type of GPS, or global positioning system. This is a ‘smarter than the average car’ sat-nav, because it’s able to share information between all the other vehicles carrying similar smart electronics in a 10-mile radius. When the system starts to detect fellow vehicles moving slowly or stopping, it smells a jam, and starts to work out the best route for the driver to take to avoid the road congestion. The driver can then decide if they want to change their route and go the way the artificial intelligence suggests.

It’s all going fuzzy

The traffic jam buster’s artificial intelligence makes use of a clever concept called fuzzy logic. Normal (that’s un-fuzzy) logic is either true or false, but when it comes to making decisions we humans tend to prefer terms like ‘maybe’, or ‘possibly’, or ‘highly unlikely but you never know’. How is a poor computer with its binary 1 and 0, true and false way of seeing things going to cope with that kind of information? Well, turns out that the ideas behind fuzzy logic allow exactly that. Fuzzy logic is called a multi-valued logic – it’s not just true or false, there are other possibilities in between.

For example you detect that cars are moving at 5mph, but is this indicative of a traffic jam? We can have two concepts, jam or no-jam, and rate the two on a sliding scale between 1 and 0. If there definitely is a traffic jam then we would have 1.0jam (or 0.0no-jam which is the same thing), if there is definitely no traffic jam then our fuzzy logic says 0.0jam, 1.0no-jam – you get the idea. This mathematical approach lets us build in a bit of good old human experience. One driver might say that the information that cars were moving at 5mph was 0.7jam and 0.3no-jam, and they clearly think a jam is more probable. Another driver might feel that 5mph wasn’t so bad, and so for them it's 0.5jam 0.5no-jam. The mathematics of fuzzy logic lets the system build in these uncertainties.

Turn left?

Combining lots of similar human-defined rules for all the other traffic variables, the artificial intelligence system can make a guess based on the available information much like a person would. But, like advice from any ‘back seat driver’, even if it’s got some smart eco-friendly electronics and software in the front seat, its up to you to decide where you’re going.