Mention neural networks, and most people think immediately of artificial intelligence. However, this technology is helping engineers solve problems in all kinds of unexpected areas.
For instance, researchers at Daimler Chrysler are currently developing a neural-net technology for a 4 cylinder petrol engine which selects the ideal ignition firing point 6,000 times a minute. Sensors measure the in-cylinder combustion pressure and networked model neurones (essentially mathematical formulae) evaluate this information and determine the best firing point for the fuel-air mixture at any given moment.
The technology may also help to prolong the life of older engines. Combustion chamber pressures vary greatly over the lifetime of a petrol engine, and, as the neural network understands the connection between these pressures and ignition timing, it is able to adapt and change the firing points accordingly, thus extending service life and using fuel more efficiently.