Modern electronically controlled automatic transmissions no longer have a switch for selecting between “Power” and “Economy”. From the sensors at hand, the control unit determines whether the driver prefers to go for power or economy. But how does the control unit decide? A co-pilot would probably be in a better position to judge from a given route section whether the driver was thinking of performance or economy. The task of fuzzy logic is to translate the rules of human thought for the computer.
The word “fuzzy” is used whenever at least two variables that are associated with each other in a special way have to be compared. Instead of ‘true’ or ‘false’, several values can be distinguished, such as ‘very high, ‘high’, ‘medium’, ‘low’ and ‘very low’. These are then weighted and, finally, balanced against each other.
Co-pilots thus have widely differing ways of coming to the conclusion, ‘power driver’. A driver may accelerate very powerfully over a brief period, or may select a more than medium gas-pedal position for a longer time. A single burst of high acceleration can then be compensated for by gentler driving. And precisely this sort of case is typical of fuzzy logic. Although it simulates human thought, it must nevertheless come to an unambiguous decision: ‘Power’ or ‘Economy’. But this is not much of a challenge, for fuzzy logic can also compare variables with each other and come up in the end with an exact figure, say for the degree of opening of a valve.