Aircraft Flight Path
Fuzzy Logic predicts Aircraft Flight Path
By Bill Edisbury, Npower Ltd. UK, May 1999
Researchers at National Air Traffic Services (NATS) in London, have demonstrated that fuzzy logic can accurately predict aircraft flight paths. Using fuzzy TECH fuzzy logic software from Manchester-based Npower, NATS were able to compare flight path predictions generated by fuzzy logic against those produced by conventional linear extrapolation. For the flights considered, fuzzy logic proved to give more accurate predictions of aircraft position 88 seconds into the future.
"Fuzzy logic is proving to be one of the more useful techniques that we're evaluating to predict flight paths," says Julia Sonander, Research Manager, Research and Development Group, NATS. "We are looking into other data mining techniques, but fuzzy logic has a number of attractions. For example, fuzzy logic generates a set of rules for prediction rather than giving a 'black box' solution. These rules could be validated and implemented in conventional software."
A number of ATC tools, including current collision avoidance systems, incorporate flight path prediction, of which aircraft height is a particularly important element. Generally, such tools predict the height of an aircraft by extrapolating from its current height and assuming it will continue to climb at its current rate. Fuzzy logic uses a set of 'rules' to predict the outcome of a climb, for example, an aircraft levelling off at a particular height, based on an analysis of the historical flight data held by NATS. If fuzzy logic proves to provide significant improvements in track prediction it could be used as a component in future ATC tools.
Initial research results have been promising. Further research is now required to see how fuzzy logic predictors can be developed and to assess their performance for a larger data set. NATS is currently undertaking an extended trial of fuzzy TECH.
NATS undertook the three-day evaluation of fuzzy TECH with consultancy support from Npower, under the DTI's Smart Software for Decision Makers programme. "We are delighted that fuzzy logic was able to demonstrate real improvements in flight path prediction," said Bill Edisbury, chairman, Npower. "Many companies are now finding that fuzzy logic has a great deal to offer when trying to make sense of a mass of data that on the surface may look pretty meaningless, for example, in analysing customers spending patterns."