Multivariate algorithms for dangerous turbulence detection in weather formations

This paper briefly describes methodologies of turbulence detection and classification. Proposed methods based on neural network approach. One method provides multivariate analysis of measured variables. Another one based on spectrum shape analysis. These methods are very flexible and allow reduction of calculation time and can be successfully implemented for warning purposes in airborne systems.