The Problem of Forming a Training Set for Multi-Valued Associative Memory Used in Trajectory Classification Depending on Linguistic Information about the Flight Situation

The problem of forming a training set for a multi-valued associative memory is considered. The concept of multi-valued associative memory is a new one; it generalizes a similar notion in fuzzy neural networks. In multi-valued associative memory, both inputs and outputs and weights are elements or subsets of a lattice of linguistic variables. A multi-valued associative memory is used to decide on the aircraft control mode.