Neural network system of operational planningfor safe bypass of a group of obstaclesduring the aircraft route flight

Abstract

Problem definition of operating control of safety and the choice of an alternative of management of flight of the closest obstacle on the way of performance of route flight of the lethal device is formulated. Usually the configuration of route flight consists not only of sites of the rectilinear movement, but also from sites of a turn therefore it is very twisting and occupies a certain space owing to what there can be a dangerous rapprochement with the forbidden zones.Basic feature of this task is need quickly enough to check an exception of hit of the aircraft in the forbidden zone for what on its board it is necessary to use high-speed control algorithms of flight control and safety control. In this work the problem is solved in the assumption that obstacles have a rectangular shape and are randomly focused concerning the set line of a way, and the number of these obstacles and coordinate of their tops are set. For the solution of a task the neural network algorithm of safe flight of obstacles supporting two simple neural networks and two program blocks which are carrying out preparatory operations is offered. At the same time the first neural network chooses the closest obstacle on the way of route flight for what the first program block estimates coefficients of danger of all obstacles of time at the moment. Then the second program block determines coordinates of tops of the chosen rectangular obstacle, and in conclusion the second neural network automatically chooses the necessary top about which it is necessary to fly. Summarizing the aforesaid, we will note that in article neural network approach to the solution of a problem of expeditious definition of a dangerous obstacle in a way of route flight and the alternative choice of option of its flight is offered. The solution of a task by means of two neural networks significantly simplified their training. Computer modeling of a neural network system confirmed its working capacity and high speed.