Research on an Embedded Intelligent Computing Technology ^†

In this paper, we study the realization in the embedded platform of an Artificial Neural Network (ANN) to explore the feasibility of a lightweight neural network. The purpose is to provide theoretical support for the intelligent information processing system based on a neural network under the airborne embedded environment, and then to support the intelligent demand of the airborne system in the future complex battlefield. By analyzing the application prospect of ANN in the aviation field, this paper will design an embedded hardware platform for UAV autonomous intelligence information processing tasks, and realize a convolution neural network (CNN) on it, using an image recognition task to train and test.