Air control voice command recognition method based on deep learning

Abstract

The invention discloses an air control voice command recognition method based on deep learning. The method comprises the following steps: acquiring a voice signal to be recognized, and converting thevoice signal into 16-bit 16-kHz PCM audio data; building a deep network model; training the deep network model by using training data to obtain a voice recognition engine; performing voice segmentation on the audio data; and inputting effective audio clips obtained by voice segmentation into the voice recognition engine, and outputting a character recognition result. According to the deep networkmodel, a convolution module is used as a feature extractor; extracted feature data is processed through a reshape layer and a full-connection layer; sequence learning is carried out through a gating circulation unit; finally classification learning and decision making are carried out through the full-connection layer, so that a prediction result is obtained. According to the method, an artificialintelligence deep learning engine is adopted as a core, so that the method has the advantages of extremely high professional applicability and accent generalization ability, and lower data quantity dependence, and is obviously superior to a general voice recognition system in air control voice recognition.

Description

[0001] The invention relates to the technical field of voice processing, in particular to a voice recognition method in the field of air traffic control based on deep learning. Background technique [0002] With the rapid development of the civil aviation industry, a large number of aircraft and flights are added every year. However, there is a shortage of air traffic controllers for a long time, and it is conservatively estimated that there are thousands of people. Even though the relevant units of air traffic control have implemented a series of methods, such as the 4+1 training mechanism and other programs, there is still a large loss of air traffic control personnel. At the same time, due to problems such as lack of experience of new recruits, lack of training time and resources, the corresponding personnel benefits cannot be brought into play. The tension in the air traffic control industry has led to the overload of air traffic control personnel, resulting in potentia...

AI Technical Summary

Problems solved by technology

[0003] In 2016, Guilin Jingzhun Measurement and Control Technology Co., Ltd. used the pre-trained voice library of controllers for voice recognition. This method is limited by the existing voice database, and the recognition effect on voice information that cannot completely match the rules is not good. Accuracy is not high

In 2018, the 15th Institute of China Electronics Technology Group built an acoustic model based on continuous hidden Markov CHMM to recognize speech. The error rate is reduced, but DNN is easy to overfit and fall into local optimum, and its recognition accuracy is still lower than that of CNN-GRU neural network model

Images