Saturated diving helium voice interpretation method based on lexicon learning

Abstract

The invention relates to a saturated diving helium speech interpretation method based on lexicon learning, which comprises the following steps of: in a system comprising a diver, a correction network and an interpretation network, establishing a working language lexicon commonly used for saturated diving operation, reading by the diver in different environments, and generating a supervision signal and a vector signal of the correction network, wherein the correction network learns helium voices of different divers at different diving depths to obtain network correction parameters, and corrects the helium voices of the divers to obtain corrected voices; and the interpretation network learns the corrected voice and completes interpretation of the helium voice. By setting individual voice signal features and working language lexicon features of the diver in different environments, machine learning is introduced into helium voice interpretation, the problem of interpretation of saturated diving helium voice is solved, the requirement of a network for the number of learning samples is reduced, the adaptivity of the helium voice interpreter in different diving depths is improved, and the method is especially suitable for complete interpretation of helium voices in different diving depth environments.

Description

technical field [0001] The invention relates to helium speech interpretation technology, more specifically to a saturation diving helium speech interpretation method based on lexicon learning. Background technique [0002] The 21st century is the era of marine economy, we will obtain more than 50% of the clean energy and production materials needed by human beings from the ocean. Saturation diving has extremely important application value in the fields of nautical operations, ocean development, military ocean, sea rescue, etc., and is an indispensable part of the development of marine economy. [0003] Due to the particularity of the deep-sea operating environment and work content, many tasks in the ocean cannot be completed by manned deep-sea submersibles or underwater robots. Divers need to go directly into the water and be exposed to deep-sea high-pressure environments using saturation diving techniques. In order to meet the physiological needs of divers in deep-sea high...

AI Technical Summary

Problems solved by technology

[0004] At present, the existing helium voice interpreters for saturation diving at home and abroad are interpreting the helium voice through the helium voice interpreter in the diving chamber and manually adjusting the frequency domain or delay characteristics of the interpreter, which cannot adapt to the depth of saturation diving operations. , and the interpretation effect is not ideal. When the diving depth is greater than 200 meters, the interpretation quality of the helium voice interpreter drops rapidly, especially when the diver’s diving depth is changing, it cannot effectively interpret the diver’s helium voice

Artificial intelligence-based speech interpretation technology for saturation diving helium is currently in its infancy at home and abroad, and practical technical solutions are rarely seen

Images