Design of Language Model for Intelligent Ship Manoeuvring Simulator

Language model is an important module in speech interaction technology. The current ship manoeuvring simulator lacks language model in the field of ship, which leads to the accuracy of speech interaction can not be guaranteed. Therefore, it is very necessary to design a language model for intelligent ship manoeuvring simulator to improve the accuracy of speech interaction. By analyzing the modeling technology of language model, a special corpus for ship manoeuvring field is built, and three language models were obtained based on the n-gram language model. The problem of data sparsity was solved by combining Good-Turing discount smoothing and Katz smoothing. Combined with the confusion test and speech recognition experiment, the performance of the three language models was evaluated, and the most suitable language model for ship manoeuvring simulator was obtained.