Automatic Deployment Method of USV Software Based on Functional Domain

摘要

In order to meet the unmanned requirements of Unmanned Surface Vessels (USV), aiming at the automatic deployment of USV software system, based on functional semantic analysis and rough set reduction, the functional domain division of USV software is realized to support automatic software deployment. The Bidirectional Encoder Representation from Transformer (BERT) model is pre-trained by the relevant standards and design documents of USV, and the semantic correlation between the function descriptions of USV software services is analyzed by using the trained model. The rough set information system for USV software system is constructed. The appropriate functional domain partition is extracted by rough set attribute reduction method, and the functional domain attribution of software service is obtained according to the equivalent partition of rough set. So USV software system can use functional domain to deploy automatically. The results show that the proposed method can meet the requirements of automatic deployment of USV software services developed from different sources and in different periods, and can realize automatic re deployment according to task changes.