The simulation and training platform for ocean unmanned systems adopts an extensible and modular design idea. Based on the use process of ocean unmanned systems (AUV, ROV, USV, etc.) and the information flow transmission in their actual work, and in combination with the use and operation training requirements of relevant personnel, a digital simulation model is built, which can effectively simulate the dynamic/kinematic characteristics of single/cluster ocean unmanned systems in the marine environment, the use of sensor loads Interaction of environmental information and marine operation process.

The seabed terrain classification and recognition system is a software suite independently developed by the POET for automatic classification and recognition of common seabed terrain and landforms, including sample classification and recognition training system as well as image display and marking system. The software suite uses machine learning algorithm to classify, recognize and display seabed targets. The system supports standard sonar data format and has a friendly visual interface.

Underwater optical identification system is an automatic identification software suite for underwater optical targets independently developed by surging ocean. The software suite uses a deep learning algorithm to identify underwater targets at close range. The system supports standard underwater video data format and has a friendly visual interface.

The POET has a domestic leading multi AUV cooperative detection technology system. The system integrates: autonomous underwater target recognition technology based on machine learning;