The process of USV swarm target defense mission is described based on OODA (Observation, Orientation, Decision and Action) ring, and the architecture model of USV swarm target defense mission is studied according to DoDAF (Department of Defense Architecture Framework) operational viewpoint and modeling language SysML. It is proved that the method proposed by this paper can optimize the top-level design of USV swarm target defense mission architecture, and facilitate the understanding and communication of researchers in different fields to the top-level mission.
To meet the requirements on both real-time and lossless radar video transmission, a method based on the combination of VarByte algorithm and RLE compression algorithm is proposed to save the storage and transmission of large capacity radar data of USV. Because the digital radar video data has the characteristics of information redundancy, so according to the principle of data compression, this paper gives a data compression storage method. The method based on the combination of VarByte algorithm and RLE compression algorithm. The storage method can reduce the amount of radar video data of each frame to about 3%. At the same time, in order to reduce the bandwidth consumption of radar video, this paper proposes an image compression transmission method based on PNG.
To address the problems of difficult automation of intelligent algorithm deployment, low deployment efficiency, and high requirements for the configuration environment on the craft-side, a complete intelligent algorithm deployment method from the intelligent platform to the craft-side is designed and implemented. Based on the TensorFlow Serving service and the Docker open source container engine, it automates the deployment of intelligent algorithms and model version management, and reduces the system overhead on the craft-side.