Cooperative Dynamic Task Allocation for Unmanned Surface Vehicle Swarms Based on CART Decision Trees

In common multi-boat task allocation, most of them adopt a global allocation method, which assigns all the tasks of the target to different unmanned boats for sequential execution at one time. However, in the ever-changing battlefield environment, there is a possibility that the unmanned boat will fail to perform the task. At this time, the target task needs to be allocated again, and a single global allocation cannot meet the task requirements. Therefore, based on the different types of target tasks, this paper adopts a cascade method and uses the Classification and Regression Tree(CART) to establish the classification model of the task target and the unmanned boat group respectively. According to the task requirements and task execution status, an event trigger is set up. When the event requirements are met, the improved crow search task allocation algorithm is called to update the task list of the specified unmanned boat. It not only solves the limitations of global allocation, but also improves the flexibility and adaptability of task allocation, ensuring that the unmanned boat group can complete the task efficiently and accurately in a complex and changing battlefield environment.