Research on the task decision algorithm for unmanned aerial vehicles based on the robot operating system and large language models

With the progress of artificial intelligence technology, the automation and intelligence of Unmanned Aerial Vehicles (UAVs) in complex environments have become increasingly significant. This paper introduces a UAVs task decision-making algorithm framework that leverages the Robot Operating System (ROS) and the large language model Qwen2-VL to enhance the efficiency and precision of UAVs task decision-making in dynamic settings. By integrating the large language model with multimodal data inputs, Qwen2-VL intelligently interprets task instructions and breaks down complex tasks into executable subtasks. To optimize the task execution process, this study employs the Moss Growth Optimization (MGO) algorithm to refine and optimize the task sequence, ensuring the efficiency and rationality of task execution. The experimental results demonstrate that the framework can effectively improve the accuracy and efficiency of UAVs task execution and possesses strong adaptability, optimizing decision-making processes in complex operational environments. Experimental validation shows that integrating Qwen2-VL and MGO algorithms within the ROS system not only enhances UAVs intelligence but also offers new methodologies for automated systems and intelligent decision-making.