Unmanned Aerial Vehicle(UAV) Task Coordination Method Study Under The Network Environment

The technologies of UAVs are developing fast, which makes the UAVs receive more and more attention. With the increasing stock and types of UAVs, a higher demand on how to use them more effectively has been advanced. The traditional usage patterns of UAVs, which can be described as “the user shall be responsible for its management”, usually for a single UAV, have seriously limited the sharing of information among the armies. It also leads to some other problems such as the management and safety of airspace and so on. To achieve unified management for UAVs in an information warfare, to make it different between the manager and the user is a trend of development for UAVs. Under the background, this paper mainly studies the classification and resolution of the order, multi-UAVs search multi-targets modeling under complex constraint, the preplanning and dynamic online replan for large-scale task coordination. The main work and contributions are summarized as follows:(1) On the basis of multiple target multi-uavs reconnaissance missions, further considering the complexity and dynamic characteristic about the tactical missions. Then presents an analysis of the whole work procedures of the MPS in details from the input of task orders to the output of the task assignment plan. Through this we will find out the correlation between tasks and UAV systems,and convert the description of tasks into the demand on the UAV systems. Owing to these works, quantized data expressions and basic models may be provided. So it is easy to summarize the military requirements on this system and build its architecture. In terms of the subsystems, it follows a comprehensive discussion on their functions and modules.(2) The format of the task order is always not unified in many researches and departments. So we actually cannot work out a task plan correctly and effectively. This paper summarizes some common types of the ISR(Intelligence, Surveillance and Reconnaissance) tasks, and explains the factors that influence the result of task allocation, for example, type, level, time window, etc. Based on that, we formulate a task order form that can be commonly used. Furthermore, a summary of the relations among these factors or with external conditions is presented, which can limits the choice of resources to actualize tasks.(3) Considering the constraints listed above, a multi-element task assignment model is built, with the maximum profit, minimum route, communication and time cost, and more tasks as accomplished taken as the performance indices. In order to reduce the planning time for large-scale task allocation, we choose a distributed computing platform and an improved genetic algorithm to realize parallel computing according to the characteristics of different Task-UAV sets. Then a global adjustment is made on the initial local plan at last. A dynamic re-plan on line is also taken into account to ensure that the plan can be carried out satisfactorily.