| With the development of today’s digital intelligence,non-cooperative target reconnaissance systems in complex scenarios have become an important research direction for civil security and military frontier defense.With the emergence of distributed platforms,reconnaissance task scheduling based on distributed platforms has become the research focus of non-cooperative target reconnaissance systems.This thesis will design and implement the reconnaissance task scheduling algorithm based on the distributed platform around the distributed platform of the cloud-side-end architecture.The main contents are as follows:(1)Combined with the requirements of reconnaissance tasks,a bidirectional auction algorithm based on state merging and multi-round allocation is proposed.The detection state of the detection source obtained by the traditional bidirectional auction algorithm is different for the target,that is,a detection source can only see the target at the same time.a goal.In order to see as many targets as possible at the same time,in this thesis,after an allocation is completed,the detection status of multiple targets whose detection source status is within the set range is merged as the action of the detection source at the next moment,which cannot be merged.Then the pairing is canceled,and the new merged state is used as the initial state of the probe source for the next round of allocation until all target allocations are completed.The algorithm uses both the target and the detection source as the allocation subjects,so that the detection source cluster can observe as many targets as possible at the same time and distribute the reconnaissance tasks in a balanced manner.(2)A reconnaissance task scheduling algorithm based on the cloud-side-end distributed platform is proposed.Based on this platform,a hierarchical collaborative scheduling structure is designed,and the collaborative tasks are divided according to the different hierarchical roles and computing capabilities.The terminal layer is mainly responsible for data collection tasks,the side layer is mainly responsible for track maintenance tasks and local reconnaissance task scheduling and allocation,and the cloud layer is responsible for global reconnaissance task scheduling and allocation.A core reconnaissance task scheduling algorithm based on track priority criterion,cooperative information processing criterion and probe source state transition criterion is proposed.Considering the actual scene,the cooperative data transmission mechanism and cooperative scheduling process suitable for various communication situations are designed.(3)A simulation scheme based on a distributed platform with multiple terminals on both sides of a cloud is designed.The feasibility of the two-level cooperative task scheduling framework is verified by the cooperative scheduling algorithm based on Markov decision process.A multi-source and multi-target distributed collaborative scheduling scenario is also designed,and the simulation verifies the effectiveness of the two-way auction algorithm based on state merging and multi-round allocation in scheduling reconnaissance tasks on a distributed platform.The distributed platform-based reconnaissance task scheduling algorithm proposed in this thesis has been verified by simulation experiments.The simulation results show that:(1)The hierarchical collaborative scheduling framework proposed in this thesis can run successfully on a distributed platform,and is compatible with centralized It can effectively reduce the running time.In the state of limited communication,the running time is less affected by the communication,which ensures the effectiveness of the decision.(2)The bidirectional auction cooperative scheduling algorithm based on state merging and multi-round assignment proposed in this thesis can observe more targets at the same time and the task assignment is more balanced. |