Design And Implementation Of Multi-uav Collaboration Platform Based On Dynamic Networking And Computing Offloading

With the continuous advancement and maturity of related technologies of the Internet of Things,applications based on UAVs are being widely concerned by various industries.However,due to the limited load and endurance of single UAV,complex tasks such as combination of video and processing cannot be performed alone.Therefore,multi-UAVs equipped heterogeneous devices work together to complete tasks has become a new application trend.However,there are two problems that need to be solved to efficiently perform multi-UAV collaboration.Firstly,constructing network communication not only among high-speed UAVs,but also between ground station and UAVs,which can transfer collaboration information among UAVs efficiently and control UAVs from remote ground station,is a good challenge.Secondly,for the heterogeneity of UAV-equipped equipment,realizing the offloading and collaborative calculation between different UAVs and then achieving the complex task goals with limited resources is also challenging.In order to solve the above challenges,this thesis has carried out related research on two key technologies,including multi-UAV network construction and multi-UAV collaborative computing.The research provides a network foundation for multi-UAV collaboration and breaks through the processing capability constraints of single UAV on computing requirements of complex task.Specifically,this thesis conducts research work from the following three aspects:(1)According to the different requirements of the network environment for the external control and internal collaboration,this thesis divides the multi-UAV network into two parts,internal communication and external communication.The ad hoc network and OLSR technology are used to realize the non-centralized communication within the UAV swarm,which breaks through the control of the number of UAV swarm,and increase the bandwidth of internal communication.The 4G LTE and other technologies are used to realize the external communication of the UAV swarm,which breaks through the constraints of the flight distance of the UAV swarm.(2)Multi-UAV collaborative computing is performed between UAVs that equipped with different devices.UAVs equipped with video equipment generate image analysis tasks continuously and UAVs equipped with computing devices perform tasks and feedback results through the internal network.This thesis designs a pipeline-based task offloading mechanism,proposes a centralized resource allocation mechanism,and proposes a computing resource allocation strategy that maximizes the utilization of computing resources of the UAV swarm.(3)This thesis designs and develops a multi-UAV collaborative computing platform based on the above research results.The platform includes flight control functions of ground station and collaborative computing functions between on-board devices.The experimental results show that the multi-UAV network architecture proposed in this thesis provides efficient internal communication of UAV swarm.The over-the-horizon control of large-scale UAV swarm can also be realized.In the meanwhile,collaborative computing between equipment carried by UAVs can significantly improve the overall computing capability of the UAV swarm,which enables UAVs to perform more complex tasks.The research work done in this thesis can provide some theoretical support and technical reference for the ultralong-range control of large-scale UAV swarm.At the same time,it has positive reference value for exploring the cooperation mode between UAVs.