基于神经网络的无人机云服务质量控制方法研究

无人机云通过动态卸载任务到云端进行高效处理，能够极大地提高无人机的智能水平。由于设计理念、任务环境、突发事件等因素，导致卸载的任务对网络服务质量（QoS）需求不尽相同。从控制角度研究无人机云系统的网络传输QoS控制问题，提出并实现了一种基于BP神经网络的双闭环接入控制方法，在最大化能量利用率的同时，实现绝对QoS和相对QoS保证。硬件实验结果表明，该方法不仅能够在任务动态变化时提供相对QoS和绝对QoS保证，并且在网络高负载下具有更高的吞吐量和能量利用率，在网络低负载下具有更低的能耗。

A Network Control-based QoS-enhanced MAC for UAV Cloud

Unmanned aerial vehicle (UAV) cloud can greatly enhance the intelligence of unmanned system by dynamically uploading the compute-intensive applications to the cloud. The different UAV missions may have different quality of service (QoS) requirements due to the uncertainty of UAV missions and the fast-changing battlefield environment. A BP neuron network-based feedback differentiated control approach for QoS-aware (BPFD)-MAC in UAV cloud is proposed, which can support both absolute and relative QoS guarantees with the consideration of energy saving. The hardware experiments demonstrate the feasibility of BPFD-MAC. Under heavy loads, BPFD has better throughput and power use efficiency; and under light load, BPFD has lower total energy consumption.