蜂群无人机自组网多优先级自适应退避算法

针对现有媒质接入控制（MAC）协议退避算法无法为蜂群无人机自组网（FANETs）提供区分服务，且在重负载时性能严重恶化等问题，提出一种多优先级自适应退避算法。采用忙闲因子自适应机制和最优竞争窗自适应机制，根据信道忙闲程度和网络状态参数自适应实时更新各优先级竞争窗口（CW）长度，从而使每次退避的竞争窗口可快速收敛到最佳状态，并实现了多业务区分服务，得到了最优的系统性能。通过建立不同优先级退避过程的三维Markov链模型求解得到了饱和吞吐量下的最优竞争窗自适应因子，并且理论推导了系统吞吐量和平均MAC时延的数学表达式。仿真结果表明，所提算法在重负载时能够实现多优先级区分服务并有效提高系统的吞吐量性能，相比区分业务优先级的自适应退避（PAB）算法和支持QoS的自适应竞争窗口退避算法（Q-ABACW），性能均有较大提升。 

An adaptive backoff algorithm for FANETs based on multiple priority

The existing backoff algorithms of the medium access control (MAC) protocols cannot provide the multiple priority differentiation, and the performance declines sharply under heavy loads in flying Ad hoc networks (FANETs), so a novel adaptive backoff algorithm based on multiple priority differentiation is proposed in this paper. The algorithm adopts a busy/idle factor adaptive mechanism and an optimal contention window (CW) adaptive mechanism, so the length of CW for each priority can be adjusted in real time with the busy degree of channels and network state parameters. Meanwhile, the CW can quickly converge to the best state in every backoff stage, and the multiple priority differentiation can be obtained. Furthermore, the best system throughput performance can be achieved by modeling. The three-dimensional Markov chain model of the backoff process for different priorities is established and the adaptive factor under the saturated throughput is solved by theory. In addition, the mathematical expressions of system throughput and mean MAC delay are also deduced. Simulation results show that the algorithm can achieve the multiple priority differentiation and availably enhance the system throughput, and its performance is superior to the priority adaptive backoff (PAB) algorithm and adaptive CW backoff algorithm for QoS (Q-ABACW).