存在时延及丢包的网络化飞行器鲁棒故障检测

考虑同时存在未知时延及随机丢包的飞行器网络环境，研究了飞行器网络控制系统的鲁棒故障检测问题。将未知时延及随机丢包的影响转变为未知扰动输入，建立飞行器网络控制系统的统一数学模型，同时考虑残差对故障的灵敏性和对未知扰动的鲁棒性，利用模型参考的思想，将故障检测问题分解为最优残差模型设计及鲁棒跟踪问题两部分，利用多目标优化方法给出最优残差模型的求解条件，并以线性矩阵不等式的方式给出鲁棒跟踪问题的存在条件，数值仿真表明，所提方法能有效抑制时延及丢包的影响，并能快速有效地检测出更小故障，验证了方法的有效性和优越性。

Robust Fault Detection for Networked Flight Vehicle with Delay and Packet Dropout

The problem of robust fault detection is investigated for networked flight control systems with unknown delays and random packet dropouts. The influences of the unknown delays and random packet dropouts are transformed to the unknown disturbance input, and a uniform mathematics model of the networked flight control systems is established. Considering the sensitivity of the residual to fault and the robustness against unknown disturbances in the same time, the fault detection problem is divided into two parts: design of the optimal reference residual model and the solution of the robust tracking. Then the solvable condition of the optimal reference residual model is obtained by using the multi objective optimization, and further the solution of the robust tracking is derived by using linear matrix inequality techniques. The numerical example show that the proposed approach can restrain effectively the impact of the delays and packet dropouts, and small faults can be detected quickly and effectively.