恶劣气象条件下的飞行员视觉受到强烈刺激，导致其生理、心理出现异常模式，甚至出现视觉失能或脑疲劳。本项目重点聚焦恶劣飞行环境下飞行员视觉刺激-脑疲劳认知联动响应耦合关系机理研究，为飞行员工作负荷评价提供支撑。建立一种三维空间内的正交、对称类Haar球面小波对飞行员虹膜进行多尺度分解，获取其瞳孔形状几何信息。设计一种基于Kaiser窗的光滑仿射伪Wigner-Ville分布时频分析方法，计算飞行员动态脑节律瞬时幅值的功率谱密度，用功率谱密度曲线面积作为四个疲劳特征向量的输入，形成一种新的脑疲劳认知评价指标。建立一种层次性狄利克雷过程-隐半马尔可夫模型，动态识别飞行员脑疲劳持续时间和认知属性。提取“光-声-振动”协同刺激下的飞行员视觉刺激-脑疲劳特征，建立基于高阶收缩深度自编码网络-三角模糊高斯支持向量机的飞行员疲劳认知状态感知模型。通过多组实验数据验证所设计的模型与算法的有效性。

The pilot's vision in severe weather conditions is strongly stimulated. This will lead to abnormal patterns in their physiology and psychology, and even visual disability or brain fatigue. This project focuses on the research on the responding and coupling mechanism between pilot visual stimulation and brain fatigue cognitive in harsh flight environment, and provides support for pilot workload evaluation. An orthogonal and symmetric Haar spherical wavelet in a three-dimensional space is established to decompose the iris of the pilot in multiple scales to obtain the geometric information of the pupil shape. .A smooth affine pseudo Wigner-Ville distributed time-frequency analysis method with Kaiser window is designed to calculate the power spectral density of instantaneous amplitude from the pilot's dynamic brain rhythm. The curve area of power spectral density is used as the input of four fatigue eigenvectors to form a new cognitive assessment index for brain fatigue. A hierarchical Dirichlet process and Hidden semi-Markov model is established to dynamically identify the brain fatigue duration and cognitive attributes of pilots. This project extracts the characteristics of visual and brain fatigue under the combined stimulation of light, sound and vibration. This study establishes a perception model of pilot’s fatigue cognitive state based on high-order contractive deep auto-encoder network and triangular fuzzy Gaussian support vector machine. The effectiveness of the designed models and algorithms is verified by multiple sets of experimental data.