飞行员负荷通过飞行数据优化识别仿真

摘要

对于航空训练中飞行员工作负荷状态的识别,可以有效的确保航空安全.通过飞行数据来对飞行员工作负荷识别,需要对飞行数据进行提取,根据飞行数据特点,建立飞行特征集,完成对负荷状态的识别.传统方法通过建立人体工作负荷状态评估模型对工作状态进行识别,但无法得到飞行特征集,导致识别精度低.提出了基于Treelets降维的飞行员工作负荷状态智能识别方法.首先引入时域信号特征均值、方差与均方根对飞行数据进行提取,根据飞行数据特点,建立飞行特征集,利用飞行特征集组建Treelets降维算法模型;然后根据布谷鸟算法优化的高斯分类模型,评估飞行员工作负荷状态;最后将结果与NASA-TILX的飞行员工作负荷结果对比,利用比对结果完成对航空训练中飞行员工作负荷状态的智能识别.实验结果表明,所提方法能够实现对飞行员的工作状态负荷状态的有效识别,且识别精度较高.%This article puts forward an intelligent recognition method for state of pilot workload based on Tresslets dimensionality reduction.Our research introduced mean value,variance,and mean square root of time-domain signal feature to extract flight data and built feature set of flight according to characteristics of flight data.The feature set was used to build model of algorithm of Treelests dimensionality reduction and state of workload was evaluated according to Gaussian disaggregated model optimized by Cuckoo algorithm.Moreover,the results were compared with workload results of NASA-TLX.Experimental results show that the method can achieve effective recognition and has higher precision.