基于BP网络的机组工作量评估方法研究

为研究机组工作量的评估问题,搭建了模拟飞行实验平台,设定了7种不同负荷的飞行任务。模拟飞行中,对被试的生理指标进行测量,并采集其主观评价值。采用BP网络进行数学建模,将生理指标测量值作为输入、NASA-TLX表法的评价值作为输出,利用输入、输出模式对对网络进行训练,并对该方法进行验证。结果表明,基于BP网络的机组工作量评估方法,较传统方法更为稳定、精确,且大大降低了被试数量。提高训练模式对的数量、简化生理指标的测量过程,并保证测量精度是该方法成功的关键。

Research on Crew Workload Base on BP Neural Network

Build a flight simulation experiment platform and set seven dif erent load missions, during the experiment, record participators＇ physical signs（heart rate, EEG, blink rate） and the corresponding subjective quality scores taking advantage of NASA-TLX. Using BP neural net builts a mathematical model as a mapping, making physical signs as inputs and subjective quality scores as outputs. Using the other three cases test the mapping model and compare their results with NASA-TLX＇s. The result shows that the new method is more steady and precise than the older and reduces the quantity of the participators. How to simplify the test of the physical signs and keep the precision is very important for this method.