航天医学与医学工程

超重环境可对人体生理、情绪及认知功能产生诸多影响，载人登月和深空探测任务的超重环境更加复杂，对航天员的心理素质和操作能力提出了更高的要求。为了更加客观和精准地评价人在超重环境下的效能，利用人工神经网络误差反向传播算法，构建超重环境下基于人的生理、情绪及认知能力的多维效能评估模型。经过验证和比较，最终选定4层人工神经网络，2个隐含层，29个节点输入，1节点输出，本试验样本的模型精度可达93.55%，数据增强后精度可达98.3%。该模型可以实现对超重环境下受试者效能的自动化测评，减少对专家主观打分的需求，减弱对主试主观判断的依赖，更加全面精准定量评估人在超重环境下的能力，为后续利用人工智能技术建立超重主试/主教员的辅助决策系统提供了技术支撑。后续可根据在超重环境下构建的基于人的生理、情绪及认知能力的效能评估模型，利用人工智能技术建立超重主试/主教员的辅助决策系统，为更加高效精准的航天员/飞行员超重选拔、训练或鉴定提供支持。

Overweight environments have many impacts on human physiology, emotions, and cognitive functions. The overweight environment of manned lunar landing and deep space exploration missions is more complex, which puts higher demands on the psychological quality and operational ability of astronauts. In order to objectively and accurately evaluate human efficacy in overweight environments, an artificial neural network Back Propagation(BP algorithm)was used to construct an efficacy evaluation model based on human physiology, emotions, and cognitive abilities in overweight environments. After verification and comparison, a four-layer artificial neural network with 2 hidden layers, 29 input nodes, and 1 output node was ultimately selected. The model accuracy of the experimental sample reached 93.55%, and the accuracy after data augmentation reached 98.3%. This model can achieve automated evaluation of subjects in overweight environments, reduce the need for subjective evaluation by experts, weaken the dependence on subjective judgments of main subjects, and provide more comprehensive and accurate quantitative evaluation of human abilities in overweight environments. It provides technical support for the subsequent use of artificial intelligence technology to establish an auxiliary decision-making system for overweight main subjects/main officials. Subsequently, an efficacy evaluation model based on human physiology, emotions, and cognitive abilities in overweight environments can be constructed, and an auxiliary decision-making system for overweight experiments/bishops can be established using artificial intelligence technology. Provide support for more efficient and accurate selection, training, or identification of overweight astronauts and pilots.