The Influence Of Stress On Performance Under Digital Interactive Interface And Its Speech Recognition

The stress state of the worker is one of the factors affecting the reliability of behavior.With the development of technology,digital human-machine interface has gradually replaced the traditional control system.But there are few related studies on the influence of stress on digital human-computer interaction behavior,mainly focusing on the fields of physiology and cognitive neuroscience.On the other side,in the field of digital human-computer interaction,there are few studies on stress monitoring methods which rely on subjective reports and objective physiological measurements.With the development of voice signal processing technology,non-contact voice monitoring has become a research hotspot.This study aims to explore the impact of stress on digital human-computer interaction behavior,and verify the validity of speech stress monitoring methods,so as to provide theoretical reference for practical application in real digital interaction scenarios.This study consists of three parts.Study 1 based on previous research on digital interfaces,two single-layer digital interface tasks were constructed to explore the impact of stress under the single-layer digital interface task.During the experiment,the reaction time,number of mistakes and other stress monitoring data were collected as evidence of the effects of stress.Study 2 explores the impact of stress on the multi-level digital interface tasks by constructing multi-level digital interface.Ask the subjects to complete a more complex problem-solving task under a shorter time limit to induce the stress state of the subjects.During the experiment,the path indicators,rating of disorientation and other stress monitoring data were collected as evidence of the effects of stress.Study 3 combined the MIST paradigm and the real nuclear power interface to design a nuclear power simulation task.After subjects completing the task,we collected the speech feature data and output Mel-frequency cepstral coefficients(MFCCs)as the main speech feature vectors.Neural network models were used to predicate the stress level.Study 3 verified that the non-contact stress monitoring index can accurately predict the stress state of digital interface operator.The main conclusion of this study are as follows:(1)In the single-layer digital interface task,participants in the stressed and unstressed conditions showed different error patterns.(2)In the multi-layer digital interface task,the subjects under the stress state produced a higher sense of disorientation and more lost disorientation behaviors.But there was no significant difference in the performance score of the two groups of subjects.(3)Based on speech feature vector data,the stress state at the current moment can be accurately predicted.Through the training of Back Propagation(BP)neural network model,the prediction accuracy can reach 95%.