Research On Evaluation Method Of Soldier’s State Based On Physiological Parameters

With the current vigorous development of military equipment and technologies around the world,it is a challenge for soldiers to work with the soldier system,which can cause some damages on soldiers.Therefore,the prediction of the soldiers’ energy expenditure and the assessment of psychological state have a broad space for development.However,the traditional methods of energy expenditure prediction and psychological state assessment have high cost and limited application range,and they cannot provide enough accuracy.Therefore,it is of great practical significance to find a low-cost,widely used,accurate and effective evaluation method to measure the energy expenditure and psychological state of soldiers in real time.At the same time,it can be a guideline in the training of soldiers,which can make outstanding contributions to the soldiers’ health.In this paper,15 male army soldiers were selected to obtain their heart rate,respiration rate,acceleration,oxygen consumption and ECG signals.And the energy consumption prediction model and mental state assessment model of 15 soldiers were established respectively.Through the simulation analysis,the model error is far less than the traditional methods,which verifies the feasibility and effectiveness of the model,and proposes a new method for the subsequent evaluation of the movement state of other soldiers.The main contents of this paper are as follows:Firstly,several commonly used extreme learning machines: Incremental Extreme Learning Machine(I-ELM)and Orthogonal Incremental Extreme Learning Machine(OI-ELM)were analyzed.Then an enhanced Orthogonal Incremental Extreme Learning Machine(EOI-ELM)was proposed to solve the problem that the original algorithm can not reduce invalid nodes.Compared with OI-ELM algorithm,the EOI-ELM algorithm has higher accuracy,stability and generalization ability through the simulation experiments.Secondly,based on the EOI-ELM algorithm,a model for predicting soldier’s energy expenditure using physiological parameters was proposed.Through the design of orthogonal acquisition experiment of physiology information of soldiers,heart rate,respiration rate,skin temperature and acceleration of soldiers were collected in the exercise.These parameters were taken as input,while the oxygen consumption as output,and the soldier’s energy expenditure was predicted.The comparison of EOI-ELM energy expenditure methods with traditional methods verifies the effectiveness and accuracy,and also shows that EOI-ELM can predict soldiers’ energy expenditure.Then,the noise in the ECG signal was removed by wavelet threshold,and the heart rate variability data of soldiers was obtained after the R wave was extracted.Then the characteristics of HRV time domain analysis,frequency domain analysis and nonlinear analysis were selected as the input,while the soldiers’ state as the output.And the classification function of EOI-ELM algorithm was used to establish the model of psychological state assessment.By comparing with the single analytic method,the advantage of this method over single analytic method is verified.At last,a low-load human flexibility test system and a soldier state evaluation software were designed.Through these two systems,the training in low-load condition can be executed,and the energy expenditure and psychological state of soldiers can be measured in real time in this process,which can realize the scientific guidance to soldiers’ training.