基于RT-Thread多线程程序的可穿戴心理健康监测设备设计研究
Research on the Design of Wearable Mental Health Monitoring Device Based on RT-Thread Multi-threaded Programme
陈桃放1
作者信息
- 1. 陕西省建筑职工大学,西安 710068
- 折叠
摘要
为了实现个体心理健康的长期监测,获得客观评估心理状况的实时数据,研究设计了一款采集用户语音、行为以及环境数据的可穿戴设备,并借助主成分分析特征压缩法与集成思想完成心理健康的评估.实验结果表明,合成加速度的时域幅值方差值和第二共振峰特征值分布情况可有效区分心理健康状态.量表低分测试者的方差值分布相对集中,密度曲线在1 000 Hz～3 500 Hz范围内呈现高斯分布,与受试者心理健康状态实际情况相符.研究设计的决策分类模型准确率较高,均方根误差收敛于0.563,召回率较高,模型综合性能较优.此次研究设计的可穿戴心理健康监测设备能够长期有效监测用户行为数据,提供了便捷、客观、综合和个性化的心理健康管理方式.
Abstract
In order to achieve long-term monitoring of individual mental health and obtain real-time data to objectively assess the psychological condition,the study designs a wearable device that collects users'voice,behavioural and environmental data,and com-pletes the assessment of mental health with the help of the Principal Component Analysis feature compression method and the integra-tion idea.The experimental results indicate that the time-domain amplitude variance of the synthesized acceleration and the distribu-tion of the characteristic values of the second resonance peak can effectively distinguish mental health states.The distribution of vari-ance values among low scoring participants in the scale is relatively concentrated,and the density curve shows a Gaussian distribution in the range of 1 000-3 500 Hz,which is consistent with the actual mental health status of the participants.The decision classification model designed for research has a high accuracy,with a root mean square error converging to 0.563,a high recall rate,and excellent overall performance of the model.The wearable mental health monitoring device designed in this study can effectively monitor user be-havioural data in the long term,providing a convenient,objective,comprehensive and personalized way of mental health manage-ment.
关键词
RT-Thread多线程程序/语音特征值/低频数据特征值/主成分分析/皮尔逊相关系数Key words
RT-Thread multithreading programme/speech eigenvalues/low-frequency data eigenvalues/principal component analysis/pearson correlation coefficient引用本文复制引用
基金项目
陕西省教育厅2022年度一般专项科研计划项目(自然科学项目)(22JK0264)
出版年
2024
NETL
NSTL
万方数据