面向在轨操作的人体负荷能力评估与预测研究

发布时间：2018-01-05 14:31

本文关键词：面向在轨操作的人体负荷能力评估与预测研究　出处：《国防科学技术大学》2014年博士论文　论文类型：学位论文

【摘要】：加强在太空飞行任务过程中人体负荷能力的客观化、定量化研究,能够更好地推进空间资源的探索与利用,以及加速空间技术发展。论文围绕评估与预测两个模型,开展在轨操作人体负荷能力的研究。首先,建立了在轨操作负荷能力评估模型。在轨操作的人体负荷能力评估方法应具有无干扰性、无侵入性、无创伤性、定量性、客观性和准确性等特性。论文首先确定了基于肌电和肌力人体负荷能力评估的模型结构。然后,在多特征理论指导下,提出了利用时域、频域、双频域分析方法对表面肌电信号的非平稳性与非线性属性进行理论解释,求解了肌电信号的幅值、频域和相位信息随负荷能力的变化规律。通过对肌力的统计分析,完成了对肌力物理量性质的数学建模,探寻了人在不同操作工况下肌力不同属性之间的相互关系,比较了人在不同操作工况下操作负荷能力的差异,分析了影响人体空间操作肌力变化的因素。在提取到肌电特征值的基础上,设计了以均方根振幅、积分肌电、平均频率、中值频率、对角切片谱能量与轴向切片谱能量结果为特征指标,并将特征指标与支持向量机结合构建操作动作分类模型,实现了较高精度的辨识率。通过仿真确定了时域、频域及双频域特征值随负荷增加的变化规律。其次,研究了在轨操作人体负荷能力的预测问题。基于Elman神经网络建立了预测模型。选取左、右上肢肱二头肌表面肌电信号的平均频率与中值频率值,以及相应的肌力作为Elman动态神经元网络的输入参数,构建了包含输入层—隐含层—连接层—输出层的四层网络结构。对具有多变量、强耦合、非线性和时变性等特性的在轨操作人体负荷能力的预测问题提出了解决的途径和方法。最后,开展了在轨操作人体负荷能力评估与预测的实验研究。实验过程中测量与采集了人体操作时三维操作力与上肢四个通道的肌电信号。人体负荷能力实验结果验证了评估与预测模型的正确性。论文针对人体空间操作负荷能力的评估与预测问题,研究了利用肌力与肌电信号诊断与测算人体负荷状态的方法,通过该研究初步系统地建立了在轨操作人体负荷能力实时监测、综合评估、有效分类与辨识,以及预先推测等方法。
[Abstract]:To strengthen the objective and quantitative study of the human body's load capacity in the course of space mission, we can better promote the exploration and utilization of space resources. And accelerate the development of space technology. This paper focuses on the evaluation and prediction of the two models to carry out the study of on-orbit human load capacity. First of all. The evaluation model of on-orbit operating load capacity is established. The evaluation method of on-orbit human load capacity should be non-interference, non-invasive, non-traumatic and quantitative. Objectivity and accuracy. Firstly, the model structure based on myoelectric and muscle strength human load assessment is established. Then, under the guidance of multi-feature theory, the use of time domain and frequency domain is proposed. The non-stationary and nonlinear properties of surface EMG signal are theoretically explained by dual-frequency domain analysis method, and the amplitude of EMG signal is solved. Through the statistical analysis of the muscle force, the mathematical model of the physical properties of the muscle force is completed. This paper explores the relationship between different properties of muscle strength under different operating conditions, and compares the difference of operating load capacity of human under different operating conditions. On the basis of extracting the eigenvalue of EMG, we designed the mean square amplitude, integral electromyography, mean frequency and median frequency. The result of diagonal slice energy and axial slice spectrum energy is the characteristic index, and the feature index and support vector machine are combined to construct the operation action classification model to achieve a high precision identification rate. The time domain is determined by simulation. The characteristic values of frequency domain and dual frequency domain change with the increase of load. Secondly, the prediction problem of human load capability in orbit operation is studied. A prediction model based on Elman neural network is established, and the left side is selected. The mean frequency and median frequency of the biceps surface EMG signal of the right upper limb and the corresponding muscle force were used as the input parameters of the Elman dynamic neural network. A four-layer network structure consisting of input layer, hidden layer, connection layer and output layer is constructed. The prediction of human load capacity in orbit operation with nonlinear and time-varying characteristics is discussed in this paper. Finally, the methods are proposed to solve the problem. An experimental study was carried out on the evaluation and prediction of human load capability in orbit operation. EMG signals of three dimensional operating force and four channels of upper limb were measured and collected in the course of the experiment. The experimental results of human body load capacity were verified. The evaluation and prediction model is correct. This paper aims at the evaluation and prediction of human spatial operating load ability. This paper studies the method of diagnosing and measuring the load state of human body by using muscle force and EMG signal. Through this study, the real-time monitoring, comprehensive evaluation, effective classification and identification of human load capability in orbit operation are established. As well as presupposition and so on method.
【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：R85

【相似文献】