乘员人体骨骼系统参数化建模研究

发布时间：2018-05-07 10:14

本文选题：骨骼系统 + 体型分析　；参考：《吉林大学》2014年硕士论文

【摘要】：随着汽车行业的快速发展,人机工效分析得到了大量学者的关注,而大量准确的人体生物力学模型是进行人机分析的前提。只有基于大量的人体测量数据,建立的合理的人体生物力学模型才能应用到车辆的设计与分析中,例如H点范围的设计与校核、视野分析和手伸及界面分析等。同时也为驾驶员的操作舒适性和乘员的乘坐方便性分析提供基础数据。因此建立参数化的人体生物力学模型,通过参数调节产生大量的人体生物力学模型是汽车设计领域的热点问题。人体骨骼作为人体体表和肌肉模型的驱动机构,建立可以缩放的人体骨骼模型是人体生物力学建模的基础,是汽车设计和仿真所必需的,也是人机工程设计与分析中必不可少的基础模型。因此本文的研究目标是建立准确的人体骨骼系统模型,并研究多种骨骼缩放技术,为以后进行驾驶姿势分析、视域分析、舒适性评价等提供数字化人体模型。获得大量的群体数据是建立参数化的人体骨骼模型的基础。现在通常采用的人体测量学方法操作繁琐、费时费力,因此本文采取蒙特卡罗仿真的方法。基于群体人体尺寸数据的数字特征,仿真产生建模所需的人体尺寸数据样本,产生的各项人体尺寸数据均服从正态分布,并对多维人体尺度的联合分布特点进行分析,便于后面建立人体尺度预测模型。获取人体测量数据样本后,进行体型分类可以有效地区分个体的体型差异,提高人体尺度预测模型的精度。本文选取了五种不同的体型分类方法,并对各种分类方法的物理意义进行说明。选取能涵盖尽可能多的人体体型特征,并保证各组内体型差别不大,组间差别较大的一种体型分类方法。体型分类完成后,采用逐步线性回归的方法,选取身高、体重和坐高作为主预测因子建立了两层的人体尺度预测模型,并对此模型进行了精度验证。通过分析不同百分位的人体尺寸预测值与样本值的误差,以及七个典型人体模型的20项宏观人体尺寸的预测值与样本值的误差,对建立的预测模型进行进一步的验证。参数化的人体骨骼系统模型作为驱动骨骼系统模型的虚拟结构,是建立缩放的人体骨骼系统模型的基础。因此本文首先基于人体尺度预测模型,建立了17段的人体骨架模型,可以驱动详细的人体骨骼系统完成姿势和尺寸的调节。然后基于人体切片建立了详细的骨骼模型,并参考美国Reed教授提出的人体骨骼装配原理,完成了人体骨架模型和骨骼的装配,最终建立了标准的人体骨骼模型。利用建立的标准人体骨骼模型,本文研究了规则骨骼和不规则骨骼的缩放方法。对于规则骨骼的缩放,本文主要探讨基于骨骼宏观尺寸的比例缩放方法；对于不规则骨骼,本文主要研究基于径向基函数的骨骼缩放方法,保留复杂骨骼的细部特征,并研究了影响不规则骨骼缩放精度的因素,并结合典型个体分析,说明该方法能够快速准确地实现骨骼的非线性缩放。本文通过选取大容量的人体尺寸样板数据进行体型分类,建立了人体尺度预测模型,并基于人体切片数据完成了人体骨骼系统建模。这种研究方法可以推广应用到人体体表和骨骼的生物力学建模过程中。同时,研究了骨骼的缩放技术,得到了合理的骨骼缩放方法,能够快速地实现骨骼的准确缩放。综合上述研究成果,参数化的人体骨骼系统模型,能够准确合理地实现不同骨骼的快速缩放,为进行人机工效分析提供了基础模型。
[Abstract]:With the rapid development of the automobile industry, a large number of scholars pay attention to the ergonomics analysis, and a large number of accurate human biomechanical models are the precondition of man-machine analysis. Only a reasonable human body biomechanical model based on a large number of anthropometric data can be used in the design and analysis of vehicles, such as the H point range. Design and check, visual field analysis, hand extension and interface analysis, etc., and provide basic data for the driver's operational comfort and passengers' ride convenience analysis. Therefore, the establishment of a parameterized human biomechanical model and a large number of human biomechanical models through parameter adjustment are the hot issues in the field of automobile design. As the driving mechanism of body surface and muscle model, the establishment of a scalable human skeleton model is the basis of human biomechanical modeling. It is necessary for the design and Simulation of automobile. It is also an essential foundation model in the design and analysis of ergonomics. Therefore, the objective of this paper is to establish an accurate human skeleton system model. We also studied a variety of skeletal scaling techniques to provide a digital human model for future driving posture analysis, horizon analysis and comfort evaluation.
Obtaining a large number of group data is the basis of establishing a parameterized human skeleton model. Now the usual anthropometric method is complicated and time-consuming. Therefore, Monte Carlo simulation is adopted in this paper. Based on the digital features of the population size data, the simulation of the human body size data is produced by simulation. All human size data are subject to normal distribution, and the joint distribution characteristics of multi-dimensional human scale are analyzed, so as to facilitate the establishment of human body scale prediction model.
After obtaining the sample of human body measurement data, the body classification can effectively divide the body shape difference and improve the accuracy of the human body scale prediction model. In this paper, five different types of body classification methods are selected, and the physical significance of various classification methods are explained. A two layer model of human body size prediction was established by stepwise linear regression, and the accuracy of the model was verified. By analyzing the size of different percentile body dimensions, the body size of different percentile was analyzed. The error between the predicted value and the sample value, as well as the error between the predicted value and the sample value of the 20 macroscopic human body sizes of seven typical human body models, is further verified.
As the virtual structure of the skeleton system model, the parameterized human skeleton system model is the basis for establishing the zoomed human skeleton system model. This paper, firstly, based on the human body scale prediction model, established 17 segments of the human skeleton model, which can drive the detailed body bone system to complete the adjustment of the posture and size. The detailed skeleton model was established in the human slice, and the skeleton assembly of the human skeleton was completed with reference to the principle of the human skeleton assembly proposed by Professor Reed in the United States. Finally, the standard human skeleton model was established.
Using the established standard human skeleton model, the scaling method of regular skeleton and irregular skeleton is studied in this paper. For the scaling of regular bones, this paper mainly discusses the scaling method based on the macroscopic size of bone. For irregular bones, this paper mainly studies the skeleton scaling method based on radial basis function to preserve the complex skeleton. The factors that affect the scaling accuracy of irregular bones are studied and combined with typical individual analysis, it is shown that this method can quickly and accurately realize the nonlinear scaling of bone.
In this paper, the human body size prediction model is established by selecting large capacity body size sample data for body size, and the modeling of human skeleton system is completed based on human slice data. This research method can be applied to the biomechanical modeling of body surface and bone. At the same time, the scaling technique of bone is studied. A reasonable skeleton scaling method is obtained, which can quickly scale the bone. The parameterized human skeleton system model can accurately and reasonably realize the rapid zooming of different bones, and provides a basic model for the analysis of ergonomics.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R336

【参考文献】