توسعه یک روش مدل نسل انسانی دیجیتال برای طراحی ارگونومیک در محیط مجازی
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|7922||2009||5 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 3187 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : International Journal of Industrial Ergonomics, Volume 39, Issue 5, September 2009, Pages 744–748
A group of digital human models (DHMs) representing the target population under consideration is used to design products and workplaces in virtual environment. The present study proposes a two-step method which generates a group of DHMs in various sizes to properly accommodate the designated level of the human size variability of the target population. In the first step, a designated number of pairs of stature and weight within a specified accommodation range are generated from the bivariate normal distribution of stature and weight of the target population. In the second step, for each pair of stature and weight, the sizes of the DHM body segments are determined using hierarchical regression models and corresponding prediction distributions of individual values. The proposed generation method was applied to the 1988 US Army anthropometric survey data and then implemented to a web-based system for passenger car interior design. This web-based generation system is capable of generating a group of DHMs as nationality, gender, accommodation percentage, and the number of DHMs required is specified. Relevance to industry A digital human simulation system has been used as an effective tool for ergonomic design and evaluation of products and workplaces in virtual environment. The human model generation method proposed in the present study is of use to efficiently generate a group of human models representing the target population.
Digital human model (DHM) simulation systems such as Jack® and RAMSIS® contribute to the efficiency of product design process. These systems have been utilized as an effective design tool to visualize the interaction of a human and workstation system (such as passenger car interior, fighter cockpit, and factory workplace) and to evaluate the human–workstation interaction from ergonomic aspects (such as reach, visibility and comfort). The ergonomic design methodology using digital DHMs makes the iterative process of design evaluation, diagnosis and revision more rapid and economical (Chaffin, 2001). Two human model generation methods (percentile and custom-built methods) are commonly implemented in DHM simulation systems (Dassault Systems, 2005 and UGS, 2006). The percentile method enables the user to generate percentile human models such as 1st, 5th, 50th, 95th, and 99th percentiles for different genders and age groups using anthropometric information stored in the simulation system. On the other hand, the custom-built method enables the user to create tailor-sized human models by specifying a set of anthropometric dimensions—missing values of anthropometric dimensions are estimated by regression equations incorporated in the simulation system. The boundary manikin methods (Bittner, 2000, Eynard et al., 2000, Kim and Whang, 1997, Meindl et al., 1993 and Reed and Flannagan, 2000) and distributed methods (HFES 300, 2004, Jung et al., 2008 and McCulloch et al., 1998) belong to the custom-built method of DHM generation because the body segment sizes of each selected representative case should be manually specified in a DHM simulation system. The existing human model generation methods are limited in terms of representativeness of DHMs for the target population and efficiency of DHM generation. In the percentile method, only a few (mostly 3) DHMs are created and then simulated for ergonomic design and evaluation (Jimmerson, 2001, Nelson, 2001, Thompson, 2001 and You et al., 1997); therefore, the simulation results may not properly reflect the variability of the target population in size and their generalizability can be significantly limited. Next, in the custom-built method, inputting the sizes of the body segments of DHMs is quite time demanding. For example, the values of 21 and 26 body dimensions need to be entered manually to generate a custom-built DHM using RAMSIS and Jack, respectively (Tecmath, 2007 and UGS, 2006). The present study proposed a novel method to generate a group of human models with various sizes which statistically represent the target population under consideration. Then, for efficient human model generation, this study implemented the proposed method into a web-based system, which automatically generates a group of human models as nationality, gender, accommodation range (the range of percentiles accommodated by a particular design) of the target population, and the number of human models required are specified.
نتیجه گیری انگلیسی
The present study developed a generation method of DHMs having various sizes of the body segments and statistically accommodating the specified percentage of the target population. The percentile method, commonly implemented in DHM simulation systems, generates only a few DHMs such as 1st, 5th, 50th, 95th, and 99th percentiles in terms of a selected anthropometric variable (mostly stature). Use of the small group of DHMs which lacks the representativeness of the target population in terms of anthropometric variability can limit the generalizability of corresponding ergonomic design and evaluation results (Reed and Flannagan, 2000). On the other hand, the two-step generation process proposed in the present study forms a better representative group by selecting DHMs within the designated accommodation percentage of the target population in terms of stature and weight and determining the sizes of the body segments using hierarchical regression models and corresponding prediction sampling distributions. By including more representative DHMs in the design and evaluation process, the variability of simulation results can be better predicted for the target population. Stature and weight were selected as primary generators in the study because their information is commonly available and specified for a particular target population; however, other anthropometric variables such as hand length and hand width can be used as primary generators depending on the design context. The proposed human model generation process takes the number of human models (n) as one of the input data. The appropriate value of n to accommodate a designate percentage of the target population can be determined by considering various technical constraints and benefits involved in the design context such as design allowance, size of anthropometric database in use, variances of anthropometric variables under consideration, and relationships between anthropometric variables. The larger the design allowance and the relationships between anthropometric variables, the smaller the required n to form a representative group of the target population; the opposite becomes true for the size of anthropometric database and variances of anthropometric variables. Further research is necessary to develop a sophisticated model which determines an optimal number of human models by compromising the relationship between technical constraints and representatives of a human model group. The proposed human model generation process is based on the multivariate normal distributions of anthropometric variables, hierarchical regression models of anthropometric variables and corresponding prediction sampling distributions. The proposed generation process can be validated by evaluating if these statistical assumptions are satisfied by the anthropometric data in use and a group of generated human models properly represents the designated percentage of the target population. Regarding the validity of the statistical assumptions, stature and weight of the US Army personnel were found to follow a bivariate normal distribution (see Section 2.1), but multivariate normality was not evaluated for the other anthropometric variables—significant normality violations would exist with some anthropometric variables as reported by Vasu and Mital (2000). Next, regarding the validity of the generated human model group, its representativeness for the target population would depend on technical parameters such as the number of human models, design allowance, and relationship between anthropometric variables, as discussed above for the required number of human models. The proposed DHM generation method was implemented into a web-based system for use in designing and evaluating a passenger car interior. The web-based system can automatically generate a group of human models that are statistically accommodating the specified percentage of the target population with simple inputs such as nationality, gender, accommodation range, and the number of models required. Also, the web-based system visualizes each human model using VRML for the ergonomic design and evaluation of a passenger car interior in virtual environment. The DHM generation system developed in the present study is useful to efficiently generate a group of human models for the target population. The custom-built method incorporated in existing DHM simulation systems requires a significant amount of time to manually input the sizes of the body segments for each human model. However, the developed web-based system automatically determines the sizes of body segments of DHMs using the proposed two-step generation method once the target population is specified. Lastly, the web-based DHM generation system uses at present only the anthropometric data of the US Army (female: 2208, male: 1774) so that extension of the anthropometric database is necessary to diverse populations. The DHM generation system can be extended to other populations by adding their anthropometric data. The extension of the anthropometric database will enable the web-based generation system to be applied for design of car interiors for populations with different nationalities.