Journal Search Engine
Search Advanced Search Adode Reader(link)
Download PDF Export Citaion korean bibliography PMC previewer
ISSN : 1598-7248 (Print)
ISSN : 2234-6473 (Online)
Industrial Engineering & Management Systems Vol.18 No.1 pp.25-34
DOI : https://doi.org/10.7232/iems.2019.18.1.025

Visualization of Skilled Worker Motion and its Effect on Product Quality and the Design of Skill Training Systems in Metallic Painting Operations

Masaki Kubo, Takuya Hida, Ryosuke Nakajima, Toshiyuki Matsumoto*
Department of Industrial and Systems Engineering, Aoyama Gakuin University, Tokyo, Japan
Corresponding Author, E-mail: matsumoto@ise.aoyama.ac.jp
August 23, 2017 November 13, 2017 January 16, 2019

ABSTRACT


Many tasks are currently automated in the manufacturing field, but some are still performed by workers, which require technical knowledge and skills. However, it is difficult for a new worker to learn skills that require precise motion and extensive experience. Conventional on-the-job training (OJT) has a problem, which is the difficulty of defining and evaluating a correct task. Therefore, this study aimed to visualize the proper skills and their effects on quality in metallic painting operations, and to design the concept for a skill training system. Product quality depends on a worker’s motion and the effects of tools. To train a skill based on these factors, it is necessary to visualize the proper skill and to devise a method of training. Therefore, changes to a worker’s motion and the quality of products were visualized using a motion capture (MOCAP) system and three-dimensional computer graphics (3DCG) software. Furthermore, such a system requires a trainee to experience the proper motion and to evaluate his or her skills. The motion of two skilled workers was therefore analyzed to extract their skills as explicit knowledge. The skilled workers’ motion was measured using the MOCAP system. The obtained data were then analyzed as positional and rotational data in three axes for each motion of the main parts of the body. Consequently, 11 types of skills were extracted by analyzing the data. Then, an experiment was conducted to visualize the quality of the metallic painting operation using film thickness as an index. Seven factors for film thickness were extracted, and these were quantified in the experiment. Moreover, these factors were applied to 3DCG, and the film thickness was reproduced by simulation. As the result of a statistical test, the reproducibility of the film thickness was guaranteed in 3DCG. Based on the results, a training system was designed for to enable a new worker to learn the skill.



초록


    1. INTRODUCTION

    Many tasks have been automated in the manufacturing field, but there are some difficult tasks that require skillful operation by workers. Such work relies on explicit knowledge so as to avoid variations in product quality caused by worker issues. In particular, small- and medium- sized manufacturing industries in Japan have an issue with technology succession against a sense of impending crisis regarding changes in production engineering systems and an unbalanced labor force with each generation of worker. In metallic painting, it is necessary to manufacture products with an accuracy that is in the order of microns, and acquiring painting skills takes considerable time for the operation considered in this study. For that reason, in recent years, automated painting operations have been implemented for mass production (Asakawa and Takeuchi, 1997;Endregaard, 2002;Baldwin, 1999;Schulz, 2013). However, skilled manual work in single product manufacturing or high-precision products still depends on human senses. Therefore, beginners are taught by skilled workers, who conduct on-the-job training (OJT) to share their experience, intuition, and know-how in the production field. However, OJT leads the trainee to a comprehension of the work action only, so it is not possible to judge whether the entire operation is performed correctly. Consequently, we considered that OJT that trained “tacit knowledge” skills in the production field was not systematized.

    In their studies on work motion, Hida and Seo (2012a, 2012b, 2013) clarified the effects that an object’s shape has on the grasping motion in a pick-and-place task. They also clarified the effects that task conditions (e.g., object size and grasping position) have on muscular load in inspection work. As for skills, Kajihara et al. (2008) applied virtual reality to the learning of skills in assembly work. Concerning tool design, Joshi et al. (2008) suggested modeling in the field of ergonomics. Finally, with respect to training, Hatamura (2006) suggested that motion and knowledge should be taught separately. However, none of these studies clarified the effects that work motion and skills have on product quality.

    For manual metallic painting operations, some studies have focused on ergonomics (Lee et al., 1997;Björing and Hägg, 2000b). They reveal the effects of spray gun type and worker behavior on musculoskeletal disorders. However, the influence that worker behavior has on the painting skill in a metallic painting operation has not been clarified.

    Small and medium sized manufacturing industries have clearly many tasks that rely on manual operations, and there is an issue of skills inheritance associated with such tasks.

    This study thus aimed to extract skilled worker motion quantitatively in metallic painting operations, to visualize its effect on product quality, and to design a concept for a skill training system. This study particularly contributed to extract the quantitative features of the motion of skilled workers, and to extract the work conditions of a spray gun to visualize film thickness for 3DCG simulation.

    2. STUDY SUBJECT

    The current working methods, quality evaluation, and product defects in metallic painting operations were investigated. From these, we clarified the product defects caused by workers’ lack of skills.

    2.1 Target Company

    This study targeted Company A, which performs precision sheet metal pressing and high-quality baked-on finishing.

    2.2 Method of Metallic Painting

    In this study, the target was a metallic painting operation using solvent-based paint. The painting method used a manual spray gun powered by air pressure. Hand painting with a spray gun requires a high level of skills to paint uniformly.

    The complete painting process included cleaning, painting, and baking. After cleaning, the worker suspended the metal sheet on a dedicated hanger. A spray gun was used for the painting. The gun had three knobs to adjust the discharged paint to conform to the object’s shape. These knobs provided adjustment of the discharge pattern, discharge quantity, and air pressure.

    2.3 Present-State Analysis

    2.3.1 Work Method

    The worker should have operated according to a manufacturing procedure and quality control (QC) process chart. However, the worker performed the painting operation using his intuition and experience because there was no pre-determined method for the specific motion required for painting work.

    2.3.2 Quality Evaluation

    Company A evaluates product quality by means of finish appearance, color tone, and film thickness. Finish appearance is evaluated by visual inspection after the baking process. Color tone is evaluated by comparing the finished product to a color chart. Film thickness is evaluated by measuring the four corners and center point on the surface of the product using a film thickness measuring instrument.

    One of the problems with this evaluation method is that quality is determined visually by the inspector. Consequently, there is a possibility of overlooked defects because the evaluation is left to the subjective judgment of the worker. Furthermore, poor uniformity of film thickness might be overlooked because the measurement points comprise only four corners and the center.

    From the above, it can be seen that the painting operation has problems in terms of the variation of motion and subjective judgment of defects. Therefore, it is necessary to analyze the painting motion quantitatively and to investigate the influence that it has on product quality.

    2.3.3 Product Defects

    According to the number of defects reported by customers within a given year, the most common one is the adhesion of foreign matter. The next most common defect is poor uniformity of film thickness. If the film is too thin, there is a lack of coverage. However, if the film is too thick, paint sag occurs. Because these two defects are caused by the worker, we believe that overlooked defects can be reduced by having workers learn the proper skills.

    3. EXTRACTION OF SKILLED WORKER MOTION

    This section specifies a formalized process of skilled worker motion. A change of worker motion affects the spray gun movement, which may affect the painting quality. Therefore, we quantified the posture and operating speed of the worker, as well as the tilt and speed of the spray gun.

    3.1 Extraction of Motion Patterns

    A digital video camera recorded the painting operation of two skilled workers. The results of the recording indicated three motion patterns, as shown in Figure 1. The first is the motion that painted the corner of the product; the second is the motion that painted the product by swinging the spray gun right and left (painting with a lateral motion); and the third is the motion that painted the product by swinging the spray gun up and down (painting with a longitudinal motion). Other motion patterns were not extracted. This study targeted only the horizontal swing condition.

    3.2 Measurement of Motion

    A motion capture (MOCAP) system (OptiTrack, NaturalPoint, Inc.) measured worker motion in a metallic painting operation. In this study, the product shape was a flat plate with dimensions of 450mm×600mm and a thickness of 0.9 mm.

    3.2.1 Target

    The subjects were two skilled workers and one beginner. The paint was a blue-colored solvent-based paint. The number of measurements varied depending on the subject: the skilled workers were measured three times, whereas the beginner was measured twice as a reference.

    3.2.2 Settings

    Here, we describe the method of marker setting and the definition of coordinates for motion capture. Each worker had 34 markers on his body segments and joints to capture his motion. Similarly, the spray gun had three markers. Furthermore, the hanger had four markers to obtain the product position.

    Figure 2 shows the layout and coordinate axes of the workplace. The central coordinate is based on the position where the worker stood.

    3.2.3 Extraction of Motion Data

    The timeline data measured by the MOCAP system is displayed in the graph of Figure 3, which shows the wrist angle of a skilled worker. The blue line represents the change of position, and the red line represents the change of wrist angle. Furthermore, we adapted the timeline data for each joint angle to the motion pattern that was extracted through video analysis. We then conducted the motion analysis using these data.

    3.3 Extraction of Detailed Features

    The overall structure of the painting motion was described in the previous section. In this section, we clarified its detailed features. It was found that the right arm motion of a skilled worker was independent with no relationship to other body sites. Therefore, the degree of freedom in the data was reduced, and the motion of the skilled worker could be analyzed. The analysis target was the joint angle and position of the right hand and the distance between the spray gun and metallic plate. This target was chosen because the motion of the right arm played a major role in all of the motions.

    3.3.1 Analysis of Motion Patterns

    The motion pattern of painting with a lateral motion was analyzed using the joint angles and positions of the skilled worker and the beginner. This motion pattern included periodic motion. Therefore, the following items were analyzed as targets: swinging width of the motion, change in rotation angle, motion speed, and distance between the spray gun and metallic plate. These items were analyzed for each axis to extract the skilled worker and the beginner motions.

    Figures 4(a)-(c) show the motion of the skilled worker with respect to the x-axis, y-axis, and the distance between the spray gun and metallic plate, respectively. Figure 4(a) shows that the skilled worker operated so as not to apply excessive paint product when changing the direction of motion. Figure 4(b) shows that the skilled worker operated the spray gun from bottom to top with a swinging motion. Furthermore, as the change of y-axis position was stepwise, the skilled worker operated so as not to change the work height when swinging right and left. Figure 4(c) shows that the distance between the spray gun and metallic plate was kept constant.

    Figures 5(a)-(c) show the motion of the beginner with respect to the x-axis, y-axis, and the distance between the spray gun and metallic plate, respectively. Figure 5(a) shows that the beginner operated unstable cycle and movements. Figure 5(b) shows that the beginner operated the spray gun from bottom to top with a swinging motion as same as the skilled worker. But, as the change of y-axis position was large stepwise. Figure 5(c) shows that the distance between the spray gun and metallic plate was not kept constant and gradually went near.

    The results clarified that the skilled workers operated with a certain regularity. Furthermore, it can be seen that the skilled workers chose the motion that was easiest to perform and with the least load.

    The paint quality was investigated with film thickness using a measuring instrument (Fisher Instrument Inc.: DIN EN ISO 2360), where 81 points on the plate. The range of film thickness for the product should be average 30 and among 20-40 micrometer. The skilled worker painted with film thickness on average 29.55 and deviation 6.48, and the beginner painted on average 25.28 and deviation 6.43.

    3.3.2 Quantification of Skills

    The following items were analyzed from the skilled worker’s data and the beginner’s data for painting operation, as shown in Table 1: wrist angle of toward painting direction (average ± standard deviation); motion speed toward painting direction (average); motion speed toward the perpendicular direction to painting direction (average); and distance between the spray gun and metallic plate (average). Specifically, we calculated the value of each analysis item for each swinging width period and took an average to extract the features of the workers’ motions. By comparing the skilled worker and the beginner, it was revealed that the skilled worker had a constant wrist angle, the range at one-time painting was narrow, and the painting speed was fast.

    3.4 Extraction of Comprehensive Features

    To understand the overall structure of the workers’ motions, the standard deviation of each joint angle in the timeline data was used to clarify the motion of each body part. Furthermore, the motion was evaluated based on a principal component analysis to extract the rough characteristics of the operation.

    3.4.1 Standard Deviation of Each Joint Angle

    We focused on the motion quantity for each body part from the standard deviation, using the timeline data, of each joint angle when painting with a lateral motion. The analyzed data comprised the angles at 18 body sites on the x-, y-, and z-axes, and one position at the hips on the x-, y-, and z-axes.

    Figure 6 shows the standard deviation of each joint of a skilled worker, who is better in quality (based on product evaluation) when painting with a lateral motion. It can be seen that the standard deviation of the right arm joints is particularly large and that the standard deviations of the trunk and left arm joints are also large. It can be assumed that painting with a lateral motion was mainly carried out using the right arm, and that the motions of the trunk and left arm were related to the motion of the right arm. Moreover, it is clear that the worker used the whole arm because the standard deviations of the angle in the x and y-directions of the right shoulder and in the ydirection of the right hand are large.

    3.4.2 Principal Component Analysis

    The entire structure of the painting motion was clarified by the timeline data of joint angles by applying a principal component analysis to clarify which parts of body used statically of the skilled worker. Using the results of the previous section, the analysis was applied to the body parts that had more than five degrees of standard deviation in their position at each joint. Specifically, the z-rotation of the neck, z- and y-rotations of the head, zrotation of the left wrist, x- and y-rotations of the right shoulder, z- and y-rotations of the right elbow, and z- and y-rotations of the right wrist had more than five degrees of standard deviation. Therefore, we applied a principal component analysis to 11 body parts, which were three body parts of the trunk, including the neck, four body parts of the left arm, including the left wrist, and four body parts of the right arm, including the right shoulder, right elbow, and right wrist.

    Table 2 presents the principal component loadings and the contribution ratio of the main principal components when painting with a lateral motion. The first principal component represents using chest, head and almost all both hands. The second principal component represents using neck, head and partly both hands. The third principal component represents using chest and partly right hand. From the contribution ratios of the main components in each motion, it can be seen that the cumulative contribution ratio up to the third main component exceeds 80 percent. For this reason, the three principal components are indicators that characterize the operation of the skilled worker, and the features of the painting motion can be extracted.

    The principal component loading during painting with a lateral motion was then calculated. From the result, it was considered that the contribution of the right arm motion was the largest with regard to the lateral painting operation because the principal component loadings and contribution rate of each segment of the right arm were larger than those of the other parts.

    4. VISUALIZATION OF QUALITY

    This section describes how we evaluated the painting quality. Therefore, we clarified the effects that the work conditions had on painting quality and simulated the painting quality using 3DCG.

    4.1 Work Condition of Spray Gun

    Table 3 shows the factors related to the discharge of paint applied with the spray gun. The work conditions were extracted from a related study (Nakanishi and Murata, 1987) and from interviews with the workers at the work site.

    4.2 Experimentation for Visualization of Quality

    This section describes the method for quantifying the above-mentioned work conditions and the results.

    4.2.1 Quantification of Air Pressure

    Adjustment of the air pressure for paint application was performed mainly using the discharge rate adjustment knobs that were built into the spray gun and air compressor. The reference values were determined from the QC process chart used by Company A. The air pressure reference values were from 0.1 to 0.3 MPa based on the chart.

    4.2.2 Quantification of Painting Pattern

    An experiment was conducted to quantify the painting pattern. Adjustment was performed using the pattern adjustment knob that was built into the spray gun. In this experiment, the discharge patterns when the adjustment knob was at the minimum and maximum settings were captured from the top and side viewpoints using a video camera. As a result, the maximum and minimum values were 10 degrees in the case of the top view, and 65 and 10 degrees, respectively, in the case of the side view.

    4.2.3 Quantification of Painting Distance

    An experiment was conducted to quantify the maximum paint discharge distance possible with the spray gun. Adjustment of the discharge distance was performed using the air quantity adjustment knob that is built into the spray gun. In this experiment, the paint discharge distances were measured when the adjustment knob was at the minimum and maximum settings. It was found that the maximum discharge distance was not related to the discharge pattern.

    4.2.4 Quantification of Film Thickness Distribution

    An experiment was conducted to quantify the distribution of paint on a metallic plate. It was difficult to obtain the data considering all work conditions because film thickness varies in different work conditions. For this reason, we set arbitrary work conditions and analyzed the change of the film thickness by means of differences in painting time.

    In this experiment, the painting time was 0.50, 0.75, and 1.00 s without moving the spray gun. The film thickness was measured at 1-cm intervals on a painted plate using a film thickness measuring device (Fisher Instruments, Inc.).

    Figure 7 shows the measurement results, which were drawn on a 3D contour line map. The film thickness distribution was analyzed for each row and column by means of these data. In the figure, a peak occurs at the center of the graphs and decreases as the work position moves away from the center. It can also be seen that the film thickness in the lateral direction is reduced rapidly when the work position moves away from the object’s center.

    4.3 Representation of Quality by 3DCG

    This section describes the process that implements the work conditions quantified by 3DCG. Specifically, we used 3DCG to indicate the modeling of tools and equipment and the quality evaluation system for the spray gun. Furthermore, the painting operation using a hand spray gun was simulated.

    4.3.1 Rendering of Tools and Equipment by 3DCG

    We carried out 3DCG modeling of the tools and equipment. The modeling targets were the spray gun, painting field, and metallic plate. Additionally, the adhesion of the paint was simulated. Specifically, this was done by generating particles from the tip of the spray gun and distributing them by blowing air through the gun. We also arranged to stop the motion of the particles once they became attached to the plate. This made it possible to simulate the paint that was discharged from the spray gun and adhered to an object.

    4.3.2 Simulation of Painting by Spray Gun

    We developed a system for outputting the number of adhesion particles in 3DCG by means of a scripting language. The system could measure the number of particles that adhered to each point in a 1cm × 1cm area on the object.

    Painting by spray gun was simulated by means of 3DCG using a work condition and film thickness distribution that were clarified in the previous section. Moreover, reproducibility of the 3DCG was improved by comparing the measured film thickness distribution data from the previous experiment with the simulated adhered particle distribution data.

    Figure 8 shows a 3D contour line map that was drawn from the simulated data during 1.00s of painting time. The results of analyzing the differences between the measured (Figure 8(c)) and simulated data show that two spots have a difference of 14.6 percent, whereas the others have differences of less than 10 percent. Considering that the film thickness measuring device had a measurement error of ±10 percent, the reproducibility of simulated painting work by spray gun using 3DCG was determined to be within the allowable range.

    5. DESIGN OF CONCEPT FOR SKILL TRAINING SYSTEM

    Based on the above eleven extracted types of skills and seven quantified factors for film thickness, as well as Takeshima’s et al. (2014) study, we designed a skill training system that consisted of the following three subsystems.

    The skill evaluation subsystem included two func-tions that could evaluate a trainee’s motion and the quality of a product, as shown in Figure 9. After measuring the motion data using the MOCAP system, a worker’s motion and the quality of products were visualized in 3DCG. Then, the ratio of realizing the proper motion patterns was calculated through an evaluation of the trainee’s motion system. Also, the virtual film thickness was calculated using the motion and spray gun data from an evaluation of the quality system.

    The skill diagnosis subsystem included two func-tions that could guide a trainee, as shown in Figure 10. By following the normal motion and streaming data, an indication of time was provided to the trainee. Also, using the motion data and evaluation system for motion, advice for the proper motion was given to the trainee.

    Implementation of the training subsystem included two functions that could provide the training procedure considering the difficulty and influencing the quality. This subsystem could train skills using the two previously mentioned subsystems, as shown in Figure 11. This system had measurement and indication by MOCAP and a head-mounted display in the training facility, with a model spray gun in 3DCG.

    Based on the above design concept, we are developing the system using augmented reality (AR) technology.

    6. DISCUSSION

    This study was implemented to visualize skilled workers’ motion and its effect on product quality in metallic painting operations in a small-sized manufacturing facility.

    In this study, the target was a metallic painting operation using solvent-based paint. The painting method used a manual spray gun powered by air pressure. There were many methods and materials used in the operation. Also, only two skilled workers were analyzed to extract the characteristics of the skill.

    Ikemoto et al. (2016, 2017) conducted an extraction of skilled workers’ motion in automobile repair painting work and developed an e-learning system for supporting skill acquisition. Regarding the extraction of skilled workers’ motion, the current study quantitatively extracted more motion than Ikemoto’s study. Therefore, we believe that it will be useful for acquiring the skilled workers’ motion. In addition, we also quantified the effect of the painting motion on product quality and designed a skill training system. Because the painting operation uses sensorimotor system skills, not only is skill acquisition supported by e-learning, as is the case in Ikemoto’s study, but the training system proposed in this study is thought to be more efficient in acquiring experienced workers’ skills. However, we believe that the procedure to visualize a skilled worker’s motion and its effect on product quality are useful, although we still need to expand the results for generalization. Furthermore, in addition to proposing the concept of a skill training system, we should develop and implement it effectively.

    7. CONCLUSION

    Many tasks are currently automated in the manufacturing field, but some are still performed by workers, which require technical knowledge and skills. However, it is difficult for a new worker to learn skills that require precise motion and extensive experience. Conventional OJT has a problem, which is the difficulty of defining and evaluating the correct task. Product quality depends on a worker’s motion and the effects of tools. To train skills based on these, it is necessary to visualize the proper skills and devise a method of training.

    Therefore, this study was performed to visualize a skilled worker’s motion and its effect on product quality in a metallic painting operation to design the concept for a skill training system. The results of this study are summarized as follows:

    7.1 Extraction of Skilled Worker Motion

    A method of analyzing skills using a MOCAP sys-tem was devised to extract the skills of an experienced worker in a painting operation. Practically speaking, we measured the motion of skilled workers using this system. Furthermore, the comprehensive features of that motion were extracted using a principal component analysis, and the skills of the workers were quantified by analyzing their features in detail.

    7.2 Visualization of Quality

    We extracted the work conditions of a spray gun to visualize film thickness, which is an index for evaluating product quality in painting work, and we quantified the work conditions. We also simulated them using 3DCG and were able to understand the incremental changes in product quality.

    7.3 Design of Concept for a Skill Training System

    We designed the concept for a skill training system that consisted of a skill evaluation subsystem, skill diag-nosis subsystem, and training implementation subsystem. Based on the design concept, we are developing the system using AR technology.

    A future issue is to develop and implement a system to train painting workers using the quantified skills and skill training system designed in this study.

    ACKNOWLEDGMENT

    This work was supported by JSPS KAKENHI Grant Number 15K21353 and the well-being project in Aoyama Gakuin University.

    Figure

    IEMS-18-1-25_F1.gif

    Motion pattern.

    IEMS-18-1-25_F2.gif

    Defined coordinate axes.

    IEMS-18-1-25_F3.gif

    Wrist angle of a skilled worker.

    IEMS-18-1-25_F4.gif

    Motion patterns of a skilled worker.

    IEMS-18-1-25_F5.gif

    Motion patterns of a beginner.

    IEMS-18-1-25_F6.gif

    Standard deviation of position and each joint in the painting with a lateral motion of a skilled worker.

    IEMS-18-1-25_F7.gif

    Measured film thickness distribution.

    IEMS-18-1-25_F8.gif

    Simulated data (Painting time: 1.00sec).

    IEMS-18-1-25_F9.gif

    Design of evaluation skill sub-system.

    IEMS-18-1-25_F10.gif

    Design of diagnosis skill sub-system.

    IEMS-18-1-25_F11.gif

    Design of implementation train sub-system.

    Table

    Characteristics of the skilled worker and the beginner motion

    Principal component loadings and contribution rate of the main principal component in the painting with lateral motion

    Work condition

    REFERENCES

    1. Asakawa, N. and Takeuchi, Y. (1997), Automation of spray-painting by an industrial robot: For the case of sculptured surface, Transactions of the Japan Society of Mechanical Engineers, Series C, 63(610), 347-352 (in Japanese).
    2. Baldwin, S. (1999), Robotic paint automation: The pros and cons of using robots in your paint finishing system, Metal Finishing, 97(5), 179-182.
    3. Björing, G. and Hägg, G. M. (2000a), Musculoskeletal exposure of manual spray painting in the woodworking industry: An ergonomic study on painters, International Journal of Industrial Ergonomics, 26(6), 603-614.
    4. Björing, G. and Hägg, G. M. (2000b), The Ergonomics of spray guns: Users’ opinions and technical measurements on spray guns compared with previous recommendations for hand tools, International Journal of Industrial Ergonomics, 25(4), 405-414.
    5. Endregaard, E. A. (2002), Paint robotics: Improving automotive painting performance, Metal Finishing, 100(5), 8-10, 12-13.
    6. Hatamura, Y. (2006), How to Transfer of Technology, Kodansha Ltd, Tokyo Japan (in Japanese).
    7. Hida, T. and Seo, A. (2012a), Effect of task conditions on grasp force during pick-and-place task, Proceedings of the 11th International Conference on Industrial Management, Tokyo, Japan, 55-59.
    8. Hida, T. and Seo, A. (2012b), Development of evaluation function for estimating upper-limb load during visual inspection, Proceedings of the 13th Asia Pacific Industrial Engineering and Management Systems Conference, Phuket, Thailand, 1773-1778.
    9. Hida, T. and Seo, A. (2013), Effect of grasp conditions on upper limb load during visual inspection of objects in one hand, Proceedings of the Institute of Industrial Engineers Asian Conference, Taipei, Taiwan, 229-237.
    10. Ikemoto, S. , Hamada, H. , and Takai, Y.(2016), Expert’s common factor of painting motion in auto repair painting process, Proceedings of the Advances in Ergonomics of Manufacturing: Managing the Enterprise of the Future, Springer, Cham, 23-33.
    11. Ikemoto, S. , Hamada, H. Takai, Y., and Kuwahara, N.(2017), Creation and usability evaluating of e-learning contents for automobile repair block painting, International Journal of Advanced Computer Science and Applications, 8(12), 333-337.
    12. Joshi, A. , Guttenberg, R. , Leu, M., and Murray, S.(2008), Modeling of the hand-arm system for impact loading in shear fastener installation, International Journal of Industrial Ergonomics, 38(9-10), 715-725.
    13. Kajihara, Y. , Tsunesada, K. , Nishimura, T., Osaki, H., and Ikeda, T.(2008), Analysis of effects of virtual reality based assembly training, Journal of Japan Industrial Management Association, 59(2), 162-172 (in Japanese).
    14. Lee, C. C. , Nelson, J. E. , Davis, K.G., and Marras, W. S.(1997), An ergonomic comparison of industrial spray paint guns, International Journal of Industrial Ergonomics, 19(6), 425-435.
    15. Nakanishi, T. and Murata, A. (1987), Japanese lacquer spray paint technology to the large product, Reports of the Gifu Prefectural Industrial Arts Research Institute, 4, 32-36 (in Japanese).
    16. Schulz, D. (2013), Painting trends in the automotive industry, Metal Finishing, 111(5), 38-40.
    17. Takeshima, M. , Yasuda, S. , Shinoda, S., Matsumoto, T., and Shinoda, M.(2014), Design of a method to analyze skill from the viewpoint of transformation from worker’s motion to objects change, Journal of Japan Industrial Management Association, 65(1), 1-8 (in Japanese).