Augmented Reality on Tablets in Support of MRO Performance

Transcription

1 Augmented Reality on Tablets in Support of MRO Performance Andrew Woo, Billy Yuen, Tim Hayes Carl Byers Eugene Fiume NGRAIN (Canada) Corporation Logres Inc. University of Toronto Vancouver, BC, Canada Ottawa, Ontario, Canada Toronto, Ontario, Canada [awoo byuen ABSTRACT Tablet computers have recently become ubiquitous, and 3D simulation technologies are now starting to be available on them. This presents new opportunities for the deployment of Augmented Reality (AR) solutions in support of Maintenance, Repair and Overhaul (MRO) performance. In this paper, we elaborate on the choice of the tablet platform for AR, and list the capabilities that are possible using the combination of tablet and AR to enhance MRO performance. A significant advantage of these capabilities on the tablet is that only a single data representation is necessary, for use in AR as well as virtual environments, enabling the technician to choose, on the fly, which environment is best suited for his or her needs. ABOUT THE AUTHORS Andrew Woo is chief technology officer at NGRAIN, a leading developer of interactive 3D visualization and simulation software tools. Prior to NGRAIN, Andrew was senior R&D manager at Alias, having worked on Sketch!, PowerAnimator, Studio, and Maya. He received a B.S. and M.S. in Computer Science at the University of Toronto. Andrew is a member of IEEE, CIPS, CHCCS, and is an ACM distinguished member. Billy Yuen is Solutions Architect at NGRAIN. He managed multiple groups including software application team and professional service team. Before NGRAIN, Billy was a software engineer in Microsoft, HSBC and team lead in Wallstreet System. Projects and software include PowerPoint 2000 and 2003, HSBC online banking, HSBC call center internal portal and Wallstreet web based cash management system. Billy developed iphone, ipad & Android tablet software since Billy received B.S. in computer science in University of British Columbia in He holds designations in CSM, Prince2 Practitioner and SCPM. Tim Hayes is currently investigating applications of AR in realm of maintenance tasks and training as a software developer at NGRAIN. Prior to this, Tim completed an M. E. Sc. in Software Engineering at The University of Western Ontario. His work here focused primarily on the use of virtual reality for surgical simulation, specifically in relation to surgical robotics. Carl Byers is President of Logres Inc., a business strategy and innovation consultancy. Carl has been an active member of the international simulation community since 1986, defining and developing strategies for simulation companies producing solutions for training, engineering, and analysis applications. Carl has founded simulation companies, led simulation software development teams, is a co-author of the High Level Architecture (HLA) standard for simulation, and has been an invited speaker and representative at major simulation events. In his consultancy, Carl works with companies to build strategies for the innovative application of emerging technologies to mobile knowledge management and learning environments. He holds a B.A.Sc. in Engineering Physics from Queen s University, an MBA from the Schulich School of Business, and a Masters in Human Security and Peacebuilding from Royal Roads University. Eugene Fiume is professor of computer science at the University of Toronto. He has long been involved in computer graphics and digital media both in university and industry. In addition to co-directing the dynamic graphics project at the University of Toronto, he participates extensively in a wide variety of advisory and consulting roles in industry. He was the recipient of an NSERC Synergy Award in 2012 for his twenty-year university/industry collaborative work with Autodesk Research. He holds a B.Math degree in computer science from the University of Waterloo, M.Sc. and Ph.D. degrees from the University of Toronto, and he held an NSERC Postdoctoral Fellowship at the University of Geneva Paper No Page 1 of 9

2 Augmented Reality on Tablets in Support of MRO Performance Andrew Woo, Billy Yuen, Tim Hayes Carl Byers Eugene Fiume NGRAIN (Canada) Corporation Logres Inc. University of Toronto Vancouver, BC, Canada Ottawa, Ontario, Canada Toronto, Ontario, Canada [awoo byuen INTRODUCTION A severe bottleneck in the efficient maintenance, repair and overhaul (MRO) of complex equipment is the lack of skilled and knowledgeable personnel. This is compounded by lack of easy access to live equipment to support training needs. Consequently, virtual training solutions have evolved and are now used to familiarize the technicians in the maintenance procedures and related equipment information, usually in the form of 3D computer graphics and simulations (Kaas 2006). A high degree of realism and effectiveness can now be achieved with virtual solutions; however, nothing can replace the experience and effectiveness of the touch, feel, and appearance of the live equipment right in front of you, whenever it is available. Furthermore, skill and knowledge fade between initial training and field maintenance activities benefits extensively from virtual task refresher solutions and job aids. Since Augmented Reality (AR) combines virtual (computer graphics and data) information within a real-world environment, AR can be an exciting and attractive extension to support hands-on MRO activities; it effectively combines the direct physical benefits of access to the real equipment with the flexibility of elements of the virtual training environment. A visual example of an AR capability in the context of MRO is demonstrated in a video produced by the automaker BMW. The video depicts a sequence of procedural task steps to properly disassemble a portion of a car engine. To see the entire video, please refer to a snapshot of the video is seen in Figure 1. In this example, a technician wears a pair of glasses, and he sees the live engine, plus an AR overlay of the virtual (3D graphics) representation of the relevant parts (highlighted in yellow) of the car engine that he needs to remove for the current task step. Once he has completed removing those parts, he says next step and the next set of relevant parts are overlaid on top of the live engine. Though the video shows only disassembly operations, note that equivalent assembly operations can be applied in the same manner. Figure 1: Snapshot of the BMW video, showing a sequence of procedural task steps to disassemble the car engine. The part shown in yellow is the part in question for the current task step. In this paper, we discuss our work to assess and develop strategies for incorporating AR in various aspects of MRO activity. The paper initially reviews prior work pertaining to AR use and experimentation in the context of MRO of complex equipment. We then review the advantages and disadvantages of the most probable platforms on which MRO-based AR capabilities can currently be implemented. We discuss our recommendation of the tablet as the primary shortterm platform, with glasses as an up-and-coming platform. Critical MRO activities are then identified and the effectiveness of the recommended AR platforms is discussed in each case. We then proceed to discuss the implementation advantages of those AR capabilities considering, in particular, the 3D modeling needs and the benefits of reusing a single data representation for both AR and virtual environments. We finish off with future work and conclusions. PREVIOUS WORK The use of AR in MRO and training has been the focus of many research projects and collaborations, such as ARVIKA (Friedrich 2002), Services and Training through Augmented Reality (Raczynski and Gussmann 2004), and ARTESAS (Siemens AG 2009). Although the benefit of incorporating AR into the process has been overwhelmingly positive (Henderson and Feiner 2009) (Haritos and Macchiarella 2005), we have yet to see any sort of large scale adoption of the technology Paper No Page 2 of 9

3 A thorough survey of AR applications in manufacturing was compiled by Ong, Yuan, and Nee (Ong, Yuan and Nee 2008) and elaborates on both the advantages and shortcomings of the available technologies. Although there is now a significant body of work on the topic of AR in MRO and training, the primary focus of many of these endeavors has been on what we will categorize as removal and install (R&I) tasks (Henderson and Feiner 2009). This focus has led many of these projects to specific types of user interface hardware. These R&I tasks require both mobility and that the user has his or her hands free to perform the operations on the physical equipment. With this in mind, it is not surprising that most research projects favor the use of Head Worn Displays (HWDs) controlled by a wearable computer. Although many convincing solutions have been developed on this or similar technology, most are not deemed commercially viable since these backpack systems are cumbersome (Ong, Yuan and Nee 2008), costly (De Crescenzio, et al. 2011), maintenance intensive, or unsuitable for many applications (Wagner, Pintaric, et al. 2005). On the other hand, there has been a great deal of commercial adoption of mobile devices such as tablets and smart phones with cameras and the processing power to support AR applications. The tremendous improvements in graphical and processing power of these devices have resulted in an explosion of AR applications in various areas such as massively multiuser games (Wagner, Pintaric, et al. 2005), electronic tour guides (Schmalstieg and Wagner 2005), instructional aides (Wagner, Schmalstieg and Billinghurst 2006), navigation (Takacs, et al. 2008), and many others. Furthermore, many of these mobile devices are now equipped with other sensors such as multi-touch, GPS, accelerometers, gyroscopes, and magnetic compasses which can be used to refine the user experience and provide novel and robust interaction methods (Billinghurst and Henrysson 2006). THE TABLET AS A VIABLE PLATFORM In Figure 2, we identify various critical parameters by which to evaluate the viability of several generic platforms as candidates for the implementation of AR in complex equipment MRO applications. The platforms considered include the tablet, a laptop equipped with a (separate) camera, semi-transparent glasses (e.g. Google glasses) and semi-transparent and immersive helmets. Although glasses and helmets are both HWD platforms requiring additional external equipment and connections such as power and computers, both are sufficiently different technically, and deliver a different user experience, to warrant independent assessment. Each parameter was assessed and given a relative rating for each platform. The results are shown in Figure 2, where green indicates good performance, orange indicates average or less positive performance, and red indicates poor or least positive performance. The figure suggests that with a simple equal weighting of parameters, the tablet emerges as a clear winner for use today. The main shortcomings of the tablet, such as lack of hands-free capability and the need for context switching, may be less relevant in some MRO scenarios, as will be discussed in the next section. In addition, while the assessment of glasses indicates that a number of limitations exist today, it is likely that negative elements such as accessibility, cost, and display resolution are likely to improve in the not-sodistant future, making glasses a clear winner tomorrow. The above conclusion is actually contrary to those of previous papers (Henderson and Feiner 2009). However, we feel that the parameters listed below cover a broader range of requirements beyond simply the R&I tasks investigated in many previous studies. Accessibility Cost Hands-free No context switching Don t need connection to power source Low pain in wearing something Mobility Ease of setup Display Res. Confined space handling Tablet Laptop with camera Glasses Helmets Figure 2: List of key parameters to access each platform s viability. Note that green indicates good performance, orange indicates average performance, and red indicates bad performance. The following paragraphs examine each of the parameters in more detail, both in terms of its importance, as well as its ranking for each platform: Accessibility: In the same way that accessibility to the live equipment has been a 2012 Paper No Page 3 of 9

4 big bottleneck preventing training, the lack of immediate availability of glasses and helmets on the market or in today s MRO facility decreases their current potential benefit. In contrast, tablets and laptops are more readily available and provide a natural, familiar environment for use. In the future, however, we expect wider availability of glasses. Cost: Cost is a significant factor, especially in wide deployment scenarios in today s budgetconscious world. Because appropriate glasses and helmets are currently much more expensive than tablets and laptops, they are less attractive alternatives for an AR platform. Again, we expect glasses to drop significantly in price in the not-so-distant future, although attachment to processing hardware remains an additional cost. Hands-free: This is an important parameter in that the technician must be entirely hands-free in order to effectively perform many MRO tasks, especially R&I. Tablets and laptops are not hands-free so rate more poorly than HWD solutions. However, with the assistance of stands attached to tablet holsters (as seen in Figure 3), tablets may be sufficiently handsfree and adequate for some MRO tasks thus the orange rating. The hands-free advantage is also very important because maintenance work is usually done in a grubby environment. Figure 3: Successful experimentation with a hands-free tablet. Experiment showed inability to reduce context switching between the tablet and the live equipment. No context switching: With tablets and laptopbased AR solutions, the technician will always need to switch viewing contexts between seeing the live equipment and observing the AR on the tablet. Context switching can often slow down or impede MRO operations. In our own experimentation, we were able to confirm the challenges of context switching. Furthermore, even when the tablet was placed directly between the technician and the live equipment, so that the user could seethrough the device (similar to a glasses-based solution), there were significant usage challenges because of the lag in the tablet camera display and lack of stereo vision for pin-point accuracy. For example, in Figure 3, the technician found it very difficult to aim the screw-driver into the proper slot. Don t need connection to power or external computing source: This is an impediment for glasses and helmets as there are always wires (connecting to the power source or computing platform) that need to be routed away from the field of activity. This can easily cause accidents during MRO operations. While wearable computers and power sources are becoming more readily available, they still represent an obstruction that results in a low rating for this parameter for HWD solutions. Low pain in wearing something: Because many MRO operations can take a long time (a 30 minute procedure is not uncommon), the comfort of the user is an important factor. This is particularly true for helmets and, to a lesser extent, glasses. In addition, if the technician is already wearing glasses, or needs to wear safety glasses, putting on another pair of glasses or a helmet (for the AR) would cause additional discomfort and increase instability in the AR viewing position. Mobility: Mobility is a particularly important parameter in terms of the ability of the user to move around large pieces of equipment. Tablets are the most mobile, but can be hampered by a stand to achieve hands-free; laptops are similar, but are also hampered by the separate camera. Glasses and helmets are mobile but require a connection to a power and computing source. Ease of setup (and calibration): The degree of predictability and amount of time to set up the AR platform is also a measurement in the efficiency of the MRO operations. In this case, the tablet, with its built in sensors and cameras, is the easiest. Display resolution: High resolution displays are necessary when displaying geometrycritical 3D graphics. Currently, glasses and helmets support lower resolution displays. However, as will be seen in the next section, not all MRO scenarios require high display 2012 Paper No Page 4 of 9

5 resolution. In addition, it is only a matter of time before glasses and helmets support higher resolution displays. Confined space handling: There are cases in which the MRO operations occur in very confined spaces, some barely permitting the technician to squeeze into the space. Under such situations, glasses would be the only potential alternative. CAPABILITIES FOR AR In many AR applications, textual information is overlaid on top of the real world image. As this pertains to MRO, an example can be seen in Figure 4, where callouts can be used to guide the technician to perform a task step, and labels can be used to provide information about the relevant parts. These are effective tools within an AR environment, and can be seen in prior papers (Henderson and Feiner 2009). It is important to note that high resolution displays would not be needed with simple textual AR. In the following paragraphs we will explore each of these MRO applications to identify potential nontextual AR components, assessing the extent to which context switching is a barrier to the effective use of a tablet based solution in the short term. Removal and Install In our vision, each task step in the R&I operations for MRO can be represented as an AR snapshot, in which the virtual 3D geometry of the parts relevant to the current task step are displayed and highlighted (see Figure 7). This goes beyond simple textual information and provides an effective way to guide the technician toward what parts need to be assembled/disassembled and in which sequence. In many of the scenarios described by prior work, the virtual 3D geometry was not considered for display. However, this can be important in some disassembly scenarios, and is absolutely critical for most assembly scenarios (i.e., the importance and need to see what the part is and where it fits with respect to the live equipment). When applying the tablet to this capability, the major deficiency is the need to context switch between directly viewing the live equipment, and the AR on the tablet when using the AR environment for R&I direction. In the medium-term future, when glasses become commonplace, cheaper, and gain higher display resolution, they will likely overtake the tablet in terms of platform effectiveness for this type of activity. Parts Familiarization and Status Check Figure 4: Displaying textual information through callouts, on top of the live equipment. There are, however, many non-textual augmentation capabilities that are particularly valuable within an MRO context, including: R&I Parts familiarization and status check Collaboration Systems understanding Annotation and reporting Furthermore, most prior AR work focused on applications where context switching was identified as a significant concern, driving the platform towards an HWD solution. In the case of parts familiarization and status checks, the AR environment displays and highlights the virtual 3D geometry for each task step or system check so that the technician can recognize and identify the part. Directional cues can be provided within the AR environment to guide the technician to the appropriate location within the AR field of view. Furthermore, the AR platform can serve as a visualization technique to expose any relevant parts that may be hidden from view. In addition, the AR environment and user interface can also be used as a connection to inventory and parts ordering systems in a seamless environment to improve the maintainer s efficiency. In this activity, it is not necessary for the maintainer to physically remove or replace parts for this capability to function. As a result, the impact of context switching and hands-free deficiencies is effectively minimized and is of limited concern. Since the main purpose is to walk around the live equipment to familiarize with and 2012 Paper No Page 5 of 9

6 identify parts from various vantage points, a tablet is a very effective AR choice. Collaboration In simple scenarios, a task is typically performed by a single person. In others, particularly those involving physically larger systems, a small team may be grouped to allow tasks to be performed in parallel. In these cases, there is interchange of information and cues between the collaborating technicians within their respective AR environments, and a supervisor may also be present to oversee the progress of the technicians. In this scenario, there is a need to see a composite of the overlay information with the live equipment as well as cues and overlays from the activities of other collaborating technicians. While context switching is of some concern for individual maintainers, the tablet is an efficient platform for multi-mode collaborative interactions that can benefit from basic capabilities such as broadband communication and GPS. Systems Understanding warning or reminder to the technician during an assembly/disassembly task. In addition, if the capacity is available, the AR view can display overlays of actual equipment read-outs, levels, and settings transmitted from the live equipment. Because there is no physical removal of parts in this capability and a primary purpose is to view and observe the live equipment from multiple locations, context switching and hands-free deficiencies are not a concern, thus making the tablet a good choice. Annotation and Reporting 3D annotations can both be visualized and created when combined with the live equipment in the AR environment. One need for 3D annotation within MRO is battle damage assessment, where damages on an aircraft are important to note and understand for stealth aircraft. That understanding then leads to an awareness of the criticality of specific damage repairs. Historical damages or fixes can be displayed and notes associated with those entries can be viewed. New annotations can be made through interaction with either the live equipment or using the virtual interface. See Figure 6 for some sample 3D annotations (in red). Another application of the AR annotation capability within the training environment is to provide instructors and students with the ability to collaborate by inserting 3D annotations on top of 3D models in AR. Figure 5: Animated arrows indicating flow direction, speed, and other characteristics, on top of the live equipment. Systems understanding is in some ways similar to parts familiarization. However, in systems understanding, the maintainer develops a deep awareness of the theory of operation of equipment and systems. Typical elements of systems understanding include fluid and air flow paths, electrical signals, and temperature gradients. An example can be seen in Figure 5 regarding air flow animation, where a virtual representation of the flow can be combined with the live equipment. Using AR for systems understanding, concepts of flow direction, speed and other characteristics can be taught to the technician. They can also be provided as a Figure 6: 3D annotations (shown in red) on top of the live aircraft. The AR environment can then be used to select, assemble, and publish reports relevant to particular equipment locations or maintenance tasks using the live equipment as the selection reference point. In both of these activities, there is no physical removal of parts so context switching and hands-free 2012 Paper No Page 6 of 9

7 deficiencies are of little concern. The sensor and userinterface capacities of the tablet make it a very attractive AR platform for these activities. ADVANTAGES OF THE AR CAPABILITIES Effective application of AR within the MRO environment has the potential to increase performance and efficiency while decreasing error rates and equipment breakage. These outcomes can be achieved as a result of better equipment and systems understanding and the immediacy of access to relevant information by the technician. In addition, with new 3D graphics techniques and improved processing environments, the generation of the virtual elements of the AR environment is more cost effective and can be developed more quickly. Better Understanding and Information Immediacy The AR capabilities described above enable the technician to explore the live equipment to become fully familiar and aware of equipment parts, systems, and theory of operations. As a result, using AR with the actual equipment in front of him or her, the technician has an unprecedented, flexible, and engaging ability to develop an increased understanding of their MRO tasks. within the virtual environment (Kaas 2006) becomes relevant for scalability in AR. Within the context of R&I and parts familiarization, the main difference between the AR and virtual environments amounts to choosing the 3D geometries to be displayed. Figure 1 shows that, in the AR world, only the current task step s relevant parts geometry need to be displayed; in contrast, in the virtual world, the entire model is usually displayed. Thus if we build our virtual training software such that the true concept of a procedural task exists (as is the case using NGRAIN Producer), and each task step contains exactly the information as to which parts need to be assembled or disassembled, then it is trivial to isolate the necessary parts per task step to display and highlight in the AR environment. As an example, see Figure 7, where the highlighted part (in yellow) in a virtual model indicates the relevant part for the current task step. In the AR environment, only that yellow part is displayed and overlaid with the live equipment; in the virtual environment, the entire model is displayed, with or without highlighting. At the same time, the AR environment provides technicians with immediate access to current and historical information about the system on which they are working. This information can also include broadcast information related to local system status and, more broadly, the extended enterprise inventory, parts ordering, and reporting system. Using the AR environment, the individual technician can be more directly and tightly integrated with the complete platform and enterprise logistics chain. Cost-effectiveness and Scalability: Re-use between AR and Virtual Environments One of the main impediments to seeing more equipment benefiting from the AR capabilities discussed above is the historical lack of a scalable approach to creating the necessary content, including 3D models and system data. However, there is currently a large and growing amount of 3D content and other equipment data that has been created during the equipment design and development phases or that has been developed for specific training applications. Much of these models and data can be re-used within an AR context. In fact, if done correctly, by using the same content within the AR environment and the virtual training environment, prior scalable successes Figure 7: The highlighted part (in yellow) can be displayed equivalently in the AR and virtual environments Paper No Page 7 of 9

8 The reuse scenario across AR and virtual environments also provides considerable advantages for the technician. Given the same data representation on the same tablet, the technician can practice the procedural task in a virtual environment (in the schoolhouse or when travelling to the field site). Then, when the technician arrives on site to fix the live equipment, the same data can be switched to view in an AR environment. Performance and Development Efficiency: Reduced Geometry Requirements If the technician has the need to view the data in an AR environment only (and not in a virtual environment), then an additional optimization can be achieved. This optimization pertains to the virtual 3D geometry. We usually have the assumption that appropriate virtual 3D geometry is readily available; however, in our experience, it is rare that existing virtual 3D geometry meets the fidelity requirement of information for training and MRO needs. As a result, the virtual 3D geometry is usually created from scratch using digital content-creation tools. When 3D geometry is deployed in a virtual environment, we often find that additional geometry is required beyond that which is practiced upon; usually that extra geometry is required to provide visual context of where the really important parts are. In the AR environment, these visual context geometries are not needed, as they are already represented live using real equipment. Thus, the amount of time needed for creating geometry can be reduced, as would the computing requirements to effectively utilize the 3D models. FUTURE WORK In order to support full scale adoption of AR, improving the usability of the system is paramount. This has many aspects that will all need to be addressed, including ease of set-up and effectiveness and scope of user interactions. Set-up, Registration, and Tracking Even with the great advances in mobile device technology, there remain areas requiring further research. For instance, more work is required in the area of camera registration. Many manufacturers are reluctant to add markers to their equipment (De Crescenzio, et al. 2011) which ultimately requires the end user to apply markers to the equipment before the system can be used. An alternative is the development and adoption of solutions based on markerless technologies. One solution to the problem lies in the area of object recognition and natural feature tracking (Lepetit 2008). However, current implementations are computationally intensive and still are based on the assumption that there are clearly defined features on a planar surface (De Crescenzio, et al. 2011). As the technology continues to improve, the algorithms are optimized, and the inputs from the other sensors on the device (GPS, accelerometer, gyroscope, magnetic compass, etc.) are included, this may become a very interesting solution. There are also many other external tracking options but they each tend to have prohibitive drawbacks (Ong, Yuan and Nee 2008). Barring some new technology in this area, it appears that optical based tracking, perhaps assisted with other on device sensors, is a viable solution. User Interface and Scope of Interactions Although there has been work on mobile AR interfaces, we have yet to see interaction metaphors fully suitable for handheld mobile devices (Billinghurst and Henrysson 2006). This remains a clear opportunity for research. In addition, as object recognition improves, we can begin to consider new applications in the area of MRO and training, including per step validation, per part tracking, damage, and deformation discovery. Display Solutions It is expected that continuing developments will result in future improvements to HWD displays and wearable computers. As these technologies evolve, we may find many AR applications returning to these systems to overcome the limitations of hand-held displays. However, with a well-designed architecture, the type of display used should be largely irrelevant and ideally can be swapped out seamlessly to provide a continuum of capability with HWD s used interchangeably within existing AR systems. CONCLUSIONS In this paper, we have explored the potential for the use of AR capabilities on a tablet platform to support MRO activities. While the tablet has been identified through our analysis as an excellent short-term AR platform, some limitations of the platform in tasks demanding limited context switching indicate that glasses are an appropriate AR platform for the future. Various critical MRO tasks can benefit today from the application of a tablet-based AR solution, and effective design of the 2012 Paper No Page 8 of 9

9 algorithms and architecture will enable the seamless use of glasses in the future. In addition, the use of a single data representation between AR and virtual environments will increase the re-use of the investment in virtual training products, increase the level of comfort and familiarity of technicians using the environment, and increase the scalability of the overall 3D model development process. ACKNOWLEDGEMENTS Thanks to Graham Smith, Gordon Wong, Jonathan Young, and Justin Quach for the many discussion ns and experimentation on augmented reality topics that led to this paper. REFERENCES Billinghurst, M., and A. Henrysson (2006). "Research Directions in Handheld AR." The International Journal of Virtual Reality, pp De Crescenzio, F., M. Fantini, F. Persiani, L. Di Stefano, P. Azzari, and S. Salti (2011). Augmented Reality for Aircraft Maintenance Training and Operations Support. IEEE Computer Graphics and Applications. Vol. 31, no. 1, pp Friedrich, W. (2002). AVRIKA - Augmented Reality for Development Production and Service. Proceedings of the 1st International Symposium on Mixed and Augmented Reality. Paper 3. Haritos, T., and N. Macchiarella (2005). A Mobile Appication of Augmented Reality for Aerospace Maintenance Training. Proceedings of the 24th Digital Avionics Systems Conference, pp. 5.B Ong, S., M. Yuan, and A. Nee (2008). Augmented reality applications in manufacturing: a survey. International Journal of Production Research, Vol. 46 no. 10, pp Raczynski, A., and P. Gussmann (2004). Services and Training Through Augmented Reality. Proceedings of European Conference on Visual Media Production, pp Schmalstieg, D., and D. Wagner (2005). A Handheld Augmented Reality Museum Guide. Proceedings of IADIS International Conference on Mobile Learning (2005), Vol. 49, Issue: Pervasive, pp Siemens AG (2009). ARTESAS - Advanced Augmented Reality Technologies for Industrial Service Applications. Takacs, G., et al. (2008). Outdoors Augmented Reality on Mobile Phone using Loxel-Based Visual Feature Organization. Proceedings of the 1st ACM International Conference on Multimedia Information Retrieval, pp Wagner, D., D. Schmalstieg, and M. Billinghurst (2006). Handheld AR for Collaborative Edutainment. Proceedings of the 16th international conference on Advances in Artificial Reality and Tele-Existence, pp Wagner, D., T. Pintaric, F. Ledermann, and D. Schmalstieg (2005). Towards Massively Multi-user Augmented Reality on Handheld Devices. Pervasive Computing, Vol. 3468, pp Henderson, S., and S. Feiner (2009). Evaluating the Benefits of Augmented Reality for Task Localization in Maintenance of an Armored Personnel Carrier Turret. Proceedings of the 8th IEEE International Symposium on Mixed and Augmented Reality, pp Kaas, E. (2006). Deployable Virtual Maintenance Trainers: Case Studies on Using Interactive 3D Simulations to Replace Hard Trainers. Proceedings of I/ITSEC, paper No Lepetit, V. (2008). On Computer Vision for Augmented Reality. International Symposium on Ubiquitous Virtual Reality, pp Paper No Page 9 of 9