Abstract: The objectives and conditions of a port and waterway study in Norfolk, Virginia, required a research methodology with two interactive bridge simulators. The study was conducted at the Computer Aided Operations Research Facility (CAORF) of the United States Merchant Marine Academy (USMMA), under the sponsorship and assistance of the U.S. Army Corps of Engineers, Norfolk District (ACOE-ND) and Engineer Research and Development Center (ERDC) Hydraulics Lab. After the project was underway it was decided that a second bridge simulator was necessary to support two-way meeting situations, with a pilot controlling each ship (bridge). Both study resources and schedule were tight. The CAORF management and engineering team developed an innovative approach to the design and construction of a second bridge. The bridge was constructed in a period of weeks, meeting the project schedule. It was effectively integrated into the simulation system, resulting in two independent ships/bridges interacting in scenarios. The second bridge had the full compliment of instrumentation, hydrodynamic and data base features, and included a 187^○ horizontal visual scene with a diameter of 16.2 feet. Although of simplistic appearance and thrifty setup, the second bridge simulator, having crisp and seamless images, received complimentary acceptance from the test pilots (some of whom noted the layout was similar to the bridges on certain vessels).

　

 This paper will describe: Research needs and design requirements for the second bridge; Project constraints in schedule and cost; Design approach (Applied/pragmatic philosophy; Available technology (Norcontrol simulation and control panel software, Commercial-Off-The-Shelf hardware, such as console displays, and projectors.)); Implementation issues (Technical challenges and solutions (Space constraint, Edge-matching, Facilitating communication with main bridge simulator and control station)); Resulting system (System configuration, Operational experience of Control Station Operator and research staff, Acceptance by pilots, Maintenance and support issues).

　

 The US Army Corps of Engineers (ACOE), headquartered in Washington, DC, has 8 Divisions all over the US with 38 district offices in the US, Asia and Europe. The Norfolk District (ND) is one of six districts within the North Atlantic Division. The ND performs many military and civil works functions within the state of Virginia. Included among the ND's civil works responsibilities is construction and maintenance of Congressionally authorized federal navigation channels servicing the Ports of Hampton Roads, Virginia.

　

 The ND and the VPA recognize that Post-Panamax container vessels will soon call on Virginia's Ports.

　

 The Norfolk Harbor Federal Navigation Channel System allowing access to Virginia's ports requires improvement to provide service to these large ships.

　

 To determine the design requirements for the channel system, the ND office partnered with the Computer Aided Operations Research Facility (CAORF) at the US Merchant Marine Academy (USMMA) in Kings Point, NY, to perform ship simulation studies to support the channel deepening effort.

　

 The District's overall study objectives were to provide safe navigation channels while looking for ways to reduce dredging costs. The District chose to investigate three primary areas: Thimble Shoal Channel (TSC), Atlantic Ocean Channel (AOC), and Norfolk Harbor Reach (NHR).

　

 The requirement for a man-in-the-loop simulation approach using an interactive second bridge to support the Norfolk Harbor Reach was the driving force for the rapid development of this new interactive bridge simulator at CAORF.

　

2.1 Background

　

 The third element of the 50' Inbound Channel slated for improvement is the NHR, which is currently maintained by the ACOE as an asymmetric channel. A 650' wide outbound element is maintained to 50' deep, and a 600' wide inbound element is maintained to 45' deep. A channel design consideration was to move the western channel boundary inward 250', to change the channel width from 1,250' to 1,000', and for incremental deepening of the inbound side of the channel to 50'.

　

 The purpose of this study was to evaluate this 1,000' wide NHR Comparative Test Channel Configuration (CTCC), when used by deep draft inbound containerships. The CTCC was tested to determine if a reduction in width would be feasible. Such a reduction in width would reduce the quantities to be removed from the channel with the future dredging contracts. This reduction in quantity would save the Corps and the local sponsor significant funds on the 50' Inbound Channel deepening project.

　

2.2 Design and Schedule Requirements

　

 This study was conducted using both the CAORF full-mission VBSS and the second bridge in an interactive mode. The requirements for the second bridge functionality and performance was driven by the research project needs and are highlighted in the following First, the overall requirement was that the second bridge be a high fidelity full mission simulator providing appropriate visual field of view, bridge controls, ARPA, ECDIS, complete ship hydrodynamics and environmental effects, etc. Second, the second bridge needed to utilize the same visual and environmental databases and ship model characteristics that were functional on the VBSS. Third, the second bridge needed to be able to operate either independent of the VBSS or in an integrated mode, interconnected with the VBSS in such a manner as to allow independent assignment of an own ship with total eye-point realism and interaction with the VBSS own-ship in two parallel man-in-the-loop simulation scenarios.

　

2.3 Design Approach

　

 The criteria for the design concepts were based on. a number of constraints and factors that would be key to the implementation methodology. They included:

2.4 Key Decisions

　

 The first key decision that had the most significant bearing on the second bridge design was where to locate the second bridge and how much space would be available for it. Several factors were involved with the decision: The second bridge needed to be in close proximity to the VBSS Control Station to facilitate set-up and the interactive operation associated with the research project. Figure 1 shows the candidate location.

　

 Second, the second bridge needed to have at least a 180 degree horizontal field of view (HFOV) to meet one of the full mission bridge criteria and provide all the visual cues needed by the Test Subject (operator/pilot).

　

 Third, confinement of the eye point of the Test Subject was deemed necessary, with the Test Subject sitting in a chair in front of the own ship controls, ARPA display and ECDIS display, in order to minimize the space required for the simulator.

　

 Fourth, the need for an operator working with the Test Subject to control the simulator was considered a needed function, although, as the use of the simulator proved out, the Test Subjects were comfortable steering the ship themselves and the resultant implementation quite user friendly.

　

2.5 Facility Constraints

　

 Because the facility issue was a key issue forcing many subsequent decisions, it was decided that the second bridge would be housed in a room on the same floor in CAORF as the VBSS and near the VBSS Control Station. This location would accommodate the key decisions mentioned above. A small video camera with audio would be utilized to show the VBSS Control Station Operator what the second bridge display was and note any comments of the Test Subject.

　

 As a result, the maximum dimension available for the second bridge simulator was 12' 10 1/2" in depth. Width was not an issue because of the rectangular layout of the room. The room was considered wide enough to probably accept a third bridge if ever deemed appropriate (layout plans were developed after second bridge implementation and did show that a third bridge could be accommodated).

　

2.6 System Configuration

　

 A key consideration for the system configuration was the utilization of multiple PC's to provide all desired functionality, interconnected by a LAN. Figure 2 shows the system configuration considered appropriate for the second bridge.

　

2.7 Projection and Display System Design

　

 With the space constraint of 12' 10 1/2" for a horizontal field of view (HFOV) of at least 180 degrees the overall diameter of the screen was determined to be 154" based on an eye point distance of 97" that still provided sufficient space for a seated test subject and a console. The decision to implement the display using flat panels rather than a curved screen facilitated using standard projectors rather than special projectors that could accommodate distortions (pincushion, barrel, prismatic, trapezoidal, etc.) as were used for the VBSS cylindrical projection screen.

　

 Since the VBSS visual database was designed for 26.66--- degrees horizontal field-of-view (HFOV) having an aspect ratio of 4:5 (height : width ratio) for each projected image. To achieve at least 180 degree HFOV would require 7 projectors, providing 186.66--- degrees HFOV overall.

　

 To maintain true perspective as seen from the eye-point, the width of the projection screen panel would be governed by that angle and the eye-point distance.

　

 Consequently, the initial solution based on an eye point distance of 97" (the radius) and a 26.66--- degree HFOV resulted in a corresponding projection screen panel width of 46". Refer to Figure 3 for the resultant overhead layout.

　

2.8 Projection and Display System Implementation

　

 Implementation was a significant challenge in such a tight configuration in finding a solution on where to put the projector, predicated on the projector's distance to the panel, which was governed by the maximum achieve a projected image width (for a HFO)V of 26.66-degrees) that would fit the projection screen panel width of 46". The use of a standard COTS projector for high brightness at minimal cost was an additional factor to be considered.

　

 A trade-off study of available technology clearly showed that DLP technology was the way to go. This was the technology that was implemented for the VBSS projection system and had the key features inherent in this technology: high relative brightness, inherent color convergence (one image surface/one lens), digital pixel resolution (sharpness), and low relative cost.

　

 A proof-of-concept set-up was created and three COTS projectors were tested in the desired configuration: a Davis Model DLP XIO projector having a high quality Zeiss zoom lens, a Toshiba DLP Model TVL projector with a standard zoom lens, and an Epson DLP projector with a standard zoom lens.

 Out of the three candidates, only the Davis projector was suitable because the projector-to-screen throw distance was significantly less than the others because of the Zeiss lens employed.

　

 That distance, with zoom set to near maximum, defined the physical location for the projector. This location near the eye-point, and as such, could not be realized unless the unit was located over the test subject. OK for one projector but 6 additional? That would require a clustered configuration, overhead mounted, with crossing image paths, and need to provide geometry correction. Mounting, alignment difficulties, and geometry correction complexities indicated that an alternate and simpler configuration was needed!

　

 The innovative idea to fold the optical path using mirrors and locate the projector(s) between the screen and the test subject station, floor mounted, and introduce angular compensation through mirror and projector tilt satisfied the geometric distortion and simplified mounting and alignment requirements. Figures 4 and 5 show the geometric configuration for the center channel and its location relative to the test subject console.

　

2.9 Simulator Setup

　

 The visual projection/display system design and implementation required accurate horizontal placement of screen, mirrors, and projectors to achieve a distortion-less view from the eye point. This required that each screen be located such that the screen center of each panel is perpendicular to the line of sight from the eye point.

　

 In addition, each mirror needed to be centered and horizontally perpendicular relative to the screen's surface.

　

 To easily accomplish this alignment a special laser tool was devised using a standard laser pointer mounted on a carpenter's level. The eye point was defined using a plumb line attached to the ceiling at the eye point distance from the screen. Second, a mirror was mounted on the surface of the center panel (a flat black surface prior to attaching the white vinyl screen surface) and using a plumb line attached at the top of the screen at the half-width point, the vertical centerline of the screen was defined and visible at the mirror. By placing the laser tool approximately 5 feet high behind the eye point with the beam passing through the eye point, the center panel was horizontally positioned precisely to achieve a reflection from the mirror's vertical centerline of the beam back through the eye point plumb line.

　

 Next, the beam was angled down sufficiently to be reflected off the screen's mirror's vertical centerline on the screen while passing through the eye-point vertical, illuminating the projection mirror thus precisely defining the horizontal placement of the projection mirror. The reflection off the projection mirror back to the screen then allowed precise angular positioning of the projection mirror to be horizontally perpendicular to the screen's surface. It also allowed correct horizontal placement of the projector lens centerline.

　

 The remaining six panels were set-up based on eye point distance to the screen s vertical centerline. The laser tool was subsequently used to locate and align the projector mirrors and projectors. After the vinyl screen was installed a geometric test pattern from the image database was used to confirm alignment and geometry. No distortion was visible.

　

 The test pattern provided boundary lines for the 26.66 - degree HFOV and boundary lines for the additional 3 degrees overlap. The projector's zoom lens was adjusted for seam-to-seam coverage of the 26.66 - degrees for each channel. Masking of the projector mirror adjacent channel sides until the overlap disappeared worked extremely well, as seen in Figure 6.

　

2.10 Simulation Platform

　

 The decision regarding the simulation platform was predicated on what was being used for the VBSS and, through its interconnection to the second bridge, how to functionally interconnect their interactive behavior.

　

 The VBSS simulation platform is a Kongsberg Norcontrol "Polaris" simulation system capable of simulating multiple own ships, bridge equipment (simulated functions as well as real bridge equipment stimulation, and operates with environmental and visual databases. Kongsberg Norcontrol also has a simulation system, "Panorama", that is essentially a "Polaris" simulator without the bridge equipment interface. "Panorama ' utilizes computer displays to provide own ship controls, ARPA radar display, and ECDIS display.

　

 It is designed to share "Polaris" databases, own ship models, visual database, etc. and operate either as a stand-alone simulator or in an interconnected mode with the VBSS "Polaris" simulator. This was the obvious solution and required no special development or integration efforts, and, as such, was the only approach worth considering. Setup and operation proved such.

　

 The "Panorama" simulator, as indicated above, was designed to provide complete bridge functionality through three CRT displays. Actual use by Test Subjects proved that the functional and ergonomic user friendly displays and controls, as well as the realism of the ARPA and ECDIS system displays, did allow realistic decision making and control of own ship. The research significance is that the research results for the Norfolk Project were deemed valid for all scenarios run.

　

2.12 Image Generation System -

　

 The development of image generation PC-based technology has progressed to such an advanced stage that consideration of such for the second bridge was essential if a low cost COTS implementation scheme were to be attempted.

　

 Kongsberg Norcontrol and others had been investigating various products for this function. At CAORF a HP Vectra 400 PC was configured with a nVIDIA g FORCE 4 MX 440 64 MB DDR AGP 4x/2x graphics display board based on advisement from Kongsberg Norcontrol that this board would appropriately provide a suitable image of a single visual database channel. Such proved out. Anti-alias functionality was less than desired and a superior display resulted with this function disabled.

　

 At a subsequent time, a board failure forced trying an available alternate, an ATI Rage 128 MB graphics display board, and it worked. Substitution did require downloading and installing ATI drivers for the Windows NT Operating System.

　

2.13 Image Display System -

　

 As mentioned above, the Davis DLP projector was the image display choice. The issue of abutment of adjacent channels with minimal seam needed to be resolved for a professional-looking display. The VBSS used projectors with special internal software that functionally varied the pixel intensity at the left and right edges. Reasonably good results were achieved after much "tweaking". Not a low cost solution since each VBSS Davis DLP projector, specially modified for this and other geometric adjustments cost about $15,000.

　

 A simple yet innovative and patent pending approach to achieve superior results using the unmodified Davis DLP projector, at a cost of $4,500 each was developed. The key concept consisted of displaying an additional 6 degrees on each side of the 26.66---- image as an adjacent channel overlap and masking each edge of the mirror to achieve an image intensity falloff from the edge of the image at the screen seam. Since the mirror was located closer to the projector lens than the screen, the falloff was gradual. Given the adjacent images were superimposed on each other yet each with their intensity falloff, the additive intensity resulted in constant intensity across the seam and a truly seamless display resulted. Figure 6 shows a photo of the achieved display result as well as the overall simulator.

　

 The projection screen presented another interesting challenge. There are many ways to build 7 panels with a white surface, tying them together and eliminating the joint seam. For the second bridge the decision to go COTS and buy a screen system made to our specified width and height resulted in a simple yet rigid and portable configuration that required only a minimal effort to assemble and attach the white vinyl surface using Velcro. Since the panels were floor standing, no support system was necessary.

 Subsequent tear-down and reassembly in an alternate location proved to be simple as expected because of the Velcro method of attaching the vinyl surface and the unique method the manufacturer used to attach each panel together (a plastic piano hinge) that allowed precise angular setup of the entire 7 segment assembly.

　

 The visual HFOV for each channel was set to be the same as that of the VBSS to achieve exactly the same view as seen from the VBSS Bridge at the preferred viewing point (eye-point located behind the bridge console on the bridge centerline). As mentioned above the projected HFOV for each channel was an additional 3 degrees on each side to achieve appropriate overlap for seamless edge matching. The resultant vertical field-of-view (VFOV) based on the 5:4 aspect ratio was 26.13 degrees.

　

 A test pattern projection simplified projector and mirror placement, angular setup, zoom lens setting. It also showed any keystone geometric distortion, which was eliminated through appropriate projector and mirror tilt in relation to the eye-point height and the horizontal mounting surface. It must be noted that the projector internal optical design includes vertical lens offset relative to the DLP surface to achieve a projected image whose bottom is at nearly the same height as the projector without visible keystone geometric distortion.

　

 The eye-point height was an important decision that affected screen height and Test Subject Console surface height. The decision to have the Test Subject sitting in a chair at all times allowed the eye-point to be considered "essentially" stationary.

　

 A console template was fabricated out of cardboard whose height could easily be adjusted. An armless vertically adjustable chair was selected for the Test Subject's use. Individuals of the CAORF staff of various heights (from 5'6" to 6'4") were used to define the desired eye-point height of 48 1/2" and console surface height of 28 1/2". Figure 4 shows the resultant vertical projection and display system layout.

　

2.15 Test Subject Console

　

 The decision to use flat-screen 17" color LCD displays for the three monitors (Bridge Controls, ARPA display, and ECDIS display) was a key decision that allowed for sufficient distance from the back of the monitors to the screen to accommodate the mirror/projection system. A graphical solution based on the projector's horizontal projection at 95% zoom (a slight amount available for fine-tuning edge matching and overlap) provided the answers as to mirror width and mirror/projector placement that would fit in the available space. The results were a compact physical layout that accommodated the facility constraints.

　

 A proof of concept for three channels was rapidly implemented using milk crates and reams of paper along with the monitors, cardboard console template, and a mirror/projector mount fabricated out of plywood and 2"x4" wood. It worked as planned and confirmed the placement of the monitors and the dimensions of the console along with the mirror/projector mount. One important factor that was addressed in setting the height of the mirror/projection system was that the mirror height would need be such as to block the light from the projectors reaching the Test Subject but not block the view of the image displayed.

　

 Another advantage of the proof of concept set-up was the use for determining monitor placement to maximize compactness, ergonomics, and Test Subject functionality and comfort.

　

2.16 Commercial Off-The-Shelf Systems / Equipments (COTS) -

　

 Throughout the conceptual, design, and implementation effort focus on the application and use of COTS was maintained. As a result the only fabricated items consisted of the console (fabricated out of 3/4" plywood using the cardboard template), the mirrors (10"x10" rear surfaced), and the plywood mirror/projector mounts. Material costs for all of this was under $300!

　

2.17 Portability-

　

 Throughout the conceptual, design, and implementation effort focus on portability was also maintained. All parts of the simulator are free-standing. Rapid tear-down and reassembly was proven when the need arose to relocate the simulator. Tear-down and removal of it all took less than 2 hours. Subsequent set-up took approximately 8 hours, primarily governed by panel assembly and alignment fine-tuning.

　

2.18 Schedule-

　

 Schedule was tight. Less than 2 months were available from determination of need to when it had to be on-line for research utilization. It was an exciting project because of the limited staff to achieve it (paper co-authors) and the uncertainties and technical challenges to be overcome. It's an easy thing to say "jam it in there"; it's an altogether different thing to conceive, design, and implement it!

　

2.19 Cost

　

 It turned out to be an amazingly low cost simulator. Cost drivers that are normally high were mitigated by maximizing the use of COTS, using PC-based simulation and visual imaging hardware, using a standard DLP projector, and a pre-fabricated projection screen configuration. Custom fabrication was limited to using plywood, 2"x4" wood, and tablecloths! Approximately three man-months total and $53,520 in materials as shown in Figure 7 was involved in this project from start to finish!