On the morning of November 7, 2007, the container ship Cosco Busan left California’s Oakland harbor for Hong Kong. The fog was so dense that sometimes the Chinese bridge crew or the American pilot could not even see the fore of the ship. Instead, the crew used the radar and the Electronic Chart and Display System (ECDIS) to navigate.
After leaving the Oakland Inner Harbor, the pilot should have aimed for the 2,200 foot span between the Echo and Delta towers of the San Francisco-Oakland Bay Bridge. Instead, he sideswiped the protective fenders on the Delta tower’s footing. The impact ripped a gouge in the side of Cosco Busan, discharging 58,000 gallons of bunker fuel into San Francisco Bay.
The pilot claimed that the radar image became “distorted” and that he had to switch to the electronic chart plotter where he could see the ship’s position overlaid on an electronic chart. However, the pilot could not make out the symbols on the chart display and had to ask the captain to point out the center of the bridge span. Evidently, there must have been a miscommunication as the pilot now set course for the Delta tower.
The ECDIS displays an electronic nautical chart (ENC) with a superimposed symbol of the ship’s position that is based on GPS data. The symbols of the ENC are standardized, but different styles of symbols and layers of information can be turned on or off, thus making it possible to customize the presentation. Using the international IHO S-57 ed. 3 standard, the ENC could, for instance, look as shown in A-C in Figure 1, depending on whether “modern” or “classical” symbols were used or whether lights were on or off. When on C, the green beacons indicate the free passage in the center of the span, and the red beacons indicate the towers. The Delta tower also has a red buoy on each side. (Compare also with the traditional paper chart at D.)
According to U.S. National Transportation Safety Board (NTSB) officials, the ship’s electronic charts may not fully comply with international standards. The transcripts from the bridge voice recorder released by the NTSB shows that there was a discussion between the pilot and the captain after the crash:
“You said this was the center of the bridge,” the pilot says.
“Yes,” the captain responds.
“No, this is the center. That’s the tower. This is the tower. That’s why we hit it. I thought that was the center,” says the pilot.
“It’s a buoy,” the captain replies.
After the collision the captain says in Chinese to a crew member: “What I said to him was not incorrect. This is the center of the bridge, not of the channel. As the pilot you should know full well.” (Note that the captain uses “you” as if he were talking to the pilot, but he is not.)
Still later the pilot apologizes to the captain. “Sorry captain, I misunderstood the chart, I thought that was the center.”
As I write this, the NTSB has not yet released its investigation of the accident. However I have no doubt that this will be yet another accident caused by “human error.” Evidently, the pilot, relying on chart and radar images in zero visibility, misunderstood the chart and mistook the center of the bridge for the center of the span under which the navigation channel extends. It also seems that he was unfamiliar with the symbols used in the onboard electronic nautical chart.
Usability on the Ship’s Bridge
“Attention narrowing” is a normal human reaction when, in a stressful situation, we try to focus on what we believe is the most relevant information. The result is that we often overlook other important information. Much can be said about the unfortunate decisions and lack of foresight and planning in this accident. However, it is a fact that a modern ship equipped with all technical means, a fully qualified crew, and a pilot with twenty-five years of experience still managed to collide with a bridge in close vicinity of the port. This kind of accident might be prevented by a new type of chart display described in this article.
Growing industrialization and increasing world trade will result in increased shipping at higher speeds. The world’s ship fleet is growing rapidly, and with that growth comes an increased shortage of ship’s officers. This will lead to a decrease in experience among bridge crews.
It is generally agreed that more than 80 percent of all accidents in the maritime industry are due to what is called “human error.” But I believe that much of what is called human error is due to design error. There is much work to do to improve usability on the ship bridge.
The Cosco Busan case is just one example that in spite of all electronic navigation devices on a modern ship’s bridge, bridge crews sometimes lose their orientation. Reasons for this might be excessive cognitive workload caused by fatigue, short decision times due to high speed, or too many instruments to read and integrate. Viewer-centered map research aims at designing navigational aids to facilitate human integration and provide cognitive off-loading.
In an information design research project, I evaluated alternative ways to display navigational information on board ships. By tradition, maps are drawn using a so-called exocentric, or bird’s-eye, view. The map depicts the world from above, using generalized land forms and abstract symbols. When the user compares the map with the real world, she has to first imagine herself standing in the map, gazing in a certain direction, and then imagine what she will see. Finally, she must then compare this imagined picture with the real world. In other words, she has to perform a mental rotation.
Mental rotations have been investigated by cognitive science, and we know that they draw heavily on working memory, especially when there is more than one rotation involved—for example, when you are travelling south using a map in a north-up orientation (see Figure 2). Since you are facing the “down” direction on the map, right on the map will be left in the real world. Most of us have experienced this as tourists in a foreign city, and we often solve the confusion by turning the map so that forward is up on the map and right on the map is also right in the real world. This display orientation is called “head-up” (not to be confused with the head-up displays used in fighter airplanes, which refers to information mirrored on the cockpit window so that the pilot can see it without looking down at the instrument panel).
“Real seamen never turn the map upside down.”
By tradition, charts and radar on board a ship are used in a north-up orientation. Sometimes both north-up and head-up are used. In Figure 3, the navigation officer in a Swedish combat boat can be seen in front of his head-up radar and his north-up electronic and paper charts. Outside the window, he has the egocentric (self-oriented) view of the world. Integrating real-time charts and radar images into coherent situational awareness as the combat boat navigates a complicated archipelago at 40 knots is cognitively a very demanding task.
My suggestion is to let the computer integrate the chart into an egocentric bridge-view of the world. This view could be used in the conning (driving) situation, eliminating the need for cognitively demanding mental rotations. To allow for such an egocentric viewer-centered display, topographical 3-D information as well as 3-D models of buildings and other landmarks may have to be added to maps. Figure 4 shows an example of such an integrated 3-D nautical chart.
The question now is whether navigating with such an egocentric perspective really is easier than navigating with traditional maps. To try to answer this question, I conducted a laboratory experiment at Malardalen and Chalmers Universities in Sweden.
Faster Decision-making with an Eqocentric View
We constructed a maze in which subjects could drive a small cart equipped with a laptop computer. Infrared cameras tracked the position of the cart and sent it to the laptop in real-time to mimic a GPS system. Through the maze, there was a winding channel of “deep water,” and on the sides were “shallow areas” where the cart would “ground.” The deep or shallow areas were not shown on the floor, only in the maps.
The only visible features in the studio were a couple of cardboard boxes, a chair, and a paper tube, which acted as landmarks. These were also depicted in the maps. Seventy-five subjects, amateurs and professional mariners, each drove the cart four times through four different but equally difficult maze designs. Each time they drove through a new maze, they were randomly assigned one of four map types: a traditional paper map, an electronic map in north-up orientation, an electronic map in head-up orientation, and the new egocentric 3-D chart (see Figure 5).
The subjects were told to drive through the maze as quickly as possible with as few “groundings” as possible. Efficiency was measured by timing the passage through the maze and logging the number of errors made. If the cart missed the deep water channel, a loud signal was heard from the laptop computer and the subject had to stop and back off the ground.
Finally, each subject ranked the user-friendliness of the systems. The results clearly showed that the egocentric 3-D view provided faster decision-making and fewer errors than the traditional map types. The 3-D map was also judged the most user-friendly. (For more details of these experiments, see http://www.upassoc.org/usability_resources/conference/2007/prp_033.pdf.)
A prototype 3-D nautical chart for the Port of Gothenburg on the Swedish west coast is presently being constructed and will be tested in a real environment (see Figure 6).
Design Implications for the Rest of Us
Awareness of perspective and the implications of mental rotations are needed not only when making different kinds of maps and way-showing, but also when it comes to assembly instructions and manuals. This awareness also affects the design and placement of controls in environments such as a ship’s bridge. The guidelines of the American Bureau of Shipping say, “the consoles, including a chart table if provided, should be positioned so that the instruments they contain are mounted facing a person who is looking forward.” However, ships do not always comply with this rule.
In car navigation systems, the viewer-centered view has become common in recent years—the map is displayed at an oblique angle, mimicking a primitive “3-D view” and changes continually while it tracks the driver’s progress.
The same principles are useful when making emergency exit maps. How should the map on the inside of a hotel room door look? If your evacuation route is to the left outside the door, but the map is oriented so that your room is shown on the top side of the corridor with the door facing down, you will likely need to think twice about which way to run.
In some situations there might not be time to think twice.