AIS Target Filter in SOB

AIS Target Filtering
When we arrived off Gibraltar in April, 2006 I first encountered the problem of chart clutter and the need to filter the target data. This page demonstrates the extent of the problem in a high traffic area and how DigiBoat developed a solution in their SOB software.
SOB Screen Capture, ALL AIS Data Displayed	This is the Strait of Gibraltar and Algeciras Bay which is shared by Gibraltar and the Spanish port of Algeciras. Sarah is berthed at Marina Bay in Gibraltar in roughly the center of the screen. How do I know? With this screen I have created an extreme example of chart clutter caused by the plotting of AIS target data. In this screen I have turned target labels with the ship's name and target tracks for the maximum time. The result is all of the chart information on Algeciras Bay and most of the Strait are obscured.
SOB Screen Capture, Target Labels and Tracks Turned Off	This is the same view with a little of the clutter removed. I turned off the target labels and tracks. Still Algeciras Bay is pretty well obscured. This illustrates the need to for effective filtering of target data by whatever package is used to plot the AIS targets. Of course the Strait of Gibraltar is one of the most dense shipping traffic regions in the world. Most areas of the world would not present this level of target clutter. Also the chart is zoomed out quite a bit. This not a zoom level that would normally be used for detailed navigation information.
SOB Screen Capture, Chart Zoomed to a Level That Can be Used for Navigation.	In this screen capture I have zoomed in to cover most of Algeciras Bay and now the screen is much less cluttered. Although this capture was taken at a different time (on my way out of the Med rather than on my way in) the number of vessels in the Bay is not that much different from that in the screen captures above. Although zooming in has restored the chart to a useful for navigation state, there is still a problem with some of the AIS targets displayed. Only a few of the targets shown are actually underway. None of the targets underway represent potential collision threats. Three targets are highlighted (Red). All of these targets appear to be anchored. The two targets in Algeciras Bay are reporting a COG that is toward Sarah's position, but the SOG values reported are less than .3 kts.
I have displayed detail the target (Clipper Lady) to the east of Gibraltar, which is another anchorage area for Gib. Actually there is a fourth highlighted target just NW of Europa Pt. at the bottom of the screen. It shows up as purple because the vessel is bunkered with a fuel ship and the two targets (one red, one yellow) are superimposed on top of each other. The problem here is that none of the red targets is actually a collision threat. They are all at anchor. However the small and changing errors in GPS positions causes the ships to report slight movement in their dynamic data, as does Sarah's GPS data. The result is SOB will extrapolate those very small movements for each target and Sarah and some of the target will appear to be on a potential collision course. Of course this is an erroneous status - i.e., SOB is crying wolf. If Sarah were not berthed in Marina Bay, but actually underway in Algeciras Bay some of the few targets that are actually underway might be true collision risks, but they could be obscured by the many targets erroneously flagged as dangerous. I encountered this situation on my initial stop in Gibraltar in April, 2006 and reported it to the DigiBoat software development team. At first they were understandably reluctant to modify their algorithms for identifying dangerous targets, but fortunately they do not appear to have any lawyers on the staff and they quickly changed their position realizing that presenting multiple targets erroneously as dangerous is as bad as missing a few real ones. Within a few days they provided me with a alpha release version of the software with a filter for slow moving targets. At the time I had a number of equipment problems on Sarah (see Log of the 2006 Cruise for details), and I was not able to thoroughly test the software for DigiBoat for several weeks.
SOB Screen Capture, Ignore Stationary Targets Turned Off	On my exit of the Mediterranean I stopped once more at Marina Bay in Gibraltar and had time to test this filter option. By that time the feature had been released in an official version of the software. I downloaded the latest version at the time (6.10) and decided to test it. Here is an initial screen capture after the upgrade to 6.10. Caleta, which is tied to the dock in Gibraltar is highlighted in red as a high collision potential.
SOB Screen Capture, Setting the Ignore Stationary Filter to 1 kt.	I set the Ignore Stationary filter (third from the bottom on the right side of the screen) to a value of 1.0 kts.
SOB Screen Capture, Anchored Targets No Longer are Flagged as High Collision Potential	And Caleta is no longer a collision threat.
SOB Screen Capture, Vessel Underway Flagged for Potential to Come Within the Safe Distance Value	I waited until one of the vessels in Algeciras Bay got underway and was headed in my general direction to insure SOB still correctly flagged the moving targets. The target at the head of Algeciras Bay is moving at 4.5kts, and if it continued on this course would come within the 1.0nm Safe Distance I had set. So I verified that SOB will not flag targets as dangerous if their Speed Over Ground (SOG) is less than the value set in the filter control. I also verified that targets moving faster than the control setting will still be flagged as dangerous if they are projected to come within the safe distance specified.