GPS is almost universally used for navigation by yachts and the use of GPS based ‘tracking’ is becoming common at events for boats of all sizes. As a result, parties frequently seek to present GPS based evidence in hearings.
GPS evidence can often be very helpful, but it must be considered in the light of the limitations of the technology used, the GPS installation and the circumstances of the incident.
Position accuracy (also known as absolute accuracy). GPS positions and tracks give an illusion of precision. Positions are usually quoted to three decimal places of a minute of latitude and longitude; for latitude, this is equivalent to about 1.85 metres. However, most modern manufacturers quote an accuracy of 3 to 5 metres to a 95% certainty level (note 1). This means that there is a 95% probability that the GPS antenna is located somewhere inside a circle of that radius with the quoted ‘position’ at the centre. There is a 5% chance that the GPS antenna is outside that circle.
The uncertainty arises from multiple sources, including atmospheric conditions, inherent uncertainty of the positions of the GPS satellites in orbit, the installation conditions of the GPS antenna, etc.
Higher accuracy is possible using techniques such as differential GPS, but the equipment and technical infrastructure required is expensive and is not currently common in marine navigation or fleet tracking systems.
Relative accuracy. If the evidence is calculated from the position of two GPS units, for example a distance between two boats, some of the uncertainties in the positions of the two units ‘cancel out’ as they affect both units in the same way. As a result, the uncertainty of such a calculation will often be less than the uncertainty in the absolute positions.
The uncertainty will depend on the installation conditions of the GPS units, their distance apart and whether or not they use the same GPS hardware and software. In ideal conditions (both antennas installed with a clear view of the sky and clear of signal reflections, identical hardware and software, and within a few hundred metres of each other) there are anecdotal statements that the uncertainty may be reduced to as little as +/- 0.5 metre. However, there is currently no published research on relative GPS accuracy in a marine environment and hardware manufacturers do not quote relative accuracy data for their equipment.
If one or both of the GPS units is moving, it is necessary also to consider the relative timing of the two GPS positions. If the ‘fixes’ were not made at the same moment the movement of the objects in the time between the fixes must be taken into account. This is a complex mathematical calculation. This is particularly important when using a tracking system as it is common for the interval between position reports to be several seconds.
In the worst case, the absolute circle of uncertainty should be applied to both positions, which increases the uncertainty of the distance between them.
Speed accuracy. The reported speed at which a GPS receiver is moving is typically accurate to within 0.1 to 0.2 knots.
SOG not speed through the water. The speed reported by GPS is speed over the ground not speed through the water. These can be significantly different if there is a strong current.
Course Accuracy. The accuracy of the course reported by GPS depends on the speed. Typically, around +/- 2.5 degrees at 2 knots, improving to +/- 0.5 degrees at 10 knots.
COG not heading. The course reported by a GPS is the course over the ground (COG) not the heading of the boat (the direction in which it is pointing) unless the tracking information is supplemented by compass information from an external system.
GPS datum. The geographic coordinate system used to reference the position. The most common standard in Europe is WGS 84. Positions can vary by hundreds of metres when referred to different datums. When comparing positions reported by two different GPS units (e.g. on different boats), or a reported GPS position to a waypoint or charted position, it is important to ensure that all positions are referenced to the same datum.
Many GPS enabled devices can record a track to show a boat’s course over time. This is usually derived from a log of timestamped positions and converted by software into a graphical display on a PC or dedicated chart plotter.
Fleet tracking systems. Increasingly used at regattas, employ GPS devices on the boat or on the sailor to transmit timestamped position and course information to a master system that can record and display the track of each boat throughout a race. Often marks of the course are fitted with a tracker. Depending on the data transmission method used, boat positions may be logged at intervals as short as 1 or 2 seconds or as long as hours between log points.
AIS tracks. Many racing yachts are equipped with AIS transponders that transmit GPS derived information over VHF. Websites exist that display this, essentially public, information in the form of vessel positions and/or tracks. Class B AIS systems, the type most often installed on racing boats, attempt to transmit a position message every 3 minutes when the vessel speed is less than 2 knots or every 30 seconds for higher speeds. However, in busy areas, class B transmissions are not guaranteed and it is not unusual for a class B track to contain segments with very long track intervals.
Important issues to consider for tracks are the logging interval (note 2) and the degree of smoothing carried out by software.
Changes of course and/or speed significantly shorter in duration than the logging interval are not usually identifiable. A high frequency non-smoothed track display will give more information about a boats course than a low frequency highly smoothed track.
The boat ‘icon’ displayed on the screen is not normally to scale with the boat and is usually displayed with the icon centred on the reported position. This can be particularly misleading for larger boats (see below) and in mark rounding situations where the tracking display shows the ‘zone’ at the mark.
When comparing multiple tracks, the track points will not usually be synchronized in time. When comparing the distance between tracks, in addition to considering the relative accuracy, consider the possible errors due to smoothing of the tracks between track points.
Tracks constructed from data transmitted by wireless networks (e.g. fleet tracking and AIS data) will often have missing data points. For display purposes, tracking systems may interpolate between the actual data points to provide a smooth track, increasing the uncertainty in a boat’s position or course.
In some cases, it may be necessary to view the raw log data.
Location of antenna. The position reported by a GPS is the position of the antenna. It is important to know the location of the antenna within the boat. This is particularly important for larger boats, or when the GPS is carried by a crew member rather than fixed to the boat.
When evaluating GPS evidence, it is important to carefully consider the inherent uncertainty of the GPS data in relation to the facts required to decide the case, the nature of the incident and the size and speed of the boats involved.
Incidents between boats. GPS tracks and positions are not usually sufficiently accurate on their own to decide close quarters incidents such as establishing/breaking an overlap, keeping clear, brief or frequent changes of course or tack, etc.
Marginal OCS calls. GPS tracks and positions are not usually sufficiently accurate on their own to decide marginal OCS calls. Specific important factors to consider are the position of the antenna on the boat, the method the boats used to identify the position of the line (e.g. manually ‘pinging’ the ends of the line, GPS trackers on the marks, etc. which introduce their own regions of uncertainty) and the ability to synchronize the boat’s track position to the starting signal
Sailing the course. GPS data and tracks will often be sufficient to show that a boat has approached close to a given mark of the course, has crossed the starting or finishing line or even that a mark has been rounded or passed on the correct side when the margin of rounding or passing is large.
Especially on smaller boats, GPS data is not usually sufficiently accurate on its own to show that a boat has passed close on either side of a mark.
Note than when a GPS track is being compared with the charted position of a mark, a major uncertainty is the actual position of the mark. Strong wind and/or tide, may move a mark significantly from its charted position. Waypoints, however, are, by definition, fixed points.
Penalty turns. Depending on the frequency of logging and the availability of heading data, a GPS track may not contain sufficient data to show that a penalty turn (or turns) has been made or completed correctly.
Determination of a boat’s tack. Particularly when sailing downwind, it may not be possible to determine on which tack a boat is sailing. For example, a boat on port tack sailing by the lee may be shown on starboard tack.
(Required for each boat if using data or tracks from boats’ own instruments)