Many maritime users today believe that GPS will always be available. This is simply not the case.
By Alan Grant, Paul Williams, George Shaw, Michelle De Voy, and Nick Ward, The General Lighthouse Authorities of the United Kingdom and Ireland
GNSS availability can be affected in many ways, through events or conditions that affect constellation health, the signal-in-space, or the reception of that signal. The primary means of positioning, navigation, and timing (PNT) employed in maritime applications, whether stand-alone or augmented, has well known vulnerabilities.
This article considers three specific threats and reports on how they may affect maritime safety: GNSS interference and jamming; constellation availability; and space weather events.
Interference and Jamming
There has been a marked increase in both the use and the availability of GPS jamming equipment in recent years. The implications are that jamming units may find their way onto ferries and around ports or harbors where they will interfere with the many systems utilizing GPS, thus affecting maritime safety.
GPS jamming units are widely available on the Internet, with current models already capable of jamming L1, L2, and L5 signals. While we report here on the jamming of GPS, all GNSS constellations would be affected in a similar manner.
To understand the effects of jamming and GPS service denial on maritime safety, the General Lighthouse Authorities of the United Kingdom and Ireland (GLAs) conducted two jamming trials, in collaboration with the UK Government’s Ministry of Defence (MOD), who provided and operated the GPS jamming units. For the safety of all GPS users, and in line with MOD regulations for the peacetime use of GPS jamming units, notice was given to all national bodies. In addition, the GLAs issued notices to mariners explaining that aids to navigation (AtoNs) using GPS in the vicinity of the trials location would be unreliable during the jamming periods.
Flamborough Head. The first jamming trial was conducted off the East coast of the United Kingdom near Flamborough Head. The aim of this trial was to understand the effect GPS jamming may have on ship-borne and shore-based equipment, GLA AtoNs, and also on the crew.
The Northern Lighthouse Board vessel Pole Star steamed between two known waypoints, through an area affected by the jamming signal. Data was recorded from two typical marine-grade GPS receivers installed on the vessel, along with an eLoran receiver that provided the true position throughout the trial.
The results identified three distinct states (Table 1) corresponding to the manner in which GPS-fed equipment responded to jamming conditions. When the jamming signal was sufficiently strong to prevent reception of GPS signals, a large number of alarms sounded on the bridge almost simultaneously, providing a potentially disconcerting and confusing environment for the mariner. However, the effect that represented the highest risk was the provision of erroneous data from some GPS receivers.
Figure 1 compares an erroneous position reported by a typical marine-grade GPS receiver with the vessel’s true location. In this figure, the light blue line shows the path taken between the two waypoints.
The colors of the plotted position points indicate vessel speed. The three states described in Table 1 can be seen.
State 1 is observed at either end of the passage where the solid blue line occurs; this is where the jamming signal strength is much lower than the GPS signal strength, and the GPS-fed systems are operating normally.
As the vessel approached the main lobe of the jamming signal, indicated by the red lines, it reached an area where the jamming signal was comparable with the received GPS signals, leading to State 2. During this state, erroneous data can be observed with the receiver reporting the vessel on land traveling at high speed.
As the vessel entered the main lobe of the jamming signal, State 3 was observed: the GPS signals were swamped by the jamming signal, and the receivers failed to provide an output. Then, as the vessel continued the passage out of the jamming area, one can observe the change in states as the ratios of jamming to GPS satellite signals decrease, and GPS is reacquired.
In the worst case, the GPS receiver reported a position some 22 kilometers away from the true location. The GPS receiver nevertheless declared the position valid. This position was made worse by the fact it was reported inland at a speed of more than 100 knots, while the trial vessel steamed steadily at 10 knots. Depending on how the resulting GPS positioning data is used, it could feasibly result in vessels changing course, through the use of an autopilot, and it could also affect the vessel’s reported position to the outside world. This would then not only affect the vessel’s situational awareness but also the situational awareness of vessels in the vicinity.
The errors observed in Figure 1 were also seen on the vessel equipment fed by the onboard GPS receivers. Erroneous positions were observed on the vessel’s electronic chart display and information system (ECDIS), on the automatic identification system (AIS) positions (where loss of position prevents the unit from calculating a range or bearing to nearby vessels, greatly affecting the crew’s situational awareness), and on the vessel’s radar (Figure 2).
The results observed during these trials gave an important example of what can happen to onboard equipment as well as the impact it can have on the mariner during periods of GPS jamming and service denial. It is clear that GPS denial caused by jamming can not only prevent PNT information from being calculated, it can also result in erroneous data being presented to the mariner.
Newcastle. A second series of demonstrations was conducted off Newcastle-upon-Tyne, on the North East coast of England, to communicate the importance of resilient PNT to a selected audience. The audience included a number of key decision-makers from European and UK governments, maritime industry, mariners, and other aids-to-navigation service providers. The demonstrations took place onboard the Trinity House vessel Galatea.
For this trial, the GPS jamming unit was installed onboard the Galatea and configured to jam GPS within a small
area around the vessel. As before, two typical marine-grade GPS receivers were installed along with an eLoran receiver; for this trial, a modified electronic chart display was also installed and altered to enable two position inputs to be displayed at the same time, to compare the reported GPS and eLoran positions in real-time.
Throughout the demonstrations differential Loran (dLoran) corrections were provided using a transportable reference station installed on the shore at South Shields, to mitigate the impact of temporal variations on the eLoran position. Differential-Loran corrections were generated by the reference station and sent to the GLAs’ eLoran transmitter in Cumbria for inclusion in the eLoran Loran Data Channel (LDC) broadcast. The eLoran receiver on the vessel received the broadcast and was able to extract and apply the corrections in order to obtain an eLoran position within 9 meters (95 percent).
One demonstration scenario showed the sudden effect of a strong jamming signal, designed to simulate a jamming unit being brought onto a ferry or other vessel. This took the vessel’s equipment directly to State 3: complete loss of GPS information with a large number of alarms sounding on the bridge. The loss of GPS data prevented the Galatea’s AIS and VHF units, among other systems, from operating correctly.
Before the second scenario was conducted, the jamming unit was stopped, and all of the GPS receivers integrated into the bridge equipment were allowed to reacquire satellites and fully recover. The second scenario was designed to reflect a vessel steaming towards a jamming source. The field strength of the jamming signal was slowly increased until State 2 was observed, with erroneous and often hazardously misleading information reported.
As with the Flamborough trials, erroneous GPS positions reporting unfeasibly high speeds were observed as shown in the OPENING Figure. However, significantly more subtle errors were seen: errors where the vessel’s reported position differed only very slightly from the true location and wandered around slowly. These subtle changes produce believable positions but hazardously misleading information (HMI). While the overall result of GPS jamming on Galatea was consistent with that observed on Pole Star, there were a few marked exceptions.
The effect of GPS jamming can be seen (Figure 3) on the erroneous positions reported by the trial vessel NLB Pole Star (center right) and also on the vessel Dutch Progress (top left).
The ECDIS onboard the Pole Star reported erroneous positions and ultimately failed with the complete denial of GPS. However the ECDIS on the Galatea continued to track the vessel’s position due to an additional position feed from the vessel’s gyro, making it more resilient to jamming, but only in the short term until the gyro requires re-calibration. This is carried out with its built-in GPS receiver! In addition, the AIS transceiver on the Pole Star reported the vessel’s position erroneously due to jamming, and this was observed at shore-based traffic monitoring stations.
During the demonstrations on the Galatea, the AIS transceiver did not provide any erroneous position information, as can be seen in Figure 4. These differences show that the impact of GPS jamming will be different for each vessel and depends on the model, installation, and configuration of the onboard systems.
Effect of Jamming on Safe Navigation
To navigate safely, the mariner needs reliable, clear and trusted information about where the ship is and what is going on around it, so that any threat can be located and identified. While consideration is often given to threats such as areas of shallow water, obstacles, or other vessels; consideration is not generally given to the loss of positional information, timing, or situational awareness.
Loss of GPS-derived PNT information at sea results in the loss of the vessel’s ECDIS, AIS, GPS, and DGPS receivers, preventing the mariner from being able to position the ship and others around it through what are nowadays regarded as the normal means. In addition, the systems one would normally expect to be independent from GPS, and as such available for use in GPS-denied conditions, are also affected; namely the vessel’s radar and gyro-compass.
The radar takes a GPS input to provide a “North-up” setting and the gyro-compass uses GPS to stabilize drift error. Under GPS-denial conditions these units also enter an alarm state and should not therefore be used in that condition.
Clearly GPS jamming can significantly affect the safety of mariners. From these trials it can be seen that the extent of the impact varies from vessel to vessel depending on the equipment installed and the configuration selected.
Satellite Constellation. From the users’ perspective, GNSS availability is the percentage of time they can receive usable data from sufficient satellites in order to calculate their position. The reduction in the number of available satellites in the constellation will have a direct impact on the system’s availability.
A report from the U.S. Government Accountability Office (GAO) in 2009 predicted “significant challenges in sustaining and upgrading widely used [GPS] capabilities” due to delays in launching modernized GPS satellites. The GAO reported the probability of maintaining a constellation of at least 24 usable GPS satellites could reduce to 80 percent or less by 2011, and not return to 95 percent probability consistently until 2015. This could lead to reduced satellite numbers causing coverage “windows” where less than four satellites could be observed and as such reduced GPS availability.
A later report by the GAO indicates that the probability of maintaining a constellation of at least 24 operational GPS satellites is now expected to be 95 percent for the foreseeable future. This figure is based on the current launch schedule, and although the U.S. Air Force Space Command (AFSPC) has provided reassurances, the satellite launch program has in recent years experienced delays, and therefore the risk of reduced satellite availability still remains.
Following the 2009 report, the GLAs commissioned a study to investigate the impact a reduced GPS constellation would have on users in their waters. This study was conducted by the GNSS Research and Applications Centre of Excellence (GRACE) and was split into two parts. The first part was to analyze the impact theoretically and found that with a 21-satellite constellation, GPS coverage “windows” (for example, fewer than four satellites) could last for several minutes and cover a large proportion of the UK and Ireland (Figure 5). This can cause reduced GPS availability and therefore increased likelihood of position errors affecting maritime safety.
The second part of the study investigated the effects further through a dynamic simulation, investigating the effects should a vessel be position
ed off the coast of Belfast during one of the coverage windows. For this a marine-grade GPS receiver and a simulator were used to observe the effects. The study found that the number of available satellites fell below four for several minutes and the reported position data from the receiver appeared to freeze for up to 10 minutes.
If a mariner was traveling at a speed of 35 knots when the position input froze, his reported position would be in error by 10 kilometers from an outage lasting 10 minutes. These outages are significant, and mariners need to be informed of such risks to GPS (and GNSS in the future) before they occur, so they are prepared for any disruptions.
Space Weather. Space-weather events are a particular concern to GNSS availability due to their random nature. It is known that GNSS signals are delayed proportionally to the number of free ions as they propagate through the Earth’s atmosphere enroute to the receiver. The amount of ions in the ionosphere, the total electron count (TEC), is dependant on time of day, latitude, and solar activity, among other factors. During high solar activity, the number of ions in the atmosphere is much higher than at any other time. The greater the signal delay, the larger the errors are in the satellite’s pseudo range and hence the position error can be significant.
Variation in electron density along the GNSS signal path causes signal refraction that produces phase scintillation, introducing group delay that may cause large errors in the pseudorange measurement. Diffraction of the signal wave front induces amplitude scintillation — variations in signal amplitude — with strong fades possible, leading to a GNSS receiver losing signal tracking, and at worst the GNSS navigation solution may be lost.
Solar activity is cyclical, peaking at a maximum approximately every 11 years, during which periods GNSS performance can be severely degraded, especially at equatorial, auroral and polar latitudes. The next solar maximum is predicted to occur during 2013.
During quiescent periods of solar activity, ionospheric effects on GNSS can be managed such that the residual errors caused by the ionosphere do not generally pose a problem to maritime navigation performance.
The GLAs’ DGPS corrections significantly reduce common mode errors, including the effects of the ionosphere. However, at the peak of the solar cycle with high levels of sunspot activity, solar storms and flares, the application of ionospheric models and differential corrections may be less effective, and this could increase position errors and introduce an integrity risk to maritime navigation.
Maritime navigation systems and services that rely on GNSS are at greatest risk of disruption from the ionosphere during the period from 2011 to 2015. Even during a quiet solar maximum, the occurrence of individual sun spots could produce significant effects for discrete events. The effects vary with latitude, season, and time of day (the hours soon after sunset being most affected).
Space weather events have the potential to affect GNSS availability, either by affecting the performance of the satellites themselves or by preventing signal reception.
Mitigation. In general, a number of steps can be taken to help reduce the impact of these threats:
- Increase awareness of GNSS vulnerabilities.
- Detect incidents and warn the mariner when they occur.
- Prevent incidents from occurring, where possible, through legislation and enforcement.
- Reduce as much as possible the effects of incidents when they occur, through the hardening of GNSS technology.
- Have alternative means of PNT, independent of GNSS.
Understanding that these threats exist and knowing what disruption they may cause is the first step to mitigating their effects, but this does not stop them happening. Being able to identify that an event is occurring and that the data being received from the receiver may not be true is an important part of mitigating the effects.
For jamming issues specifically, the use of GPS jamming units is illegal in the UK and Ireland; however, preventing them from being used is very difficult to achieve. Jamming units are small and easily hidden; however, port-side security and vessel security procedures should prevent jamming units from being used in these locations.
It is a different case, however, to prevent a jamming unit from being used at a coastal location or headland due to the remote nature of these areas.
Mitigating the effect of jamming can be achieved in a number of ways: by limiting the effect within the receiver by using anti-jamming techniques, or by hardening GNSS receivers. Ultimately the best mitigating activity is to not rely on GNSS PNT once the integrity of the data has been compromised.
For space weather events or cases of reduced satellite numbers, there is very little action the mariner can take to remedy the problem or stop it happening. The mitigating action here is one of awareness — information forewarning the mariner that such a condition is imminent, for example.
Monitoring and detection networks can assist in providing such notifications and real-time information on GNSS problems. The need for such a network across the UK and Ireland is the subject of a different GLA publication, but the GLAs support the discussion on a body to monitor GNSS performance and to take the lead in the dissemination of key information.
For periods where GNSS availability has been affected by mutual interference, jamming, space weather events or constellation issues, the best mitigating action is to use PNT information from a second source, one with dissimilar failure modes.
Mariners need to be prepared for GNSS failures and have access to PNT information through dissimilar systems. In addition, procedures covering what to do in the case of GNSS unavailability should also be provided and rehearsed. It is with this view that the GLAs firmly promote the use of all available means of navigation.
All three threats to GNSS availability reviewed here could affect maritime safety. The two trials observed presentation to the mariner of erroneous data, some of which could be considered hazardously misleading, along with the degradation of crews’ situational awareness. The main effects observed were:
- The presentation of random errors leading to hazardously misleading information that could, depending on installation, cause a vessel to move off course.
- The presentation of erroneous and potentially misleading data to other vessels and shore-based infrastructure.
- The sheer number of alarms on the bridge of the vessel could be disconcerting and distracting for the mariner.
- The loss of GPS-fed systems, which can create an unfamiliar bridge situation and remove safety-critical systems from operation.
- A large number of bridge systems are integrated with GPS and enter an alarm state during periods of GPS outage.
The loss of GPS or a lack of integrity in the reported information leads to an unfamiliar situation on the bridge.
The crews of the Pole Star and the Galatea were expecting to lose GPS, were well-trained, and had primed other systems so they could navigate safely. In real life, there would be no advance notice, and the impact on the crew would be more severe.
The impact of low satellite numbers, as predicted in the 2008 GAO report, could produce poor constellation availability and a loss of PNT information for a considerable period of time. This could result in the same outcome as observed in the GPS jamming trials when entering State 3, where many systems on the bridge failed and entered an alarm condition.
Space weather events are difficult to predict both in terms of when they may occur and their severity. Events could affe
ct satellite positions, their operation, and the reception of their signals by the user, and are clearly a threat.
The GLAs strongly support the need for a resilient PNT solution, one that could continue to provide reliable information during such threats for the safety and benefit of all mariners.
This article is based on a paper given at the Institute of Navigation’s 2011 International Technical Meeting.
Alan Grant is a principal engineer for the Research and Radionavigation Directorate of the GLAs of the UK and Ireland, technical lead and project manager for all GNSS projects there. He has a Ph.D. from the University of Wales.
Paul Williams is a principal development engineer with the Directorate and currently technical lead of the GLAs’ eLoran Work Programme. He has a Ph.D. in electronic engineering from the University of Wales.
George Shaw is an engineer at the Directorate and holds a master’s degree in mathematics from the University of Cambridge.
Michelle De Voy is a development engineer for the Directorate, with an MSc in oceanography from the University of Southampton and an MSc in satellite positioning from the University of Nottingham.
Nick Ward is research director of the General Lighthouse Authorities of the UK and Ireland, with responsibility for strategy and planning of research and development.