Log in

Print Email

Home > GPS
Related topics: GPS,Innovation

GPS

Jamming GPS

January 1, 2003 By: Dr. Borje Forssell, Lt. Trond Birger Olsen GPS World

Susceptibility of Some Civil GPS Receivers

The vulnerability of civil GPS receivers has long been known, but it has rarely been taken into account by receiver manufacturers or users. Only some five years ago, when the U.S. Department of Defense started comprehensive activities related to GPS and warfare (navigation warfare or NAVWAR for short), did it become publicly clear that jamming of civil receivers should be taken as a serious issue.

Figure 1 and 2.

The definition of NAVWAR reads: "an environment in which

1. friendly forces maintain their ability to use satellite navigation,
2. satellite navigation is denied to hostile users,
3. there is no effect upon civilian applications."

The civil GPS community got an eye-opener in 1997 as well. First, the Russian company Aviconversias announced in September that it could deliver a commercial GPS/GLONASS jammer capable of blocking civil GPS receivers within a radius of 200 kilometers. Then, military GPS testing in the New York area in December caused a number of GPS receivers in civil aircraft to lose track of GPS signals during approach into Newark International Airport. Thus, it was confirmed that civil receivers were vulnerable to jamming, and at the same time, that jamming equipment was commercially available.

As a consequence of these and other events, several analyses of the vulnerability of GPS-based transport systems have been carried out. One of the most important studies in this field, and - coincidentally - with very good timing, was the so-called Volpe report on the vulnerability of GPS which concluded that, like other radionavigation systems, GPS is vulnerable to jamming, and that jamming of GPS could jeopardize safety and have serious environmental and economic consequences. The report also concluded that increased use of GPS in civil infrastructure makes it an increasingly attractive target for hostile activities by individuals, groups and states. At the same time, the analyses underlined the commercial availability of equipment for jamming purposes.

Figure 3.

Our Research Goals

The purpose of the work described in this article was to investigate how civil GPS C/A-code receivers react to certain types of interfering signals. It should be remembered that, after all, most interference with GPS signals is unintentional, coming from transmitters established for other purposes. For this reason, we investigated certain generic classes of interference.

We wanted to answer the following questions:

1. What signal types are most "efficient" as interferers/jammers?
2. How does a GPS receiver react to different interference power levels?
3. Can receiver responses be theoretically predicted with high reliabilty?

Thus, the most efficient interfering signal is one which needs the lowest power level to make the receiver lose navigation ability and at the same time prevents the receiver from regaining that ability.

Figure 4a and 4b.

Equipment

The interfering signals were generated by a commercial interference generator connected to a GPS signal simulator. The generator output signal, that is, a combined RF signal consisting of simulated GPS signals plus interference, was used to feed several commercially developed GPS receivers.

We used three receiver types of different makes and different levels of sophistication for the investigations: a receiver introduced in 1994 using arrow-correlator technology but no longer on the market, an OEM sensor currently on the market, and a recently introduced receiver for machine control.

Figure 4c.

Scenarios

We designed seven different jamming scenarios including one without jamming. In order to isolate the effects of the jamming signal types, as many variables as possible were kept identical in the different scenarios. These "static" variables include, among others, receiver position, GPS signal level, UTC of the scenario runtime, and the ionospheric and tropospheric models. (These models were in accordance with NATO Standardization Agreement (STANAG) 4294, Issue 1.) Modeling of multipath effects is possible with the simulator system, but was not utilized during the simulations. Only the jamming signal types were changed from scenario to scenario.

The different jamming signals used were:

1. Non-Coherent Continuous Wave (NCW) Frequency: 1575.42 MHz
2. Coherent CW (CCW) Frequency: 1575.42 MHz Swept CW (SCW) Center frequency: 1575.42 MHz
3. Sweep waveform: Triangle Repetition rate: 1 kHz Frequency deviation: 650 kHz
4. Amplitude Modulation (AM) Carrier frequency: 1575.42 MHz Modulation waveform: Sine Modulation frequency: 1 kHz Modulation depth: 50.0 percent
5. Frequency Modulation (FM) Carrier frequency: 1575.42 MHz Modulation waveform: Sine Modulation frequency: 1 kHz Frequency deviation: ±50 kHz
6. Band-limited White Noise Center frequency: 1575.42 MHz Bandwidth: 20 MHz

(See "Signal Modulation - The Basics" sidebar on page 54 for a description of modulation types.)

Parameters. The jamming signal level was varied from 2122 decibels referenced to 1 milliwatt (dBm) to 259 dBm and back to 2122 dBm in 0.5 dB steps in all the jamming scenarios (see Figure 1). The seventh scenario in which no jamming signal was introduced provided a reference for the various jamming scenarios.

Signal Modulation - The basics

In all scenarios, the following parameter values were used:

Latitude N 638 259 6.77450, longitude E 108 239 57.21800, height 106 meters;

Power received at L1 from each GPS satellite: 2130 dBm;

Length of each session: 2 hours, 20 minutes (see, also, Figure 1);

8 dB, JSR , 71 dB (that is, 2122 dBm , J , 259 dBm), where J is the jamming signal power and JSR is the jamming signal to GPS signal power ratio.

Simulations

The number of satellites tracked was one indicator used for assessing receiver vulnerability to jamming, as this parameter is important with regard to position accuracy. In general, the more satellites in view, the better the possibilities for good position accuracy. This is due to the fact that a receiver having many satellites accessible can avoid tracking an excessive number of low-elevation-angle satellites with low signal-to-noise ratio (SNR). (Some low-elevation-satellites must always be tracked for geometrical reasons.)

In tracking mode, it is particularly interesting to observe receiver transitions from four to three satellites (from 3D to 2D navigation, with the receiver height held constant) and transitions from one to zero satellites (that is, from some tracking to no tracking).

Figure 5a and 5b.

When a receiver is in search mode, it is correspondingly interesting to observe transitions in the opposite direction, that is, when the receiver is regaining navigation capability.

The SNR deteriorates in the presence of interfering signals. There is a variation in SNR during a jamming test because of the varying instantaneous interference level, leading to corresponding variations of position accuracy. As an example, Figure 2 shows the SNR values of all tracked channels during a simulation interval (CCW jamming of the narrow-correlator receiver). SNR values are those provided by the receiver log file. The SNR values in this figure and elsewhere in the article are actually the SNR in a 1 Hz noise bandwidth.

Figure 5c.

When a "good" satellite (one with favorable geometry and a healthy SNR) is lost, there is a sharp decline in position accuracy which continues until a replacement satellite has been acquired or the lost satellite has been reacquired (see Figure 3).

Test Results

As we mentioned above, the reference scenario contains no jamming/interfering signals. Figure 4 illustrates such normal performance for the OEM receiver. It is seen that tracking for 3D navigation is achieved a few seconds after start. The corresponding position errors are shown in Figure 4c.

Referring to the discussion of scenarios above, the effects of AM, FM, and noise jamming on the same (that is, OEM) receiver are used as examples. Numerical values are given in Tables 1-3.

Table 1a - 1c.

For the sake of comparison, corresponding diagrams and tables are shown below for the machine-control receiver (Figures 5 a-c and Tables 2 a-c).

Except for differences in specific threshold values, the OEM and machine-control receivers largely show similar behavior. SNR values decline with increasing interfering signal level and rise with decreasing interference. Position errors are maximum just before the loss of lock and just after regaining lock.

However, there is sometimes a considerable difference in error magnitudes between the receivers during such a transition, as the OEM receiver's errors can be many kilometers in spite of the receiver indicating a valid position, whereas the machine-control receiver errors are only a few meters. A peculiar similarity (see Table 3, next page) is that both receivers fail to regain navigation capability within the interval after having been jammed by a swept continuous wave. This is in contrast to the narrow-correlator GPS receiver. (4/3 means loss or regaining of 3D navigation, 1/0 means loss/acquisition of the last/first satellite.)

Table 2a - 2c.

The narrow-correlator receiver has a special feature which disables position updates if more than four channels have SNR values less than 30 dB. The "last valid position" and "first valid position" columns in Table 3 show at what JSR level four or more channels had SNR values of 30 dB or higher after losing/regaining lock.

Discussion and Conclusions

The narrow-correlator receiver does not deliver position data if fewer than four satellites have signal-to-noise ratios exceeding 30 dB. Thus, the last valid position really defines the upper JSR limit for navigation with this receiver and not the loss-of-lock value which is used for the other two receivers.

The narrow-correlator receiver has an anti-jam mode which, however, did not detect the AM signal. This is assumed to explain why it was most vulnerable to the AM signal (see Table 3), in contrast to the other receivers, where AM is listed in fifth place with regard to efficiency of jamming.

Table 3.

The transitions between tracking of three and four satellites and between one or no satellite show that the FM signal is most efficient (that is, requires the smallest JSR) for the narrow-correlator and OEM receivers. For the machine-control receiver, the SCW signal is the most efficient one, although the FM signal is only 0.5 dB worse. All in all, we therefore conclude that all three receivers as a group are most vulnerable to the FM jamming signals. However, we also conclude that the JSR values causing loss of navigation capability are quite different for the three receivers as shown by Table 3.

Plain noise turned out to be the least efficient type of jamming signal as it requires the highest JSR to cause loss of navigation capability for all three receivers. Noise jamming requires 13-15 dB higher JSR values for loss of lock than FM jamming of the narrow-correlator and OEM receivers, whereas the difference in the machine-control receiver case is about 8 dB.