Analysis of GNSS Interference Events Based on TRITON GNSS-R Measurements

Jyh-Ching Juang; Yung-Fu Tsai

doi:10.33012/navi.720

Abstract

Although the issue of global navigation satellite system (GNSS) interference has been addressed in the community of satellite navigation, the extent of GNSS interference in the past couple of years has raised serious concerns for air and marine transportation. This paper assesses GNSS interference based on observations of the TRITON GNSS reflectometry (GNSS-R) payload. The TRITON GNSS-R payload contains a navigation unit and a science unit that are designed to receive direct line-of-sight and scattered GNSS signals, respectively. In the presence of radio-frequency interference, these two units experience different phenomena, including navigation disruptions, spoofed localization, and degradation of GNSS-R performance. This paper describes the effects of interference, analyzes the observation data, and elucidates the interference characteristics. Analyses of GNSS interference based on space data are believed to be instrumental for providing information concerning the frequency, location, and severity of interference and for developing interference-resistant techniques.

Keywords

1 INTRODUCTION

Jamming and spoofing are concerns for global navigation satellite system (GNSS)-based positioning and timing services, as GNSSs are potentially vulnerable owing to the weak signals received from satellites and the use of publicly available signal formats. To provide resilience against interference, many different interference detection and mitigation methods have been developed and utilized in GNSS receivers. Indeed, because satellite navigation is conceived as a dual-use system, means for interference resistance have been well studied for both military and civilian user equipment, as discussed by Spilker and Natali (1996), Dovis (2015), and Gao et al. (2016). However, in the past couple of years, the issue of GNSS interference has become a greater concern, as multiple incidents have been reported, particularly in geographical areas surrounding conflict zones; see websites such as GPSJAM (2022), Flightradar24 (2024), and SkAI Data Service (2024). In response to the growing severity and safety risk of GNSS jamming and spoofing incidents, the European Aviation Safety Agency (EASA) (2022) has issued a Safety Information Bulletin addressing the degradation of communication/navigation/surveillance service due to GNSS outages. GNSS interference can be considered as a threat or attack on the positioning, navigation, and timing (PNT) system. To better mitigate interference and protect the operation from being disrupted, it is imperative to have an understanding of the scope of GNSS interference that is characterized in terms of time, location, occurrence, severity, and type. Indeed, an assessment of GNSS interference is essential for the operation of satellite-based navigation systems and for the design of interference-resistant GNSS receivers. In this paper, in-orbit GNSS measurement data from the TRITON GNSS reflectometry (GNSS-R) payload are analyzed to elucidate the scope of GNSS interference to pave a way for the design of appropriate countermeasures.

Methods for the detection, identification, classification, and localization of GNSS interference have been investigated in the community; see, for example, reports by Bartl et al. (2017), Qiao et al. (2023), and Dempster and Cetin (2016). In the work of Liu et al. (2023), Chen et al. (2023), and Liu et al. (2024), data from the Automatic Dependent Surveillance-Broadcast (ADS-B) system (ADS-B Exchange, 2024) were utilized for the detection of interference events and the localization of interference sources. More recently, ADS-B data have been analyzed by Felux et al. (2021) to assess jamming from the perspective of aviation. Observations of radio-frequency interference in aircraft have been reported by Querol et al. (2018) and Novak et al. (2018). GNSS interference can also be monitored by using low-Earth orbit (LEO) satellites. The detection and localization of interference from space have unique advantages in terms of coverage and observation repetition. In the past, spaced-based detection of GNSS interference has been achieved at the International Space Station (Murrian et al., 2021), CYGNSS satellites (Chew et al., 2023), and FORMOSAT-7/COSMIC satellites (Roberts et al., 2022). Moreover, the mean and variance of the ratio of range-compensated carrier to interference-plus-noise has been assessed by Murrian et al. (2021) for the detection of interference. Likewise, Chew et al. (2023) accounted for the mean noise to render a characterization of hot spots. In the work by Roberts et al. (2022), the efect of interference on the voltage signal-to-noise ratio was analyzed. These investigations have stimulated interest in the deployment of observation constellations, as discussed by Ellis et al. (2022) and Dutsch et al. (2024), as well as the design of more advanced instruments, as reported by Gamba et al. (2024).

Recall that spoofing is a malicious attempt to manipulate and overlay GNSS signals to mislead the receiver. Techniques for spoofing detection can be found in the works by Psiaki and Humphreys (2016) and Meng et al. (2022). A spoofing event over the Black Sea has also been reported by Liu et al. (2024). More recently, the vulnerability to and mitigation against GPS spoofing were documented by OPSGROUP (2024).

The TRITON satellite was launched into a sun-synchronous LEO at 600 km in October 2023. The primary payload of TRITON is a GNSS-R receiver, which is designed to receive and process scattered GNSS signals from the surface of the Earth for remote sensing applications such as probing surface roughness, retrieving ocean wind speed, and detecting sea ice (Jin et al., 2014; Zavorotny et al., 2014). As the operation of the GNSS-R payload involves the reception and processing of GNSS signals, any GNSS interference may potentially affect the operation of the payload. This paper attempts to analyze some anomalous behaviors of the payload and to provide information for the monitoring and classification of GNSS interference.

In addition to evaluating more up-to-date and globally distributed observation data on GNSS interference, the main contributions of this paper include the following:

Existing LEO-based observations of GNSS interference typically address the degradation of the signal-to-noise ratio and the variation of noise statistics. Because of the increased complexity of interference, the navigation unit of the TRITON is routinely jammed, resulting in a failure to operate. A method of combining GNSS navigation and orbit propagation is used to provide a complete determination of satellite orbit for a better understanding of the interference. This approach allows the collection of GNSS outage data in order to statistically characterize the repetition and location of GNSS interference.
This paper reports a spoofing event from its onset to demise. Under this spoofing event, the GNSS navigation unit is forced to switch among different operations. The cause and consequence of the spoofing are analyzed in terms of carrier-to-noise ratios with respect to legitimate GNSS satellites and the spoofer, leading to an enhanced understanding of the features of the spoofer. Notably, although spoofing incidents have been reported based on ground and aircraft observations, to our knowledge, this is the first reported spoofing incident in space.
The GNSS-R payload is able to provide a delay-Doppler map (DDM), which reveals the power of the scattered signals at specified time delay and Doppler frequency bins. In the presence of radio-frequency interference, the DDM reveals some interference fringe patterns. A model is developed to elucidate the interference fringe pattern and to relate this phenomenon to the presence of continuous-wave interference.
This paper analyzes the effects of interference by using data from the navigation unit and science unit of the TRITON GNSS-R payload. In the presence of interference, the navigation unit renders information at the navigation data level while the science unit reveals information at the post-correlation level. The former addresses features such as outages of navigation service, loss of lock of navigation satellites, deviation of position, and degradation of the signal-to-noise ratio. In contrast, the latter accounts for statistics (mean, variance, and kurtosis) of noise, distortion of the correlation function, and the pattern of the DDM. By combining this information, the interference situation can be better assessed and characterized.

It is believed that the analysis of GNSS interference based on space data will pave the way for a better understanding of interference behavior and will shed light on the development of mitigation schemes for improving the resilience and accuracy of PNT systems.

The remainder of this paper is organized as follows. In Section 2, the TRITON GNSS-R mission and payload are introduced. The operations of the navigation unit and science unit are described. Methods for the analysis of GNSS interference are then delineated. In Section 3, GNSS interference events, as observed and recorded by the navigation unit, are discussed. By combining GNSS outage and signal-to-noise ratio data, the time and location of the interference or jamming incidents are assessed. Furthermore, a spoofing event is analyzed in detail. In Section 4, some unique signatures of GNSS interference, as observed by the science unit, are discussed. A model is proposed to explain the observed interference fringe pattern. Further, simulation and template matching are used to verify the model. Finally, concluding remarks are provided in Section 5.

2 BACKGROUND

In this section, the TRITON GNSS-R mission is discussed, the operating principle of the GNSS-R receiver payload is explained, the generation of the DDM is formulated, and the methods for analyzing GNSS interference are provided.

2.1 TRITON GNSS-R Mission

with a weight of 241 kg and dimensions of 100 × 120 × 125 cm³, the TRITON satellite is a microsatellite designed to perform a GNSS-R mission and technology demonstration mission (Taiwan Space Agency (TASA), 2024). The former employs a GNSS-R receiver payload to collect GNSS signals reflected from the Earth’s surface for research on air–sea interactions and typhoon intensity prediction whereas the latter aims to verify/validate some domestically developed satellite subsystems/modules in space to raise the technology readiness level.

Figure 1 depicts a block diagram of the GNSS-R receiver payload. A zenith antenna is used to receive direct line-of-sight signals from GNSS satellites, and a nadir antenna is used to receive GNSS signals scattered from the surface of the Earth. The received signals are, respectively, amplified by low-noise amplifiers (LNAs) and then passed to a radio-frequency front end (RFFE) for filtering, additional amplification, down-conversion, and digitization. During normal operation, the digitized signals received from the right-handed circularly polarized zenith antenna are processed by the navigation unit to perform acquisition, tracking, decoding, and positioning tasks and to determine the position, velocity, and clock bias of the receiver. In contrast, the science unit performs a series of correlations on the signals received from the left-handed circularly polarized nadir antenna to generate the DDM, which reveals the scattered signal power for a set of code phase delays and Doppler frequencies.

FIGURE 1

Block diagram of the TRITON GNSS-R payload

The outputs of the navigation unit are sent to the science unit for preparation of the telemetry (TM) data. The TM data contain position, velocity, and timing solutions of the navigation unit, the state of each tracked satellite, and other status data. The TM data are passed to the TRITON satellite and then down-linked to the ground station for monitoring of the receiver. The satellite operator can also up-link commands to the satellite for control of the receiver. Commands from the satellite to the payload are referred to as telecommand (TC) data, which contain the initialization, operation assignment, parameter setting, and data request of the payload. In addition, the DDM data from the science unit are sent to the satellite bus through the data channel and down-linked to the ground station. Thereafter, the data from the navigation unit and science unit can be obtained from the ground data center for processing and analysis. Currently, some TRITON data products can be obtained from the website TACC (2024).

2.2 Delay Doppler Map

The science unit prepares the DDM observation. To this end, with respect to the incoming signals from the nadir antenna, a series of coherent integration and noncoherent accumulation at specified code phase delays and Doppler frequencies are performed. Let T_s be the sampling interval and u(t_n) = u(nT_s) be the digitized signal from the nadir antenna at time t_n = nT_s. The coherent integration evaluates the correlation between the incoming signal and the local replica over an integration interval:

$y_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}; m) = \frac{1}{N} \sum_{n = (m - 1) N}^{m N - 1} u (t_{n}) g_{i} (t_{n} + {\hat{τ}}_{i, l}) \exp (- j 2 π {\hat{v}}_{i, k} t_{n})$ 1

where $j = \sqrt{- 1}$ and N is the number of coherent accumulations. The local replica in Equation (1) at instant t_n is given by $g_{i} (t_{n} + {\hat{τ}}_{i, l}) \exp (- j 2 π {\hat{ν}}_{i, k} t_{n})$ , where g_i (⋅) is the spreading code, ${\hat{τ}}_{i, l}$ is the estimate of the code phase delay of the l-th code phase bin, and ${\hat{ν}}_{i, k}$ is the estimate of the Doppler frequency of the k-th Doppler bin with respect to the i-th satellite. Note that a set of code phase delays and Doppler frequencies are prepared in an attempt to evaluate the correlation result at different code phase delays and Doppler frequencies. In the TRITON GNSS-R payload, 128 code phase delays and 64 Doppler frequencies are processed. The value of N in Equation (1) is related to the integration time by T_int = N T_s. Typically, T_int is selected to be an integer multiple of the period of the spreading code. In the TRITON GNSS-R payload, the coherent integration time is 1 ms, which is the same as the period of the Global Positioning System (GPS) L1 coarse/acquisition (C/A) code. The sampling rate of the TRITON GNSS-R payload is 16.368 MHz, and as a result, N = 16368. Because the scattered signal is weak, noncoherent accumulations are further employed to enhance the signal and to account for the uncertainties of navigation data transition. The noncoherent accumulation computes the average sum of squares, leading to the following:

$z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}; p) = \frac{1}{M} \sum_{m = (p - 1) M}^{p M - 1} {| y_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}; m) |}^{2}$ 2

where M is the number of noncoherent accumulations. Thus, the index p of the noncoherent accumulation corresponds to the time epoch at p M N T_s. In the TRITON GNSS-R payload, the nominal number of noncoherent accumulations M is 1000, that is, the DDM is updated every second. The science unit of the TRITON GNSS-R payload is designed to provide up to 8 DDMs with respect to different satellites. Details about the requirements, specifications, design, and implementation of the TRITON GNSS-R payload can be found in the reports by Juang et al. (2016) and Juang (2023).

The above coherent/noncoherent integration resembles a match filter (Betz, 2015; Kaplan & Hegarty, 2017). In PNT applications, the input signal can be expressed as $u (t_{n}) = \sum_{i} d_{i} (t_{n}) g_{i} (t_{n} + τ_{i}) \exp (j 2 π ν_{i} t_{n} + j θ_{i})$ , where d_i(t_n) denotes navigation data, τ_i is the actual code phase delay, v_i is the actual Doppler frequency, and θ_i is the phase of the received signal of the i-th satellite. Then, the expected value of $z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}; p)$ in Equation (2) is approximated as the squared ambiguity function (Betz, 2015; Zavorotny & Voronovich, 2000):

$E {z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}; p)} \approx R^{2} (τ_{i} - {\hat{τ}}_{i, l}) {sinc}^{2} ((ν_{i} - {\hat{ν}}_{i, k}) T_{int})$ 3

where R(⋅) is the auto-correlation function of the spreading code. Under the binary phase shift keying modulation, this function is approximated by a triangle function:

$R (\tilde{τ}) = {\begin{matrix} 0, & | \tilde{τ} | > T_{c} \\ 1 - \frac{| \tilde{τ} |}{T_{c}}, & | \tilde{τ} | \leq T_{c} \end{matrix}$

where T_c is the chip interval. For the GPS L1 C/A code, the chip interval is given by 1/1023 ms. The sinc function in Equation (3) is defined as follows:

$sinc (x) = {\begin{matrix} 1, & x = 0 \\ \frac{\sin (π x)}{π x}, & x \neq 0 \end{matrix}$

Note that with the use of the mutually uncorrelated property of spreading codes, the coherent integration result with respect to the i-th satellite is not significantly affected by signals from other satellites. Further, through the square operation in forming $z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}; p)$ , the dependency on navigation data and carrier phase is removed. Thus, the expected value in Equation (3) primarily depends on the code phase error and Doppler frequency error.

Therefore, with respect to a direct line-of-sight or coherently reflected GNSS signal with the same spreading code, if ${\hat{τ}}_{i, l}$ is close to τ_i in the sense that $| τ_{i} - {\hat{τ}}_{i, l} | < T_{c}$ and ${\hat{ν}}_{i, k}$ is close to v_i in the sense that $| ν_{i} - {\hat{ν}}_{i, k} | < 1 / T_{int}$ , then the coherent/noncoherent integration leads to a peak in the DDM at the bin $({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k})$ . In contrast, if the incoming signal is dominated by noise or the estimates are beyond the above bounds, the resulting DDM is noise-like. During the processing of reflected signals for GNSS-R, because the scattered signals are relatively weak and the power may spread over some delay and Doppler bins, the science unit specifies some combinations of delays and Doppler frequencies to perform a series of coherent and noncoherent accumulations to enhance signal quality. Indeed, with respect to GNSS signals scattered from the surface of the Earth, the resulting DDM is a combined effect of the propagation path, antenna gain, bistatic cross-section coefficient of the reflecting surface, and the above ambiguity function as discussed by Zavorotny and Voronovich (2000) and Voronovich and Zavorotny (2018). This situation typically leads to a horse-shoe-shaped DDM. Some snapshots of the DDM from the TRITON GNSS-R payload are depicted in Figure 2. In the figure, the left panel depicts the DDM obtained when GNSS signals scattered from the ocean surface are processed, and the horse-shoe shape is clearly visible. In the plot, the vertical axis is the code phase delay, the horizontal axis is the Doppler frequency, and the value of the DDM or the power of the received signal can be found from the bar chart below the plot. The amplitude, waveform, or even the image of the DDM is used for remote sensing applications; see works by Zavorotny et al. (2014) and Jin et al. (2014). The middle panel of Figure 2 presents a DDM obtained when a coherently reflected GNSS signal is received in which a peak is clearly observed. For comparison, the right panel of the figure shows the DDM obtained when the scattered signals are not sufficiently strong and the DDM is primarily affected by noise. We note that the signal power received from coherent reflection (middle panel) is stronger than that from diffuse scattering (left panel). These DDMs are considered as good DDMs in remote sensing data analysis. In comparison, the noise-contaminated DDM (right panel) is considered as a bad DDM; such DDMs are often disregarded. Methods for the retrieval of remote sensing parameters based on good DDMs have been reported by Garrison et al. (2002), Jin et al. (2014), Zavorotny et al. (2014), and Ruf and Balasubramaniam (2019).

FIGURE 2

Snapshots of TRITON GNSS-R DDMs

The left panel presents a DDM obtained from diffuse scattering, the middle panel corresponds to coherent reflection, and the right panel displays a DDM obtained in the presence of severe noise. The bar chart beneath the plots depict the count through noncoherent accumulation. The peak of the coherently reflected signal is stronger than that from diffuse scattering. In the noise-dominated DDM, the count is the smallest.

2.3 Analysis of Interference

As mentioned above, the two antennas of the TRITON GNSS-R payload are designed to receive direct line-of-sight and scattered GNSS signals, respectively. In practice, the antennas may also pick up unintentional and intentional radio-frequency interference. In the presence of GNSS interference, the operation and output data of the navigation and science units are affected. The signal processing chain of the GNSS-R payload may perceive interference at three stages:

Post-correlation stage: The correlation results are distorted by interference. For the TRITON GNSS-R payload, this means that the observed DDMs are subject to variations.
Signal detection stage: The carrier-to-noise ratio is affected in the presence of interference.
Navigation stage: The position report may be disrupted or misled if the navigation unit is jammed or spoofed.

A challenge in interference detection and characterization is the lack of ground truth. Indeed, once a GNSS receiver is jammed, the receiver may fail to provide positioning results. Further, a spoofed receiver simply reports the erroneous positioning data, and it is challenging to obtain the true position if there are no additional references. Fortunately, the situation can be somewhat relaxed for a satellite. One can obtain satisfactory position- and signal-level results based on analyses of orbital mechanics and communication links. Consequently, determining the position and assessing the carrier-to-noise ratio can facilitate the analysis of GNSS interference.

For cases in which the GNSS receiver is jammed, the position report is not available from the navigation unit. The position of the receiver, however, can still be estimated. Indeed, in-orbit objects are continuously tracked by major space organizations for fight safety and situational awareness. For example, Space Track (2024) tracks space objects and provides orbital element sets to users worldwide. Orbital elements, including designation, epoch, mean motion, inclination, right ascension of the ascending node, eccentricity, argument of perigee, mean anomaly, derivatives of mean motion, and drag coefcients, are listed in the so-called two-line elements (TLEs), which can be accessed from websites such as Space Track (2024) and CelesTrak (2024). By using TLE data, one can compute the position and velocity of a satellite at a specifed epoch via simplifed perturbation models. For LEO satellites, the simplifed general perturbation model SGP4 (Vallado et al., 2006) can be used. As a result, the use of TLE/SGP4 can be regarded as an independent method for estimating the satellite position. In the presence of jamming or spoofing, the position estimated via the TLE/SGP4 approach can serve as a reference. Indeed, the TLE/SGP4 approach can be used to interpolate missing data when the navigation unit is jammed. This approach can also be used to determine the true position when the navigation unit is spoofed.

The positions of GNSS satellites can be obtained in a similar manner. However, for GPS satellites, the U.S. Coast Guard facilitates almanac and precise orbit parameters at the website Navigation Center (2024) for determining the position of GPS satellites. Therefore, in the presence of jamming or spoofing, one can determine the position of the transmitter (GNSS satellite) and receiver (TRITON satellite). As a result, the received carrier power and carrier-to-noise ratio can be evaluated. As reported by Misra and Enge (2006), the received carrier power is given by the following:

$C = \frac{P_{t} G_{t} G_{r}}{L_{a} L_{r}} {(\frac{λ}{4 π d})}^{2}$ 4

where P_t is the GNSS transmitted power, G_t is the gain of the transmitter antenna, G_r is the gain of the receiver antenna, L_a represents the atmospheric loss, L_r is the loss in the receiver, λ is the wavelength, and d is the distance between the transmitter and receiver. The noise power density N₀ can be expressed as follows:

$N_{0} = κ T_{e q}$ 5

where κ is the Boltzmann constant and T_eq is the equivalent temperature. From Equations (4) and (5), the carrier-to-noise-density ratio C/N₀ with respect to each observable satellite at any epoch can be evaluated. Indeed, information on the GPS transmitted power P_t and gain G_t can be obtained from the interface specification (IS-GPS-200N, 2022). Here, the wavelength λ for the GPS L1 signal is known to be 19.05 cm, and the loss L_a is taken as 0.5 dB as suggested by Misra and Enge (2006). The temperature T_eq, gain G_r, and loss L_r are variables related to the navigation unit and are available from the design and test data. Further, the distance d between the navigation unit and the corresponding GPS satellite can be computed based on the positions of the navigation unit and the GNSS satellite. Hence, one can assess the interference by comparing the carrier-to-noise ratios from the measurement and from the aforementioned model.

3 GNSS INTERFERENCE OBSERVED BY THE NAVIGATION UNIT

Orbiting around the Earth at 600 km with a zenith antenna, the TRITON navigation unit is supposedly not vulnerable to terrestrial interference. Yet, ever since its commission, the navigation unit has been routinely affected by GNSS interference. Recall that the data of the navigation unit are recorded, packed, and down-linked to the ground stations as TM data for health monitoring. The TM data are down-linked in packet formats. Each packet contains the report of the navigation unit for a period of time. If the navigation unit fails to perform its function because of, for example, jamming, the data report is subject to gaps. We note that missing data do not necessarily imply that the navigation unit is jammed, as other issues such as data transmission errors or even receiver anomalies may result in data gaps as well. However, by scrutinizing the TM data packets, some analyses on data gaps can still be performed to render statistically meaningful results. Furthermore, the navigation unit may be subject to spoofing. The TRITON navigation unit is an unsophisticated GNSS receiver that does not employ advanced spoofing detection or protection schemes. As a result, in the presence of spoofing, the navigation unit may be misled to render an erroneous position report. In this section, the interference observed and reported by the navigation unit will be discussed and analyzed.

3.1 Jamming

During normal operation, the TM data at each epoch contain sequence number, GPS week/second, satellite position, satellite velocity, satellite attitude, operation information, and other variables related to the status of the navigation unit. The operation information includes the pseudorandom numbers (PRNs), position, velocity, and signal-to-noise ratios of the satellites in the navigation processing. If the navigation unit fails to perform positioning, navigation, and timing functions at certain locations or time points, some data entries in the TM data will be blank even though the sequence number remains continuous. Thus, a lack of data serves as a clue in the analysis of interference.

To illustrate and explain the observations, an example is provided. The light blue curves in the left panel of Figure 3 depict the latitudes and longitudes reported by the navigation unit over a single day (Day 325, 2023). As shown in the figure, the navigation unit provides position solutions most of the time. However, it is observed that some portions are missing in the reported data. Such an event of failing to report position is likely caused by the effect of interference. The figure illustrates the onset of each event by a red dot and the resumption of navigation function by a blue dot. The dots are indexed to represent the event number. The purple curves in the left panel in Figure 3 depict the interpolated result based on TLE/SGP4 processing. It is clear that the missing data are filled, implying that the TLE/SGP4 procedure can provide position estimates in the presence of data gaps. With the reported (healthy) data and interpolated data, one can proceed to analyze the spatial and temporal distributions of GNSS interference.

FIGURE 3

Observation of interference effects and interpolation of data gaps

The left panel depicts the latitudes and longitudes on Day 325, 2023. The light blue curves represent the reported data, the purple curves represent interpolated results with respect to missing data, and the onset and resumption of events are marked by red and blue dots, respectively. The right panel illustrates the interpolated results of the missing data from Day 321 to Day 327, 2023.

The above method was applied to missing data in the week from Day 321 to Day 327, 2023, and the interpolated results are depicted in the right panel of Figure 3. It is observed that GNSS outages are encountered in the Mediterranean region, eastern Europe, Middle East, Baltic Sea, Arctic area, and even the South China Sea. These results bear some similarity to the results derived from data collected by aircraft and ADS-B data reported by GPSJAM (2022) and Flightradar24 (2024). We note that the ADS-B data collected by aircraft can be geographically restricted, as aircraft may not be permitted to fly into certain zones. In comparison, satellites such as TRITON are not limited, and the observations can cover the no-fly zone of aircraft, as depicted in Figure 3, making the overall observation data more complete.

In addition to observing data outages and interpolating missing data, it is desirable to perform an in-depth analysis to elucidate the cause and effect of these phenomena. Recall that the navigation unit provides the signal-to-noise ratios of tracked satellites; thus, these data are analyzed below to further assess the situation. Figure 4 depicts the carrier-to-noise-density ratios (C/N₀) of different GPS satellites, as measured by the navigation unit, for Event 1 shown in the left panel of Figure 3. For clarity, the horizontal axis is the relative time, with the onset of the GNSS outage occurring at 600 s. The navigation unit ceases to provide a position solution from 600 s to 1112 s, as the number of tracked satellites during this period is less than 4. Indeed, as shown in the figure, only three satellites (PRNs 7, 9, and 20) are tracked during this period. Figure 5 presents sky plots of GPS satellites at the onset (600 s) and resumption (1112 s) of the event, respectively. It is noted that the above three GPS satellites are at high elevation, making them less vulnerable to terrestrial interference.

FIGURE 4

Carrier-to-noise-density ratios for the first interference event on Day 325, 2023 Between 600 and 1112 s (dashed lines), only three satellites are tracked, and the navigation function is compromised.

FIGURE 5

Sky plots of GPS satellites

The left panel corresponds to the onset of the interference event (600 s), and the right panel corresponds to when the navigation function has resumed (1112 s). Note that the satellites with PRNs 7, 9, and 20 are at high elevation.

The blue curves in Figure 6 depict the carrier-to-noise-density ratios of admissible satellites that are reported by the navigation unit. The figure indeed reproduces some of the C/N₀ data shown in Figure 4. Referring to the sky plots in Figure 5, the admissible satellites are satellites that could have facilitated signals for positioning during which the period under consideration. The red curves in Figure 6 represent the computed C/N₀ based on Equations (4) and (5). In the plots, both the reported C/N₀ and computed C/N₀ are subject to gaps, resulting from the inability of the navigation unit to track the satellite and the fact that the GPS satellite is below the zero-degree elevation of the navigation unit, respectively. For clarity, the period during the navigation unit does not provide a position report (between 600 and 1112 s) is shaded in yellow. It is found that even though satellites with PRNs 7, 9, and 20 are tracked, the reported C/N₀ is less than the computed C/N₀, implying that the tracking performance is degraded. More importantly, the degradation of C/N₀ occurs earlier (at approximately 200 s) than the onset (at 600 s). Indeed, as shown by the curves for PRNs 4, 7, 9, 11, 16, and 20, the simultaneous decrease in the reported C/N₀ and the eventual loss of lock trigger the disruption of the navigation function. Thus, the interference takes place at an earlier time or location. After 1112 s, the interference has weakened, authentic GPS satellites (PRNs 4, 6, 11, and 30) are rapidly acquired/tracked, and the navigation function resumes. Notably, the reported C/N₀ and computed C/N₀ show close agreement when the interference effect is absent. In addition, the C/N₀ values of the three tracked signals (PRNs 7, 9, and 20) during the period when the navigation function is compromised appear to be correlated. More precisely, at approximately 756 s and 949 s, there are two simultaneous drops in C/N₀, as highlighted by vertical dashed lines in the corresponding plots, implying that the interference may be from a common source.

FIGURE 6

Reported and computed carrier-to-noise-density ratios for the first interference event on Day 325, 2023

The reported results are shown in blue, and the computed results based on a link budget analysis are shown in red. In the absence of interference, the results are similar. When the navigation unit experiences interference, discrepancies can be observed.

The above results clearly show that by analyzing the data gaps for the GNSS outage and applying some justifications to the signal-to-noise ratio, it is possible to assess the temporal and spatial distribution of the GNSS interference by processing satellite observation data.

3.2 Spoofing

The navigation unit of the TRITON GNSS-R payload is also subject to spoofing. In this subsection, a spoofing incident is analyzed and discussed to enable an understanding of spoofing behavior in space.

The left panel of Figure 7 depicts position reports of the navigation unit for Day 84 (March 24) of 2024, which reveal the presence of spoofing. In the figure, the blue curves depict the reported latitude φ, longitude λ, and altitude h of the navigation unit over a period of 1000 s. The horizontal axis is the relative time, which is offset from the GPS time by 26,100 s. It is noted that in some regions (shaded in yellow), the position reports are missing, owing to the fact that the navigation unit is not able to track and decode a sufficient number of signals. The region shaded in red corresponds to the duration when the navigation unit is spoofed. As depicted, the navigation unit is operational until 387 s, at which time the navigation unit experiences interference. At 488 s, the navigation unit appears to be functional, as position solutions are being provided. The reported solutions, however, are erroneous, implying that the receiver is spoofed. This spoofing incident lasts until 626 s, at which point the position reports are lost again because of a weakening of the spoofing signals. The navigation unit then reacquires and retracks authentic GNSS signals and resumes its normal operation at 668 s. During the period from 387 s to 488 s (shaded red region), it is obvious that the receiver is spoofed because the position reports in the period remain the same. More precisely, the position is at latitude 33.8183°, longitude 35.4908°, and altitude 225.5 m, which corresponds to the Beirut airport (Lloyd’s List, 2024). Further, the velocity components, as depicted in the right panel of Figure 7, indicate that the receiver is misled to be stationary when it is spoofed.

FIGURE 7

Reported latitude, longitude, and altitude (left panel) and velocity components (right panel) during the spoofing incident

The period during which the navigation unit is spoofed is shaded in red, and the period during which the navigation data are missing is shaded in yellow.

Figure 8 depicts the ground track of the TRITON satellite as it proceeds from normal operation to the spoofed situation, followed by a recovery to normal operation. The spoofing occurs when the satellite is over the Caspian sea. The portion of the ground track corresponding to the transition and spoofing situation is reconstructed by the aforementioned TLE/SGP4 procedure. The carrier-to-noise-density ratios are shown as a function of time and satellite PRN in Figure 9. It is noted that before 387 s, the C/N₀ values of the observable satellites gradually decrease, eventually causing the navigation unit to fail to lock legitimate GNSS satellites. The receiver is then tricked to acquire and track spoofed signals and render erroneous position reports. The right panel of Figure 9 shows the C/N₀ values of the admissible satellites during the period when the navigation unit is spoofed. Some additional observations of the spoofing event are as follows:

FIGURE 8

Position reports of the spoofing event

The ground track of the satellite runs from north to south, and the reported position is spoofed to the Beirut airport when the satellite is over the Caspian sea.

C/N0 values with respect to satellites/spoofers in the spoofing event When the navigation unit is spoofed, the reported C/N0 values of different PRNs are highly correlated (right panel).

FIGURE 9

C/N₀ values with respect to satellites/spoofers in the spoofing event

When the navigation unit is spoofed, the reported C/N₀ values of different PRNs are highly correlated (right panel).

When the receiver is spoofed, the C/N₀ values of the spoofed signals with different PRNs are highly correlated, as shown in the right panel of Figure 9. This result implies that the spoofing may originate from a single spoofer or from a set of coordinated spoofers.
The PRNs of the spoofed signals and the signals from admissible satellites overlap. The sky plot viewed from the Beirut airport at the onset of the spoofing period (388 s) is depicted in Figure 10. Note that the PRNs of the tracked signals are identical to the PRNs in the sky plot.
The spoofing events are not as persistent as the jamming issues discussed in the previous section. Figure 11 depicts heat maps of interference when the navigation unit fails to report data as well as heat maps of spoofing when the navigation unit reports erroneous data for March 2024. By assessing the time and location for which the navigation unit is compromised and spoofed, respectively, we find that the affected region of spoofing is more localized.
The exact location of the spoofer is not known. However, it is conjectured that the spoofing originates from the Israel region in an attempt to provide air defense against objects from the northeast direction (NPR, 2024). As the spoofing signal is beaming into the sky and of sufficient power, the LEO satellites are spoofed.

FIGURE 10

Sky plots at the beginning of the spoofing incident, as viewed from the spoofed position

FIGURE 11

Heat maps for March 2024, as observed by the TRITON navigation unit

The number of interference events is indicated by color. The left panel shows the interference (jamming) map for when the navigation function is compromised, and the right panel shows the spoofing map.

4 INTERFERENCE OBSERVED BY THE SCIENCE UNIT

The TRITON GNSS-R payload is capable of producing a DDM, which is an array of measurements related to the received power of the scattered signals for different code phase delays and Doppler frequency shifts. Because the scattered GNSS signals are extremely weak, a series of coherent integration and noncoherent accumulations is employed to enhance the signal in the construction of the DDM, as discussed in Section 2. Thus, the science unit that is responsible for generating the DDM is a highly sensitive receiver. Further, the nadir antenna used for receiving the scattered signals has a peak gain of 15 dB. Therefore, a terrestrial interference with sufficient power can be detected by the science unit even though the antenna is left-handed circularly polarized. Moreover, in the presence of interference, because the science unit is an array receiver, some signatures of the interference can be observed. In this section, the detection of interference by the science unit and the resulting DDM data are discussed.

In the presence of interference, the TRITON science unit may yield DDMs that are drastically different from those depicted in Figure 2. An example is displayed in Figure 12, which shows four DDMs for an epoch on Day 330, 2023, when the TRITON is over the Korean peninsula. The DDMs reveal fringe-like patterns along the Doppler frequency axis. All DDMs are subject to similar patterns, yet the patterns of the different PRNs are different. In contrast, the values of the DDM along the code phase axis vary only slightly. It is desirable to explain how these fringe-like patterns form.

FIGURE 12

TRITON GNSS-R DDMs in the presence of interference

Let T_p be the period of the spreading code, that is, g_i(t) = g_i(t + rT_p) for any integer r. The periodic spreading code can be represented in terms of a Fourier series expansion as follows:

$g_{i} (t) = \sum_{q = - \infty}^{\infty} G_{q}^{(i)} \exp (j 2 π q t / T_{p})$ 6

Here, the $G_{q}^{(i)}$ terms are the complex Fourier coefficients, which can be computed as follows:

$G_{q}^{(i)} = \frac{1}{T_{p}} \int_{0}^{T_{p}} g_{i} (t) \exp (- j 2 π q t / T_{p}) d t$ 7

The variable q/T_p in Equation (7) is the frequency of the q-th harmonic, and the fundamental frequency is 1/T_p. For the GPS L1 C/A code, one can represent the spreading code in terms of Fourier coefficients as follows:

$g_{i} (t) = \sum_{q = - Q}^{Q} G_{q}^{(i)} \exp (j 2 π q t / T_{p})$ 8

where Q = 511, as the length of the spreading code is 1023.

If the incoming signal u(t_n) is a continuous wave expressed as u(t_n) = exp(j2πν t_n + jφ) for some frequency v and phase φ, then under the assumption that the coherent integration time is 1 ms (i.e., m = 1), the coherent integration result in Equation (1) is as follows:

$\begin{matrix} y_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}) & = \frac{1}{N} \sum_{n = 0}^{N - 1} \sum_{q = - Q}^{Q} \exp (j φ) G_{q}^{(i)} \exp (j 2 π q {\hat{τ}}_{i, l} / T_{p}) \exp (j 2 π q n T_{s} / T_{p}) \exp (j 2 π (ν - {\hat{ν}}_{i, k}) n T_{s}) \\ = \exp (j φ) \sum_{q = - Q}^{Q} G_{q}^{(i)} \exp (j 2 π q {\hat{τ}}_{i, l} / T_{p}) h_{q} (ν - {\hat{μ}}_{i, k}) \end{matrix}$ 9

where:

$h_{q} (ν - {\hat{ν}}_{i, k}) = \frac{1}{N} \sum_{n = 0}^{N - 1} \exp (j 2 π q n T_{s} / T_{p}) \exp (j 2 π (ν - {\hat{ν}}_{i, k}) n T_{s})$ 10

Note that the magnitude of $y_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k})$ does not depend on φ. Thus, the DDM or noncoherent accumulation result in Equation (2) with M = 1 is as follows:

$z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k}) = {| \sum_{q = - Q}^{Q} G_{q}^{(i)} \exp (j 2 π q {\hat{τ}}_{i, l} / T_{p}) h_{q} (ν - {\hat{μ}}_{i, k}) |}^{2}$ 11

The above result implies that the DDM can be regarded as a weighted sum of the basis functions $h_{q} (ν - {\hat{μ}}_{i, k})$ in which the weighting coefficients are given by $G_{q}^{(i)} \exp (j 2 π q {\hat{τ}}_{i, l} / T_{p})$ . Manipulating $h_{q} (ν - {\hat{ν}}_{i, k})$ in Equation (10) yields the following:

$h_{q} (ν - {\hat{ν}}_{i, k}) = \frac{1}{N} \frac{\sin (N π (ν - {\hat{ν}}_{i, k} + \frac{q}{T_{p}}) T_{s})}{\sin (π (ν - {\hat{ν}}_{i, k} + \frac{q}{T_{p}}) T_{s})} \exp (j (N - 1) π (ν - {\hat{ν}}_{i, k} + \frac{q}{T_{p}}) T_{s})$ 12

The function $h_{q} (ν - {\hat{ν}}_{i, k})$ possesses the following properties:

The peak value of $| h_{q} (ν - {\hat{μ}}_{i, k}) |$ is 1 and is obtained when ${\hat{ν}}_{i, k} = ν - \frac{q}{T_{p}}$ .
The magnitude of $h_{q} (ν - {\hat{μ}}_{i, k})$ is governed by the envelope $\frac{1}{N} \frac{\sin (N π (ν - {\hat{ν}}_{i, k} + \frac{q}{T_{p}}) T_{s})}{\sin (π (ν - {\hat{ν}}_{i, k} + \frac{q}{T_{p}}) T_{s})}$ , which decreases as $ν - {\hat{ν}}_{i, k} - \frac{q}{T_{p}}$ increases. The main lobe of the envelope has a null-to-null bandwidth of $\frac{2}{T_{p}}$ .
The two functions $h_{q_{1}} (ν_{1} - {\hat{μ}}_{i, k})$ and $h_{q_{2}} (ν_{2} - {\hat{μ}}_{i, \bar{k}})$ are the same provided that we have the following:
${\hat{μ}}_{i, \bar{k}} = μ_{i, k} - (ν_{1} - ν_{2}) - \frac{q_{1} - q_{2}}{T_{p}}$

The fringe-like patterns of the DDMs shown in Figure 12 can be explained in terms of the functions $h_{q} (ν - {\hat{ν}}_{i, k})$ , as verified by simulation. The real and imaginary parts of the Fourier series coefficients of GPS PRN 10 C/A code with q from −20 to +20 are depicted in Figure 13. The real parts are symmetric whereas the imaginary parts are antisymmetric, as the spreading code is real. Moreover, the coefficient $G_{0}^{(i)}$ is nearly zero, owing to the randomness or near-zero-mean property of the spreading code. Figure 14 depicts the function $h_{q} (ν - {\hat{ν}}_{i, k})$ versus ${\hat{ν}}_{i, k}$ for v = 0, T_s = 1/(1.023 × 10⁶) s, and T_p = 1.0 ms. As shown in the top panel, the real part is symmetric, and the imaginary part is antisymmetric. The bottom panel shows that the magnitude of the function decreases as the magnitude of ${\hat{ν}}_{i, k}$ increases. For a different q, the function is shifted by q/T_p. Thus, the functions $h_{q} (ν - {\hat{ν}}_{i, k})$ act as gating functions along the frequency axis, and the actual DDM or fringe-like pattern is jointly governed by the weighting coefficients. The simulated DDM in Equation (11) for the above spreading code when the frequency and phase of the incoming signal are both zero, i.e., a constant-amplitude signal, is shown in the left panel of Figure 15. In the simulation, the integration time T_int is selected to be the same as the period of the spreading code T_p, which is 1 ms. The result clearly depicts fringe-like patterns similar to the DDMs shown in Figure 12. The right panel of Figure 15 further depicts the DDM when the input frequency is 5 kHz. It is observed that the DDM $z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k})$ with input frequency v is the same as the DDM $z_{i} ({\hat{τ}}_{i, l}, {\hat{ν}}_{i, k} - ν)$ under a zero-frequency input. In other words, a change in the frequency of the incoming signal results in a shift of the DDM along the frequency axis. The analysis and simulation results shed light on our understanding of the observed DDM. The DDMs in Figure 12 can be attributed to the presence of continuous-wave interference.

FIGURE 13

Fourier coefficients of the GPS PRN 10 C/A code

The function hq(ν−ν^i,k) as a function of ν^i,k when v = 0

FIGURE 14

The function $h_{q} (ν - {\hat{ν}}_{i, k})$ as a function of ${\hat{ν}}_{i, k}$ when v = 0

FIGURE 15

Simulated DDM under a constant-amplitude, zero-frequency input (left) and under a continuous wave with a frequency of 5 kHz (right)

The gray dashed line depicts the input frequency.

The DDM from the science unit is obtained for a certain combination of code phase delays and Doppler frequency shifts. For the TRITON, the measured DDM window size is 128 (code phase delays) × 64 (Doppler frequencies). In contrast, one can simulate the above model for a wide range of code phases and Doppler frequencies. The measured DDM can be matched to the simulated DDM to verify the model and estimate the frequency. To achieve this goal, one can use the template matching technique, which is a common image processing method, as discussed by Gonzalez and Woods (2018) and Brunelli (2009). Given that the measured DDM is susceptible to gain variation and noise, adjustments to the bias and scaling factor are implemented to determine a match. Figure 16 shows the results obtained by matching the measured DDM in Figure 12 to the model. The measured fringe-like patterns are satisfactorily approximated by the models. Therefore, the interference fringe pattern in the DDM can be attributed to the presence of continuous-wave interference.

FIGURE 16

Match between the measured DDM (left) and the model (right) for different PRNs

5 CONCLUSIONS

The presence of GNSS interference has become a major concern in satellite-based PNT service. To effectively mitigate the consequences of GNSS interference and achieve a high level of resilience, it is crucial to have a deep understanding of the scope of the interference phenomena. This paper shows that even at LEO, GNSS interference incidents are real, persistent, and severe in some regions. It is believed that the TRITON GNSS-R data discussed in this paper will facilitate our understanding and complement existing data for enhanced monitoring and characterization of GNSS interference. Specifically, existing GNSS interference data such as that reported by GPSJAM (2022), Flightradar24 (2024), and SkAI Data Service (2024) are collected based on terrestrial and aircraft observations. Observations at LEO may augment out understanding of the scope and significance of GNSS interference. In addition, the repetition of the satellite orbit yields observations with regular temporal and spatial resolution. With the availability of orbital elements, one can still establish the satellite position even when the GNSS navigation function is compromised.

However, the effects of GNSS interference observed at the two units are different. Roughly, the navigation unit observes interference at the navigation data level while the science unit perceives interference at the post-correlation level. It is observed that the TRITON navigation unit, orbiting at 600 km with a zenith antenna, is still vulnerable to GNSS interference. The interference not only affects the signal-to-noise ratio but also causes the navigation unit to lose lock and fail to provide a position solution. By combining outage data, signal-to-noise ratio data, and orbit dynamics computations, we find that the GNSS interference appears to be strong and persistent in the affected areas. In areas subject to GNSS jamming, some spoofing events are occasionally observed. One spoofing event is documented in this paper to provide an understanding of the spoofing and its consequences. At the post-correlation level, interference fringe patterns are observed in the DDM. A model based on Fourier series expansion is proposed to describe the phenomena. This model is verified through simulation and template matching with real data.

This paper provides insights about GNSS interference around the world based on an analysis of TRITON GNSS-R data. This work will be continued for a better understanding and characterization of jamming and spoofing. More exactly, the accumulated data will be analyzed to evaluate the change in the heat map with respect to jamming and spoofing. Subjects such as spoofer localization will be investigated as well. The analysis results are also being used to adjust the operation and data flag of the TRITON GNSS-R payload, as the primary mission objective is to provide good DDMs for the purpose of remote sensing. In the future, assimilation with other observation data will be pursued to provide a more comprehensive characterization of GNSS interference around the world. In addition, the overall processing techniques will be automated for the development of a near-real-time GNSS interference warning system.

HOW TO CITE THIS ARTICLE

Juang, J.-C., & Tsai, Y.-F. (2025). Analysis of GNSS interference events based on TRITON GNSS-R measurements. NAVIGATION, 72(4). https://doi.org/10.33012/navi.720

CONFLICT OF INTEREST

The authors declare no potential conflicts of interest.

ACKNOWLEDGMENTS

This study was supported by the National Science and Technology Council (NSTC) of Taiwan under Grant NSTC 113-2218-E-006-020.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

REFERENCES

↵
ADS-B Exchange. (2024). World’s largest source of unfiltered flight data. https://www.adsbexchange.com/
Google Scholar
↵
Bartl, S., Berglez, P., & Hofmann-Wellenhof, B. (2017). GNSS interference detection, classification and localization using software-defined radio. Proc. of the 2017 European Navigation Conference (ENC), Lausanne, Switzerland. https://doi.org/10.1109/EURONAV.2017.7954205
Google Scholar
↵
Betz, J. W. (2015). Engineering satellite-based navigation and timing: Global navigation satellite systems, signals, and receivers. Wiley-IEEE Press. https://doi.org/10.1002/9781119141167
Google Scholar
↵
Brunelli, R. (2009). Template matching techniques in computer vision: Theory and practice. Wiley. https://doi.org/10.1002/9780470744055
Google Scholar
↵
CelesTrak. (2024). NORAD GP element Sets. https://celestrak.org/NORAD/elements/
Google Scholar
↵
Chen, Y. H., Liu, Z., Blanch, J., Lo, S., & Walter, T. (2023). A RFI testbed for examining GNSS integrity in the various environments. Pro. of the 2023 International Technical Meeting of the Institute of Navigation, Long Beach, CA, 1032–1042. https://doi.org/10.33012/2023.18681
Google Scholar
↵
Chew, C., Roberts, T. M., & Lowe, S. (2023). RFI mapped by spaceborne GNSS-R data. NAVIGATION, 70(4). https://doi.org/10.33012/navi.618
Google Scholar
↵
Dempster, A., & Cetin, E. (2016). Interference localization for satellite navigation systems. Proceedings of the IEEE, 104(6), 1318–1326. https://doi.org/10.1109/JPROC.2016.2530814
Google Scholar
↵
Dovis, F. (2015). GNSS interference threats and countermeasures. Artech House. https://us.artechhouse.com/GNSS-Interference-Threats-Countermeasures-P1710.aspx
Google Scholar
↵
Dutsch, N., Ernest, H., Pany, T., Campello, A. P., Borheck, D., Speidel, J., & Sunay, H. (2024). GNSS interference detection and geolocation from LEO satellites - satellite formation and payload design specific considerations and their impact on the detection sensitivity and geolocation accuracy. Proc. of the 37th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2024), Baltimore, MD, 737–768. https://doi.org/10.33012/2024.19677
Google Scholar
↵
Ellis, P., Irisov, V., Pojani, G., & Breeze, C. (2022). GNSS interference monitoring from LEO. Proc. of the ESA’s Navigation Innovation and Support Program (NAVISP) Element 2 EL2-117. https://navisp.esa.int/uploads/files/documents/NAVISP%20EL2-117%20Presentation%20Final.pdf
Google Scholar
↵
European Aviation Safety Agency (EASA). (2022). Safety information: Global navigation satellite system outage and alterations leading to communication/navigation/surveillance degradation. https://ad.easa.europa.eu/ad/2022-02R3
Google Scholar
↵
Felux, M., Fol, P., Figuet, B., Waltert, M., & Olive, X. (2021). Impacts of global navigation satellite system jamming on aviation. NAVIGATION, 71(3), 673–685. https://doi.org/10.33012/navi.657
Google Scholar
↵
Flightradar24. (2024). GPS jamming map. http://flightradar24.com/data/gps-jamming/
Google Scholar
↵
Gamba, M. T., Polidori, B. D., Minetto, A., Dovis, F., Banfi, E., & Dominici, F. (2024). GNSS radio frequency interference monitoring from LEO satellites: An in-laboratory prototype. Sensors, 24(4), 508. https://doi.org/10.3390/s24020508
Google Scholar PubMed
↵
Gao, G. X., Sgammini, M., Lu, M., & Kubo, N. (2016). Protecting GNSS receivers from jamming and interference. Proceedings of the IEEE, 104(6), 1327–1338. https://doi.org/10.1109/JPROC.2016.2525938
Google Scholar
↵
Garrison, J. L., Komjathy, A., Zavorotny, V. U., & Katzberg, S. J. (2002). Wind speed measurement from forward scattered GPS signals. IEEE Transactions on Geoscience and Remote Sensing, 40(1), 50–65. https://doi.org/10.1109/36.981349
Google Scholar
↵
Gonzalez, R. C., & Woods, R. E. (2018). Digital image processing (4th ed.). Pearson. https://www.pearson.com/en-us/subject-catalog/p/digital-image-processing/P200000003224/9780137848560?srsltid=AfmBOoqSpF3IRoOd3VilGOk_hhg0VmNCBYmpzXaCslCZ_98OmLnxOIaP
Google Scholar
↵
GPSJAM. (2022). GPS/GNSS interference map. http://gpsjam.org/
Google Scholar
↵
IS-GPS-200N. (2022). Interface specification IS-GPS-200, Navstar GPS space segment/navigation user interface. https://www.gps.gov/technical/icwg/IS-GPS-200N.pdf
Google Scholar
↵
Jin, S., Cardellach, E., & Xie, F. (2014). GNSS remote sensing: Theory, methods, and applications. Springer-Verlag. https://doi.org/10.1007/978-94-007-7482-7
Google Scholar
↵
Juang, J. C. (2023). Determination and sensitivity analysis of the specular reflection point in GNSS reflectometry. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 16, 974–982. https://doi.org/10.1109/JSTARS.2022.3233044
Google Scholar
↵
Juang, J. C., Ma, S. H., & Lin, C. T. (2016). Study of GNSS-R techniques for FORMOSAT mission. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 9(10), 4582–4592. https://doi.org/10.1109/JSTARS.2016.2575069
Google Scholar
↵
Kaplan, E. D., & Hegarty, C. J. (Eds.). (2017). Understanding GPS/GNSS Principles and Applications (3rd ed.). Artech House. https://us.artechhouse.com/Understanding-GPSGNSS-Principles-and-Applications-Third-Edition-P1871.aspx
Google Scholar
↵
Liu, Z., Blanch, J., Lo, S., & Walter, T. (2023). Investigation of GPS interference events with refinement on the localization algorithm. In Proceedings of the 2023 International Technical Meeting of the Institute of Navigation, Long Beach, CA, 327–338. https://doi.org/10.33012/2023.18627
Google Scholar
↵
Liu, Z., Lo, S., Blanch, J., Chen, Y. H., Walter, T., & Pullen, S. (2024). GNSS spoofing and jamming in Eastern Europe. Inside GNSS, 19(2), 28–35. https://insidegnss.com/gnss-spoofing-and-jamming-in-eastern-europe/
Google Scholar
↵
Lloyd’s List. (2024). War zone GPS jamming sees more ships show up at airports. https://www.lloydslist.com/LL1148748/War-zone-GPS-jamming-sees-more-ships-show-up-at-airports
Google Scholar
↵
Meng, L., Yang, L., Yang, W., & Zhang, L. (2022). A survey of GNSS spoofing and anti-spoofing technology. Remote Sensing, 14(19), 4826. https://doi.org/10.3390/rs14194826
Google Scholar
↵
Misra, P., & Enge, P. (2006). Global Positioning System: Signals, measurements, and performance. Ganga-Jamuna Press. https://www.navtechgps.com/global-positioning-system-signals-measurements-and-performance-revised-second-edition-paperback/
Google Scholar
↵
Murrian, M. J., Narula, L., Iannucci, P. A., Budzien, S., O’Hanlon, B. W., Psiaki, M. L., & Humphreys, T. E. (2021). First results from three years of GNSS interference monitoring from low earth orbit. NAVIGATION, 68(4), 673–685. https://doi.org/10.1002/navi.449
Google Scholar
↵
Navigation Center. (2024). GPS MANUS, almanacs, OPS advisories, & SOF. https://www.navcen.uscg.gov/gps-nanus-almanacs-opsadvisories-sof
Google Scholar
↵
Novak, A., Havel, K., & Bugaj, M. (2018). Measurement of GNSS signal interference by a flight laboratory. Transportation Research Procedia, 35, 271–278. https://doi.org/10.1016/j.trpro.2018.12.011
Google Scholar
↵
NPR. (2024). Israel fakes GPS Locations to deter attacks, but it also throws off planes and ships. https://www.npr.org/2024/04/22/1245847903/israel-gps-spoofing
Google Scholar
↵
OPSGROUP. (2024). GPS spoofing. https://ops.group/dashboard/wp-content/uploads/2024/09/GPS-Spoofing-Final-Report-OPSGROUP-WG-OG24.pdf
Google Scholar
↵
Psiaki, M. L., & Humphreys, T. E. (2016). GNSS spoofing and detection. Proceedings of the IEEE, 104(6), 1258–1270. https://doi.org/10.1109/JPROC.2016.2526658
Google Scholar
↵
Qiao, J., Lu, Z., Lin, B., Song, J., Xiao, Z., Wang, Z., & Li, B. (2023). A survey of GNSS interference monitoring technologies. Frontiers in Physics, 1133316. https://doi.org/10.3389/fphy.2023.1133316
Google Scholar
↵
Querol, J., Onrubia, R., Pascual, D., Castellvi-Esturi, J., Park, H., & Camps, A. (2018). RFI analysis and mitigation in airborne GNSS-R campaign. Proc. of the 2018 IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2018), Valencia, Spain, 1237–1240. https://doi.org/10.1109/IGARSS.2018.8518827
Google Scholar
↵
Roberts, T. M., Meehan, T. K., Tien, J. Y., & Young, L. E. (2022). Detection and localization of terrestrial L-band RFI with GNSS receivers. IEEE Transactions on Geoscience and Remote Sensing, 60, 1–11. https://doi.org/10.1109/TGRS.2021.3109524
Google Scholar CrossRef
↵
Ruf, C., & Balasubramaniam, R. (2019). Development of CYGNSS geophysical model function for wind speed. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(1), 66–77. https://doi.org/10.1109/JSTARS.2018.2833075
Google Scholar
↵
SkAI Data Service. (2024). Live GPS spoofing and jamming tracker map. https://spoofing.skai-data-services.com/
Google Scholar
↵
Space Track. (2024). SSA sharing & orbital data requests (ODR). https://www.space-track.org/
Google Scholar
↵
Spilker, J. J., & Natali, F. D. (1996). Interference effects and mitigation techniques. In J. J. Spilker, P. Axelrad, B.W. Parkinson, & P. Enge (Eds.), Global Positioning System: Theory and applications, Vol. 1, 717–771. American Institute of Aeronautics and Astronautics. https://doi.org/10.2514/5.9781600866388.0717.0771
Google Scholar
↵
TACC. (2024). TRITON data. https://tacc.cwa.gov.tw/v2/triton_download.html
Google Scholar
↵
Taiwan Space Agency (TASA). (2024). TRITON. https://www.tasa.org.tw/en-US/missions/detail/TRITON
Google Scholar
↵
Vallado, D. A., Crawford, P., Hujsak, R., & Kelso, T. S. (2006). Revisiting spacetrack report #3. Proc. of the AIAA Astrodynamics Specialist Conference, Keystone, CO. https://doi.org/10.2514/6.2006-6753
Google Scholar
↵
Voronovich, A. G., & Zavorotny, V. U. (2018). Bistatic radar equation for signals of opportunity revisited. IEEE Transactions on Geoscience and Remote Sensing, 56(4), 1959–1968. https://doi.org/10.1109/TGRS.2017.2771253
Google Scholar
↵
Zavorotny, V. U., Gleason, S., Cardellach, E., & Camps, A. (2014). Tutorial on remote sensing using GNSS bistatic radar of opportunity. IEEE Geoscience and Remote Sensing Magazine, 2(4), 8–45. https://doi.org/10.1109/MGRS.2014.2374220
Google Scholar
↵
Zavorotny, V. U., & Voronovich, A. G. (2000). Scattering of GPS signals from the ocean with wind remote sensing application. IEEE Transactions on Geoscience and Remote Sensing, 38(1), 951–964. https://doi.org/10.1109/36.841977
Google Scholar

PDF

Previous Next