GNSS Monitoring · RF Evidence and Receiver Diagnostics
GNSS Interference Monitoring: Correlating RF Evidence with Receiver Behavior
GNSS interference monitoring combines observations of the RF environment with receiver measurements, status and navigation output. Its value comes from knowing what was observed, where and when—and how strongly the evidence supports a receiver-impact or cause-attribution claim.
A spectrum feature, a falling C/N₀ trace and an interference alarm are different observations. A useful monitoring record preserves all three when available, together with their coverage, timing and configuration.
Separate four conclusions: an RF feature was observed; a detector reported an event; receiver impact was demonstrated; a cause was established. Evidence for one does not automatically prove the others.
Use jamming versus spoofing for threat mechanisms and GNSS Simulation 101 for the wider test-method guide. This article focuses on correlating RF evidence with receiver behavior.
1. Choose complementary evidence streams
Define the bands, signals, antenna ports and receiver functions of interest. Separate a monitoring receiver from the target receiver when they are different devices, and identify which records each produces.
| Evidence | Useful observation | Required qualification |
|---|---|---|
| Spectrum or RF samples | Energy distribution or waveform behavior over the captured band and interval. | Measurement point, frequency coverage, units, gain, processing, timing and overload status. |
| Front-end diagnostics | AGC/gain state, attenuation, clipping or overload indicators where available. | Field definition, direction, scale and automatic/manual mode; not automatically calibrated input power. |
| C/N₀ and tracking | Signal-specific quality changes, lock events and channel availability. | Satellite/signal identity, geometry, eligibility, receiver settings and lost-channel reporting. |
| Raw observations | Pseudorange, carrier-phase and Doppler consistency or discontinuities. | Epochs, validity and clock behavior. Derived observations are not sampled RF. |
| PVT or timing output | Navigation validity, availability, output age and independently measured error. | A suitable truth/reference and application criteria; a valid flag alone does not establish accuracy. |
| Detector and collector status | Alarm state, configuration changes, dropped records and collection faults. | Enabled paths, trigger/clearing rules and logging health; alarm absence is not proof of a clean environment. |
Scroll horizontally to view all columns. Confirm actual fields and their meanings for the selected hardware and firmware.
Compare C/N₀ within compatible signal and geometry conditions. If weaker channels disappear, the mean of the surviving channels can rise even while reception worsens. Keep the eligible population, missing channels and lock losses visible instead of averaging only available values.
Investigate installation effects as well as external emitters. Septentrio’s mosaic hardware guidance discusses interference from nearby digital electronics and the use of spectrum observations during integration. [5] A C/N₀ change alone does not identify the source.
2. Define the RF measurement boundary before comparing traces
Record antenna location and orientation, RF routing, cable/splitter losses, active gain, filtering and antenna-power arrangements. Two receivers connected to different antennas do not necessarily experience the same interference. Even a shared antenna with split paths requires a documented path comparison.
Identify whether each diagnostic observes an early RF stage, a decimated signal or a signal after particular filters. A reduced peak in a filtered view may describe the receiver’s processing rather than the disappearance of an external emission.
- Amplitude: preserve units and reference plane. Do not convert a relative spectrum or AGC code into an absolute DUT input level without an established method.
- Dynamic range: check the monitoring path for overload or clipping. A distorted capture cannot establish an accurate amplitude or spectrum shape.
- Frequency scope: record the observed span and path bandwidth. A clean selected in-band view does not rule out an unobserved blocker or front-end overload mechanism.
- Receiver configuration: retain filter, mitigation, tracking and logging settings and their changes. Compare like configurations or explain the difference.
For power reference planes and path losses, see GNSS RF power calibration . Spectrum and waveform capture still require their own settings, scaling and validity checks.
3. Interpret native spectrum and detector records correctly
As one implementation example, NovAtel OEM7 ITPSDFINAL
reports pre-decimation spectrum power in absolute dBm, with values
dependent on resolution bandwidth (RBW). Its post-decimation and
post-filter views use relative dB.
[1]
Do not label every spectrum as dBm/Hz or subtract incompatible views
to claim a mitigation attenuation.
Preserve the spectral status/configuration, first frequency, RBW and complete sample sequence. OEM7 can split one spectrum across multiple log messages; reconstruct the complete group before treating it as a spectrum. [1] Record incomplete groups explicitly.
The OEM7 ITSPECTRALANALYSIS command configures the view,
update period, FFT size and averaging. Its post-filter mode refers to
the bandpass/notch-filtered spectrum. NovAtel also notes that increasing
spectral processing load can affect receiver idle time.
[2]
Preserve the configured filters, for example through the documented
ITFILTTABLE output.
[4]
OEM7 ITDETECTSTATUS is an asynchronous detection report for
enabled active paths and is intended for change-triggered logging.
Spectrum-analysis entries and statistical-analysis entries have different
field meanings; the latter’s four detection parameters are reserved.
[3]
Silence between reports can mean unchanged state, so retain initial state
and separate collection-health evidence.
Use the DUT’s definitions. These OEM7 examples are not a universal GNSS protocol or an MGR8000 interface specification. Confirm whether an AGC field represents gain, attenuation or raw counts before assigning a direction to its change.
4. Check what the monitor could actually observe
Record actual acquisition windows, band switching, rearm/transfer gaps, averaging and retention behavior. Display refresh, log-delivery rate and RF observation coverage are separate quantities.
Illustrative blind interval: suppose the actual capture windows are [0, 20) ms and [100, 120) ms. A 5 ms burst during [45, 50) ms falls entirely between them. The two clean captures cannot establish the RF conditions during that gap. These arbitrary windows do not describe a particular product.
Tektronix’s guidance on transient EMI measurement explains that stepped wide-span observation can miss events outside the band/time being processed. [6] Assess detection coverage against the required event duration, frequency range and amplitude; do not borrow another instrument’s probability-of-intercept specification.
If waveform preservation is required, specify continuous or triggered acquisition, supported pre-event history, data format, sustained throughput and file rotation. Confirm each capability. Saving observation logs or spectrum snapshots does not automatically preserve the original RF waveform.
5. Align time and carry uncertainty into the conclusion
Keep the RF acquisition interval, detector processing/report time, receiver measurement epoch and complete-record arrival time separate. An asynchronous log header is not automatically an RF onset timestamp, and packet arrival order is not physical event order.
Preserve native timestamps and quality flags alongside a host monotonic clock or acquisition counter where available. Document clock mappings, resets and discontinuities. Verify the actual source of any reference described as independent: a network clock or another receiver can share the same upstream GNSS dependency.
When a reference enters holdover, record its last trusted state and the uncertainty growth. NIST describes holdover capability as dependent on the oscillator and the implementation. [7] A holdover label alone does not establish the timing accuracy required for this comparison.
Worked example: nominal order is not established order
Illustrative interval arithmetic, not measured performance. Assume two defined event timestamps have already been mapped to one clock. Their conservative bounds include the relevant observation-window, event-definition and clock-mapping uncertainties; they are not one-standard-deviation error bars.
| Event | Mapped estimate | Permitted time interval |
|---|---|---|
| RF event | 100.000 s ± 0.080 s | 99.920 to 100.080 s |
| Tracking event | 100.050 s ± 0.100 s | 99.950 to 100.150 s |
Scroll horizontally to view all columns. The bounds are assumptions for this example, not instrument specifications.
Nominal difference: 50 ms.
Conservative interval: 50 ms ± (100 + 80) ms = −130 to +230 ms.
The interval includes zero and both signs, so these records do not establish which event came first. Even resolved ordering would not prove causation. If only report-arrival times are available, state a delivery-time comparison; do not rename it detector latency.
6. Correlate events without turning every anomaly into a diagnosis
Use the same signal identities, comparison windows and declared validity rules across streams. Match event intervals with their uncertainty, not just rows that share a displayed second.
| Evidence combination | Supported statement | Next check |
|---|---|---|
| RF feature; no observed output breach | An RF feature occurred in the observed band/window; the declared output criteria showed no observed breach. | Coverage, uncertainty and sensitivity; do not generalize to every receiver or operating condition. |
| Receiver anomaly; no matching RF record | The receiver changed behavior; an RF cause is not established. | Observation gaps, path mismatch, installation, blockage, clock and collector faults. |
| RF and receiver changes overlap | The evidence is temporally associated within the stated bounds. | Compatible RF paths, alternative explanations and controlled or independent corroboration. |
| Alarm; no obvious spectral peak | The configured detector reported its condition. | Detector type, acquisition coverage and relevant diagnostics; a narrow peak is not required for every event. |
| RF event; no alarm | RF observation and alarm outcome differ. | Enabled bands, thresholds, dwell, sensitivity, processing mode and alarm-delivery integrity. |
| Overload, missing data or poor time quality | The evidence chain has a stated limitation. | Repair or quantify it; leave amplitude, ordering or impact claims not verified where evidence is insufficient. |
Scroll horizontally to view all columns. These are analysis categories, not standardized receiver alarm codes.
Use independent truth for position/time error and review validity and age with the PVT log validation method . A missing record is a collection question until tracking evidence establishes otherwise.
For suspected integration noise, a controlled equipment-state comparison can strengthen attribution if antenna, RF path, geometry and other conditions are documented. State the evidence and remaining alternatives. A local spectrum by itself neither identifies intent nor locates an emitter.
7. Validate the monitoring workflow and preserve the event record
Before relying on unattended monitoring, verify the collection chain with representative baseline and known-event cases. Agree which functions belong to the receiver, instrument, collector or downstream analysis.
- Baseline and benign variation: establish behavior across relevant geometry, antenna and equipment states using the declared eligibility rules.
- Coverage: verify required bands, event durations, capture settings, overload reporting and observation gaps.
- Event delivery: check trigger/clearing behavior, clock mapping and complete alarm delivery, including collector interruption or restart.
- Receiver impact: compare tracking, observations and fresh navigation/timing outputs with the applicable reference and requirements.
- Retention: verify sustained logging, storage limits, file rotation, export and preservation of raw/native records.
For each event, retain a stable ID, configuration/version, RF path, acquisition support, native records, clock mapping, collector health and the analysis-rule version. Preserve pre-event and post-event context when available, including mitigation changes and clearing/recovery evidence.
Report measured coverage and unresolved intervals. Detection-rate or false-alarm claims need a defined, labeled event population and counting rules; an unlabeled field alarm list is insufficient. Keep incomplete captures and unresolved cases in the report instead of silently removing them.
8. Match the required evidence to the monitoring configuration
Start with the GNSS reception and monitoring product family and the reception and analysis solution . Define the monitor, target receiver, RF measurement/capture equipment, references and collector as separate roles.
The MGR8000 is positioned for live GNSS reception, raw observations, signal-quality monitoring and anomaly awareness. Its published RF acquisition/storage capability requires external storage configuration. Confirm the selected capture format, bands, throughput, duration and export path.
Request representative output records and their timestamp definitions. Confirm native diagnostics and alerts separately from any required spectrum analysis, waveform retention, trigger integration or downstream classification. Hardware connectors alone do not establish those functions or their timing performance.
GNSS interference monitoring FAQ
Does a C/N₀ drop prove GNSS jamming?
No. Review the affected signals, antenna/RF path, geometry, front-end status and RF evidence. The quality change is an observation; its cause requires additional evidence.
Is a receiver spectrum always an absolute power measurement?
No. Confirm the view, reference, scaling and bandwidth. Some documented receiver views are relative. A diagnostic spectrum also requires appropriate calibration and validity checks for the claimed measurement.
Does a clean spectrum snapshot rule out short interference?
Only the captured band and interval were observed. Review acquisition windows, retuning and processing gaps before interpreting an absent feature.
Can raw GNSS observations be replayed as the original RF?
Pseudorange, carrier phase and Doppler are derived measurements. Preserving the original waveform requires a suitable RF-sampling and storage arrangement, with any replay compatibility confirmed separately.
Can one monitoring receiver locate the interference source?
Its local observations do not by themselves establish emitter location. Direction finding or localization requires a separately verified spatial measurement arrangement and method.
Define your GNSS interference monitoring evidence
Share the required bands, installation, target receiver and the question you need to answer: RF awareness, alarm review, receiver impact or a defined diagnostic study.
Request a configuration review covering available diagnostics, time references, collection interfaces, any RF acquisition/storage, and the evidence needed for acceptance. Sample records and detailed field lists can follow during the review.
Technical references
Implementation examples explain measurement boundaries; they do not transfer capabilities to another product. The timing and capture-window examples above are illustrative calculations.
- NovAtel OEM7: ITPSDFINAL spectrum output and scaling .
- NovAtel OEM7: ITSPECTRALANALYSIS configuration and processing modes .
- NovAtel OEM7: ITDETECTSTATUS event reporting .
- NovAtel OEM7: ITFILTTABLE filter configuration .
- Septentrio mosaic hardware manual: EMC considerations .
- Tektronix: transient EMI and real-time spectrum analysis .
- NIST TN 2189: GPS dependencies and timing-system holdover .