GNSS Anti-CRPA System: Carrier Phase Anti-Interference Technology That Keeps Precision Alive

Carrier phase tracking is what gives GNSS receivers centimeter-level accuracy. It is also the first thing to die when interference shows up. The carrier loop has a bandwidth of 10 to 20 Hz — narrow enough to reject noise, but wide enough to let a strong jammer pull it off lock in milliseconds. A CRPA with adaptive nulling protects the signal before it reaches the tracking loops, but residual interference still gets through. That is where carrier phase anti-interference technology inside the anti-CRPA system becomes the difference between a working receiver and a bricked one.

Why Carrier Phase Is Both the Prize and the Vulnerability

The carrier phase measurement tracks the fractional wavelength of the GNSS signal. One cycle at GPS L1 is roughly 19 centimeters. By tracking the phase to a fraction of a cycle — typically 1 to 2 millimeters — the receiver achieves millimeter-level precision in differential mode. This is what makes GNSS useful for surveying, autonomous vehicle lane keeping, and precision agriculture.

But that precision comes at a cost. The carrier tracking loop operates at a very narrow bandwidth. A 15 Hz bandwidth rejects most thermal noise, but it cannot reject a jammer that is 40 or 50 dB above the signal. The loop dynamics simply cannot pull the phase estimate back to the true value fast enough. The phase error grows, the loop loses lock, and the receiver falls back to code tracking — which gives you meter-level accuracy instead of centimeter-level.

A CRPA suppresses the jammer spatially, but suppression is never perfect. A 7-element array might give you 50 dB of null depth against a strong CW jammer. If the jammer is 80 dB above the signal, you still have 30 dB of jammer energy at the correlator input. That is enough to wreck a carrier phase loop. The anti-interference technology built around the carrier loop is what handles that residual threat.

How Residual Interference Attacks Carrier Phase

Narrowband Jammer Leakage Through the Spatial Filter

The most common residual threat is narrowband CW jammer leakage. The CRPA places a null toward the jammer, but the null has finite width. A perfectly synthesized null would be infinitely deep at one angle and zero everywhere else. Real nulls have a beamwidth — typically 5 to 15 degrees depending on array size and element spacing.

If the jammer sits exactly at the null center, suppression is maximum. But if the jammer is slightly off-center — even by a degree or two — the suppression drops by 10 to 20 dB. That leaked energy sits right on top of the carrier frequency. The carrier loop sees it as in-band interference and cannot filter it out without also filtering out the satellite signal.

This is the core problem. The carrier loop and the jammer occupy the same frequency band. You cannot separate them with a frequency filter. You can only reduce the jammer power enough that the loop can track through it — and that requires a combination of spatial nulling and carrier-domain anti-interference techniques working together.

Pulsed Jammer Energy During Blanking Gaps

Pulse blanking kills the jammer during its on periods, but it does nothing during the off periods. If the blanker is not perfectly synchronized — and it never is — there are gaps where pulsed jammer energy slips through. Even a 1 microsecond gap lets enough energy through to disturb a 15 Hz carrier loop.

The carrier loop integrates phase over time. A single microsecond of jammer energy creates a phase jump that takes tens of milliseconds to settle. During that settling time, the phase measurement is corrupted. If the pulses come faster than the loop can settle, the phase output becomes a random walk — useless for any precision application.

Carrier Domain Anti-Interference Techniques

Phase-Locked Loop Bandwidth Adaptation

The simplest and most effective carrier-domain defense is dynamic loop bandwidth control. Under clean conditions, the carrier loop runs at a narrow bandwidth — 10 to 15 Hz — for maximum noise rejection and phase precision. When interference is detected, the bandwidth widens to 30 to 50 Hz.

A wider bandwidth lets the loop track through higher-power interference because the loop can correct phase errors faster. The tradeoff is noise — a 50 Hz bandwidth lets in 3 to 4 times more thermal noise than a 15 Hz bandwidth. But in a jammed environment, noise is not your problem. The jammer is. Widening the bandwidth buys you survival at the cost of precision.

The adaptation must be fast. The loop bandwidth should increase within a few milliseconds of jammer detection and return to narrowband once the jammer is suppressed. This requires a real-time interference monitor that feeds the loop filter controller. The monitor watches the carrier-to-noise ratio and the phase error variance. When either exceeds a threshold, the bandwidth widens automatically.

This technique is used in virtually every military GNSS receiver and is increasingly appearing in automotive and survey-grade equipment. It is not elegant, but it works — and it costs almost nothing in computation.

Frequency Lock Loop as a Carrier Phase Shield

A Frequency Lock Loop (FLL) tracks the Doppler frequency of the carrier without relying on phase continuity. When the carrier loop loses lock due to interference, the FLL keeps the frequency estimate alive. Once the interference drops below a threshold, the carrier loop reacquires lock using the FLL's frequency estimate as a starting point.

In an anti-CRPA system, the FLL runs in parallel with the PLL at all times. During interference events, the FLL output feeds the navigation filter while the PLL is suspended. The position solution degrades slightly because FLL-based measurements are noisier than PLL-based ones, but the solution does not jump or diverge.

The handoff from FLL back to PLL must be smooth. A hard switch causes a phase discontinuity that corrupts the carrier phase history. The best implementations use a gradual handoff — blending FLL and PLL outputs over a 100 to 200 millisecond window until the PLL reacquires and the blend factor shifts entirely to PLL.

Cross-Correlation Carrier Smoothing

Carrier smoothing uses the code measurement to clean up the carrier phase. The code measurement is noisy but unambiguous. The carrier measurement is precise but ambiguous by integer cycles. By combining them, you get a precise and unambiguous range estimate.

Under interference, the code measurement degrades but does not fail as catastrophically as the carrier measurement. Cross-correlation carrier smoothing exploits this by weighting the code measurement more heavily when the carrier loop is stressed. The smoothed output is less precise than pure carrier phase, but it is continuous and does not cycle slip.

This technique is particularly valuable in multi-jammer environments where the carrier loop is constantly under attack. The smoothing filter acts as a low-pass filter on the phase output, rejecting the high-frequency phase jumps caused by residual jammer energy. The effective bandwidth of the smoothed output can be as low as 1 to 2 Hz, which rejects almost any residual interference that leaks through the CRPA.

Advanced Carrier Recovery Methods

Decision-Directed Carrier Tracking

When the signal-to-noise ratio drops below the point where a conventional PLL can lock, decision-directed tracking takes over. Instead of using a clean reference signal, the receiver generates its own local replica based on the best available navigation data and uses that replica as the reference for phase comparison.

The loop is now tracking against its own estimate rather than the true signal. This sounds dangerous, and it is — if the estimate is wrong, the loop locks onto the wrong phase and the position error grows. But under heavy interference, having a wrong but stable phase estimate is better than having no phase estimate at all.

Decision-directed tracking is used as a last resort. It activates when the carrier-to-noise ratio falls below 20 dB-Hz and stays active until the CRPA suppresses the jammer enough for conventional PLL tracking to resume. The navigation filter must be told when decision-directed mode is active so it can inflate the measurement noise covariance accordingly.

Vector Tracking Across Multiple Satellites

Conventional receivers track each satellite independently. Vector tracking ties all satellite tracking loops together into a single navigation-aided filter. The filter uses the platform dynamics and the known satellite geometry to predict what each carrier phase should be. The actual measurements are compared against the predictions, and the loops are pulled toward the predicted values.

Under interference, vector tracking is dramatically more robust than independent tracking. If one satellite's carrier loop loses lock, the filter keeps predicting its phase based on platform motion and the other satellites' measurements. The lost satellite reacquires faster because the filter gives it a good initial estimate.

In a multi-jammer environment, vector tracking is not optional. Independent tracking loses satellites one by one as jammers sweep across the sky. Vector tracking loses them more slowly because the filter holds the solution together even when individual loops fail. The computational cost is higher — a vector tracking filter for 12 satellites runs a 24-state Kalman filter at 50 Hz — but modern embedded processors handle it without breaking a sweat.

The Interaction Between CRPA Nulling and Carrier Protection

Null Depth Versus Loop Bandwidth Trade-Off

There is a direct trade-off between how deep the CRPA null is and how much the carrier loop bandwidth must widen. A 60 dB null lets the carrier loop run at 15 Hz. A 40 dB null forces it to 30 Hz. A 30 dB null requires 50 Hz or wider. The carrier anti-interference techniques are not a replacement for good nulling — they are a complement to it.

System designers use this trade-off to set performance requirements. If the application demands centimeter-level accuracy at all times, the CRPA must deliver 50 dB or more of null depth so the carrier loop can stay narrowband. If meter-level accuracy is acceptable during jamming, a 30 dB null with aggressive bandwidth widening is sufficient.

This is why carrier phase anti-interference technology cannot be designed in isolation from the CRPA. The two systems must be co-optimized. A great carrier loop with a weak CRPA is useless. A great CRPA with a fragile carrier loop wastes the spatial filtering gain.

Calibration Impact on Carrier Phase Integrity

Element calibration affects carrier phase more than it affects code tracking. A 2-degree phase mismatch across a 7-element array creates a null pointing error that degrades carrier phase coherence across the array. The carrier measurements from different elements no longer align, and the combined carrier phase output has increased noise.

For carrier phase applications, calibration tolerance is tighter. Gain mismatch must stay below 0.5 dB. Phase mismatch must stay below 1 degree. This is harder to achieve in the field, especially across temperature extremes. A CRPA that meets code-tracking calibration specs may still fail carrier phase specs under thermal stress.

Built-in self-calibration that runs continuously is the only reliable way to maintain carrier phase integrity in a deployed anti-CRPA system. Periodic calibration is not enough — thermal drift happens continuously, and the carrier loop notices every degree of phase error.