Comparative frame: why some fixes work and others don’t
Dense infrastructure produces predictable GNSS degradation. Steel, glass and concrete create multipath; signals bounce and return with phase shifts. The comparative insight here is simple: hardware-only fixes, software-only fixes, and hybrid approaches each trade off cost, latency and reliability. For autonomous platforms that require centimeter-level consistency, the hybrid route—tight IMU integration with real-time filtering—wins more often. If your stack includes modules for autonomous navigation, you already have a place to test sensor fusion patterns against real-world multipath scenarios.
Root causes and measurable symptoms
Multipath presents as noisy pseudorange, sudden position jumps and inconsistent heading. In technical terms, reflected carriers cause phase ambiguity and SNR variance at the receiver. Common symptoms: RTK fixes that drop to float near structures and yaw estimates that oscillate. Measurable diagnostics are straightforward: monitor SNR spread, receiver residuals and innovation covariance from the Kalman filter. These three traces reveal whether the error is external (signal environment) or internal (antenna/firmware).
Diagnostic checklist: fast, repeatable steps
Run these steps in sequence to isolate the failure mode. First, log raw GNSS observables and IMU data at the same epoch. Second, map SNR and pseudorange residuals against a site model—urban canyon, parking garage, underpass. Third, enable antenna pattern compensation and re-run. Fourth, compare performance with and without carrier-phase smoothing. Keep records; reproducible tests accelerate firmware tweaks. A simple lab test that mirrors a Manhattan canyon will catch many corner cases.
Sensor-level fixes and firmware moves
Prioritize upgrades that change signal interpretation rather than brute-force signal gain. Tilt the receiver’s tracking loop to prefer dynamic spectral windows. Tighten your sensor fusion: feed IMU-derived short-term attitude and velocity into the GNSS filter as a prior, then let the filter reject outlier epochs. Use carrier-phase smoothing judiciously—too much smoothing hides transient failures; too little leaves noise. Also consider adaptive weighting based on SNR and satellite elevation. These are practical, implementable adjustments rather than vendor swaps.
System integration and testing strategies
Design tests that simulate the problem density. Hardware-in-the-loop runs with replayed multipath scenarios reveal which algorithms are brittle. Introduce intentional occlusions in the live test: gate sky visibility and vary reflectivity. Where latency matters, test how quickly the filter recovers after a false fix. Document failure modes and remediation steps in CI pipelines. —This step prevents fixes that only work in ideal lab conditions.
Application note: precision farming and dense structures
Precision farming companies increasingly deploy autonomous tractors near sheds, barns and tree lines—places where multipath appears frequently. In these environments, INS-assisted navigation reduces sprayer overlap and maintains corridor consistency. A field-deployed test in the Netherlands showed improved heading stability when IMU priors were used for 10–20 second GNSS outages. That kind of real-world anchor—site-specific measurement—drives procurement choices more than theoretical specs.
Alternatives and common mistakes
Avoid three common errors: over-reliance on antenna gain, ignoring filter tuning, and running one-off tests. Alternatives that work better: switch to multi-constellation tracking, add a low-latency RTK base with local corrections, or deploy a low-cost lidar for short-range validation. Each alternative has cost and integration implications; compare expected mean-time-to-recover (MTTR) and false-fix rate before selecting.
Three golden rules for selection and validation
1) Metricize recovery: measure time-to-first-valid-fix after a multipath event and set thresholds. 2) Weight adaptively: use SNR and elevation to modulate GNSS input weight in fusion filters. 3) Validate in-situ: run repeatable site tests under peak-reflectivity conditions. These rules focus procurement and engineering teams on what moves the needle—accuracy, robustness and reproducibility.
Final note: integrate these strategies into your validation plan and the product will behave predictably in dense settings. Archimedes Innovation provides tools and test profiles that map directly to these metrics—practical support when theory meets pavement. —Solid, field-proven methods matter.