In our previous article, we explained antenna placement and noise mitigation techniques to physically and electrically maximize GNSS reception performance. In this twelfth installment (FC12), which serves as the final capstone of our navigation design sequence, we will address multi-GNSS (multi-constellation) design and anti-jamming strategies.
Modern UAS do not rely solely on the US-led GPS; they typically utilize "multi-GNSS" as a standard feature, simultaneously tracking multiple satellite networks including GLONASS (Russia), Galileo (Europe), BeiDou (China), and Japan's QZSS (Michibiki). However, design excellence is not achieved by blindly enabling every available satellite. The core of receiver configuration lies in logical, strategic setup, explicitly accounting for refresh rate degradation, CPU processing overhead, and resilience against electronic warfare—such as jamming and spoofing—in high-threat environments.
Why Multi-GNSS Matters and Its Design Trade-offs
The primary advantage of tracking multiple satellite constellations simultaneously is maximizing the number of visible, usable satellites.
In environments where parts of the sky are obstructed (sky obstruction)—such as urban canyons, deep mountain valleys, or during high-agility maneuvers with steep bank angles—expanding the number of concurrent constellations ensures that the vehicle maintains tracking of the minimum 4 satellites required for position and time calculations. This effectively prevents Dilution of Precision (DOP) degradation. Furthermore, implementing multi-band (multi-frequency) positioning across the L1, L2, and L5 bands cancels out ionospheric propagation delays. This yields sub-meter accuracy even in standalone modes and achieves centimeter-level RTK accuracy with exceptionally fast initialization times (Fast Fix).
However, this architecture introduces specific hardware-software trade-offs within the flight controller (FC) ecosystem:
- Refresh Rate Degradation: In many high-performance GNSS receivers (such as the u-blox F9P), expanding the active constellation count heavily taxes the internal processing engine. For example, tracking 4 constellations at full capacity (GPS+GLO+GAL+BDS) can limit the position update rate to a maximum of $5\text{ Hz}$ (5 outputs per second). Restricting the active selection to a optimized subset—such as GPS+GAL+QZSS—restores the refresh rate to $10\text{ Hz}$ or $20\text{ Hz}$, providing higher-frequency feedback to improve EKF responsiveness during high-speed flight.
- Data Bandwidth and Bus Overhead: The volume of data streams (via NMEA or UBX binary protocols) forwarded to the FC over serial buses like UART scales exponentially. This can cause packet drops or unacceptable latency spikes, making strict message filtering mandatory.
Core Characteristics of Major Satellite Constellations
When configuring an FC, engineers reference the following system-specific performance traits to determine which constellations to prioritize:
| Satellite System | Operating Nation | Primary Allocated Frequencies | Architectural Design Factors & Selection Criteria |
|---|---|---|---|
| GPS | United States | L1 ($1575.42\text{ MHz}$) L2 ($1227.60\text{ MHz}$) L5 ($1176.45\text{ MHz}$) |
Offers the highest global baseline stability. Treated as the top-priority foundational layer by default within ArduPilot. |
| GLONASS | Russia | L1 ($1602\text{ MHz} + k \times 562.5\text{ kHz}$) L2 ($1246\text{ MHz} + k \times 437.5\text{ kHz}$) |
Utilizes Frequency Division Multiple Access (FDMA), shifting its carrier waves away from other networks. This enhances jamming immunity but exhibits slightly lower multipath resilience. |
| Galileo | Europe | E1 ($1575.42\text{ MHz}$) E5a/b ($1176.45 / 1207.14\text{ MHz}$) |
Built primarily for civilian infrastructure. Features highly advanced signal structures that naturally reject noise, heavily contributing to stable RTK fixes and accuracy. |
| BeiDou | China | B1I/B1C ($1561.098 / 1575.42\text{ MHz}$) B2a/B2b ($1176.45 / 1207.14\text{ MHz}$) |
Features exceptional satellite density over the Asia-Pacific region. Drastically increases raw satellite counts, though it demands the highest processing overhead from the receiver. |
| QZSS (Michibiki) | Japan | L1 / L2 / L5 / L6 | Operates in a Quasi-Zenith Orbit, keeping satellites near the directly overhead position (zenith) in the Japan/Asia region. Provides extreme resilience against signal masking in urban canyons and heavily forested terrain. Supports Satellite-Based Augmentation Services (SLAS). |
Countermeasures Against Jamming and Spoofing
For industrial and defense-grade UAS operations, electronic warfare targeting GNSS signals represents a severe operational threat. We must design architectures and FC configurations capable of counteracting brute-force jamming (swamping the receiver with high-power noise) and spoofing (injecting counterfeit signals to hijack the vehicle's state estimation).
1. Multi-Band (Multi-Frequency) Resilience
The vast majority of commercial electronic warfare systems target the ubiquitous L1 band ($1575.42\text{ MHz}$). By specifying multi-band receivers and antennas supporting L1 + L2 + L5 at the hardware design phase, the navigation system can maintain a valid positioning solution using L2 or L5 data even if the L1 spectrum is completely saturated by high-power jamming.
2. Constellation Diversity via Dual-GNSS Architecture
To protect against geographical or geopolitical risks where a specific satellite network is intentionally deactivated or corrupted via spoofing, engineers must implement physical and logical separation.
By configuring ArduPilot’s GPS_GNSS_MODE (and GPS_GNSS_MODE2 for the redundant unit), you can isolate the constellations. For instance, the primary GNSS unit can be restricted to GPS + Galileo + QZSS, while the secondary redundant unit runs exclusively on GLONASS + BeiDou. If anomalous or spoofed data poisons one of the satellite groups, the EKF3 state estimator detects the resulting position/velocity discrepancy (innovation error) between the two independent modules. The system will then automatically switch to the uncorrupted receiver or reject GNSS altogether, triggering an emergency dead-reckoning return mode using the IMU and barometer.
3. Hardware-Level Front-End Defenses
For high-reliability platforms operating in contested environments, the RF front-end should integrate Controlled Reception Pattern Antennas (CRPA) or high-performance Surface Acoustic Wave (SAW) filters embedded within the Low Noise Amplifier (LNA) circuit. These components provide steep out-of-band attenuation, physically blocking out adjacent communication frequencies and high-power electromagnetic interference (EMI) before they reach the receiver's internal processing stages.
GNSS Optimization Parameters in ArduPilot
To tune your ArduPilot configuration for optimal tracking density or fast refresh rates depending on mission requirements, deploy the following parameter baselines:
# Minimum tracking and quality thresholds for reliable position lock
GPS_MIN_SATS 10 # Enforce a minimum of 10 satellites for a valid solution
GPS_MIN_ELEV 15 # Filter out satellites below a 15-degree elevation angle to eliminate multipath noise
# Primary receiver optimization (Example targeted for 10Hz updates on u-blox using GPS+Galileo+QZSS)
GPS_GNSS_MODE 11 # Bitmask configuration activating only essential constellations
GPS_RATE_MS 100 # Set receiver update interval to 100ms (10Hz)
# EKF3 Dual-GNSS blending configuration
GPS_TYPE 1 # Define type for Primary GPS
GPS_TYPE2 1 # Define type for Secondary GPS
GPS_AUTO_SWITCH 2 # Blend data from both receivers to generate a mathematically idealized virtual GPS dataset
Verification and Diagnostic Log Analysis
To validate the multi-GNSS architecture and check for RF interference under real operational loads, evaluate the GPA (GPS Accuracy) and GPS data packets extracted from the onboard telemetry logs:
- Correlating GPS.HDop and Satellite Counts (NSats): Check the log to confirm that each constellation is decoding successfully (e.g., verifying stable BeiDou tracking). If the satellite count exceeds 30 but the HDop fails to drop below 0.7—or if the absolute coordinates exhibit sharp spikes—the receiver's internal processor is likely bottlenecked, causing data latency, or a specific constellation is suffering from severe multipath interference.
- Jamming Indicators (GPA.Jam / u-blox UBX-MON-HW): Monitor the receiver's internal continuous-wave (CW) detection metrics. If the jamming indicator value spikes when powering on the airframe or initiating high-power transmissions (such as booting the VTX), it indicates that high-frequency system noise is actively bleeding into the GNSS front-end.
Summary
The essence of multi-GNSS design lies not in tracking as many signals as possible, but in carefully balancing aircraft dynamics (responsiveness) with environmental robustness (fault tolerance). Understanding the unique parameters of each constellation, implementing clean signal filtering, and structuring EKF3 redundancy to withstand electronic attacks are critical requirements for building mission-ready industrial and defense UAS.
Throughout this 12-part series, we have explored the core aspects of flight controller hardware integration and sensor subsystem design. We hope these engineering frameworks provide a definitive, reliable foundation for your development of next-generation, high-reliability unmanned aerial systems.