7.1 · Intermediate

GNSS Receiver Architecture: From Antenna to Position

Introduction

"Inside every GNSS receiver, there is a complex chain of hardware and algorithms working together to turn radio signals into position."

A GNSS receiver is far more than a simple radio. It is a carefully engineered system that must detect and decode signals arriving from satellites over 20,000 km away, at power levels so low they are buried beneath the noise floor. Understanding how this works - from the first contact with the radio wave to the final position solution - is essential for any GNSS engineer.

The Four Core Stages

Every GNSS receiver, from a consumer-grade smartphone chip to a geodetic survey instrument, shares the same fundamental architecture divided into four sequential processing stages.

1. RF Front-End: Signal Capture and Conditioning

The first stage is the radio frequency (RF) front-end. This is the analogue hardware that interfaces directly with the antenna. Its job is to take the extremely weak L-band signals (in the 1.2 to 1.6 GHz range) and convert them into a form that digital processing circuits can work with.

The front-end performs three critical functions. First, a Low-Noise Amplifier (LNA) - often located inside the antenna itself - boosts the signal while adding as little noise as possible. Second, bandpass filters reject out-of-band interference from other radio systems. Third, a down-conversion stage mixes the RF signal with a local oscillator to shift it down to an intermediate frequency (IF) or directly to baseband. Finally, an Analogue-to-Digital Converter (ADC) digitises the resulting in-phase (I) and quadrature (Q) samples, typically at rates between 2 MHz and 40 MHz depending on the signal type being tracked.

Key Concept: The noise figure of the RF front-end directly determines receiver sensitivity. A front-end that adds too much noise will struggle to track weak signals, particularly at low satellite elevations or under foliage.

2. Signal Acquisition and Tracking

Once the signal is digitised, the baseband processor takes over. It must first acquire the satellite signals - a process of searching through possible code delays and Doppler frequency offsets simultaneously. Each GNSS satellite transmits a unique Pseudo-Random Noise (PRN) code, and the receiver generates local replicas of each code. When the local replica aligns with the incoming signal, a correlation peak is detected. This peak identifies both the satellite and its approximate code phase and Doppler shift.

After acquisition provides a coarse estimate, the tracking loops take over to maintain continuous lock. A Delay-Lock Loop (DLL) tracks the code phase to extract pseudorange measurements, while a Phase-Lock Loop (PLL) or Frequency-Lock Loop (FLL) tracks the carrier frequency and phase. These loops continuously adjust the local replica to stay aligned with the incoming signal as satellite geometry changes, and they output the raw observables - pseudorange, carrier phase, and Doppler - that are used to compute position.

Note: Signal tracking quality is the single most important factor in positioning accuracy. Poorly tuned tracking loops, or signals weakened by obstructions, directly degrade pseudorange and carrier phase measurements.

3. Measurement Generation: The Observables

The tracking stage produces three primary observables for each satellite signal being tracked.

ObservableDescriptionTypical Precision
PseudorangeCode-based distance measurement to the satellite0.1 – 1 m
Carrier PhasePhase of the carrier wave - precise but ambiguous1 – 5 mm
DopplerRate of change of range - used for velocitycm/s

The carrier phase observable is the most precise but contains an unknown integer ambiguity - the number of whole wavelengths between satellite and receiver. Resolving this ambiguity is the basis of RTK and PPP-AR positioning techniques that achieve centimetre-level accuracy.

4. Position Engine: Computing the Solution

The position engine, sometimes called the navigation processor or PVT engine (Position, Velocity, Time), collects observables from all tracked satellites and computes the final solution. Using the known satellite positions (from broadcast or precise ephemeris data) and the measured pseudoranges, it solves a system of equations to determine the receiver's position in three dimensions, along with a receiver clock correction.

Modern position engines go well beyond simple single-point positioning. They apply atmospheric correction models, weight observations by satellite elevation, perform integrity monitoring to detect faulty signals, and - in RTK or PPP mode - process differential corrections or precise orbit and clock data to achieve centimetre accuracy.

Key Concept: A minimum of four satellites is required to solve for four unknowns: X, Y, Z position and receiver clock offset. Additional satellites improve accuracy, redundancy, and the ability to detect outliers.

Hardware vs Software: Both Matter

Modern GNSS receivers increasingly blur the line between hardware and software. Software-Defined Radio (SDR) receivers perform much of the signal processing in software running on general-purpose processors, offering flexibility to add new signal types or update algorithms without hardware changes. Even traditional hardware receivers rely heavily on firmware algorithms for tracking loop design, multipath mitigation, ambiguity resolution, and integrity monitoring.

Vital Points

  • GNSS is both a hardware and software system - both layers must be well-engineered for accurate results.
  • Signal tracking quality is the critical link between signal reception and positioning accuracy; any disruption propagates directly into the solution.
  • Poor tracking caused by obstructions, interference, or poor antenna placement will produce poor positioning regardless of how sophisticated the position engine is.
  • Understanding receiver architecture helps engineers diagnose problems and select the right equipment for each application.