9.6 · Advanced

RAIM and ARAIM: How GNSS Detects Its Own Failures

Introduction

GNSS receivers are designed to detect when their own navigation solution has become unreliable - without requiring input from external monitoring networks. This capability is called Receiver Autonomous Integrity Monitoring (RAIM). As multi-constellation GNSS has become universal, Advanced RAIM (ARAIM) has extended these concepts to take advantage of the statistical redundancy that multiple independent constellations provide. Understanding RAIM and ARAIM is essential for any engineer working on safety-rated GNSS applications.

The Principle of RAIM

RAIM exploits the redundancy that comes from tracking more satellites than the minimum needed for a position solution. With exactly four satellites, the receiver has just enough observations to solve for its four unknowns (X, Y, Z, clock offset). There is no redundancy - if one satellite provides an erroneous pseudorange, the error propagates directly into the position solution with no means of detection.

With five or more satellites, the system is over-determined: there are more observations than unknowns. The receiver can compute a position solution and examine the post-fit residuals - the differences between the measured pseudoranges and the ranges predicted from the computed position. If all satellites are providing valid measurements, the residuals should be small and consistent. A single faulty satellite will cause an inconsistent residual that can be detected through statistical testing.

Key Concept: RAIM requires at least five satellites for fault detection (FD) and six satellites for fault detection and exclusion (FDE). With FDE, the algorithm can identify which satellite is faulty and remove it from the solution, recovering a valid position using the remaining satellites.

RAIM Algorithms

The most common RAIM algorithm for aviation is snapshot RAIM (or least-squares RAIM), which processes all satellites together in a single epoch:

  1. Compute a position solution using all available satellites.
  2. Form the post-fit residual vector - the difference between observed and predicted pseudoranges.
  3. Compute a test statistic (typically the squared sum of normalised residuals).
  4. Compare the test statistic to a threshold derived from the target false-alert probability.
  5. If the threshold is exceeded, a fault is declared. If enough satellites are available, sequentially exclude satellites and test whether exclusion of one satellite brings the statistic below threshold - if so, that satellite is excluded.

Limitations of Classic RAIM

Classic single-constellation RAIM has important limitations. It can only detect and exclude one faulty satellite at a time. It performs poorly when satellite geometry is poor (high DOP), as the protection levels widen and may exceed alert limits. It also cannot provide integrity at the precision approach level for aviation because the precision approach vertical alert limit (10–15 m for LPV-200) is tighter than what GPS-only RAIM can reliably guarantee.

Advanced RAIM (ARAIM)

ARAIM exploits the combination of multiple independent constellations to provide integrity at precision approach levels without requiring a ground-based augmentation system. The key insight is that GPS, Galileo, GLONASS, and BeiDou are operated independently by different authorities and subject to independent failure modes. The probability of simultaneous failures across multiple constellations is extremely small, which allows ARAIM to compute protection levels that meet aviation requirements even without a SBAS ground network.

FeatureClassic RAIMARAIM
ConstellationsSingle (typically GPS)Multi-constellation
Fault modelSingle-satellite faultMultiple simultaneous faults, constellation-level faults
Performance levelEn-route to non-precision approachPrecision approach (LPV-200 and below in development)
InfrastructureReceiver onlyReceiver + Integrity Support Message
StandardisationMature (DO-229)Under development (ICAO working group)

ARAIM requires an Integrity Support Message (ISM) - a broadcast message that provides the statistical parameters characterising the expected performance of each constellation (satellite fault probabilities, signal-in-space errors). The receiver uses these parameters to compute protection levels that account for both user-level satellite faults and constellation-level events.

Note: ARAIM is a developing standard. As of 2025, it is not yet certified for operational aviation use, though extensive research and field validation have been completed. Its deployment represents a significant step toward GNSS-based precision approach navigation without requiring dense ground monitoring networks.

Vital Points

  • RAIM uses redundant satellite measurements to detect faulty signals autonomously - it requires a minimum of five satellites for detection and six for exclusion.
  • Fault detection and exclusion (FDE) allows the receiver to identify and remove a faulty satellite and continue with a valid solution.
  • Classic RAIM is limited to single-satellite fault detection in single-constellation environments - it is insufficient for precision approach integrity requirements.
  • ARAIM leverages multi-constellation independence to provide tighter protection levels suitable for precision approach navigation without ground infrastructure.