GNSS-denied appears in robotics briefs and procurement specs all the time, usually without a precise definition. In practice it covers three distinct situations that impose very different requirements on a localization system, and conflating them is a reliable way to design something that fails in production.
Signal absence
Underground environments (tunnels, mines, basement facilities) have no GNSS signal at all. The system must operate without any satellite-based position reference from the moment it enters the environment. This is the clearest case and the one most teams design for when they hear the term.
Signal degradation
Urban canyons, dense industrial facilities, and environments with significant metal infrastructure receive a signal that is intermittent, multipath-corrupted, or geometrically degraded. The system may get a position estimate, but that estimate has unpredictable error: sometimes 2 meters, sometimes 20. A navigation stack that treats this as reliable will fail in ways that are extremely difficult to reproduce in testing and deeply unpleasant to debug in the field.
Signal denial
Defense, critical infrastructure, and some commercial environments operate under the assumption that GNSS signals may be intentionally jammed or spoofed. The system must be designed expecting that any satellite-derived position is potentially adversarial. Most civilian localization stacks are not built for this, and the gap matters more than teams realize until they need the contract.
Deploying in GNSS-denied environments? Bring your specific constraints to the conversation.
Book a demo →