Building a Shared Coordinate System for GPS-Denied Operations Across Air and Ground
At the recent DGI 2026 conference in London, experts from Niantic Spatial and Vantor came together to discuss one of the most urgent challenges facing modern operations: what happens when GPS becomes a vulnerability? Responding to a live poll, 59% of the audience said that jamming and challenging topography caused signal loss in their operations.
From electronic warfare to urban environments, GPS disruption is increasingly common. When satellite signals are jammed, spoofed, or blocked, autonomous systems can lose the ability to navigate. Ground teams lose orientation. And the shared situational awareness that complex operations depend on breaks down.
To address this vulnerability, Niantic Spatial and Vantor are integrating technologies to build a mission-ready joint capability: a shared coordinate system for GPS-denied operations powered by visual positioning, enabled by advanced georegistration, and anchored to a global 3D spatial foundation.
During the panel, Charlie Houseago, Research Productization Lead for Spatial AI at Niantic Spatial, moderated a conversation with Tory Smith, Director of Product Management at Niantic Spatial, and Josh Sisskind, Principal Product Manager at Vantor, about how this integrated system works and what it could unlock in the coming years.
Below is an edited Q&A from that discussion.
Q: Why is GPS denial becoming such a critical problem?
Josh: GPS denial can happen in many ways. It could be deliberate jamming or spoofing, but it can also occur for more mundane reasons like signal obstruction. Urban environments, dense infrastructure, and rugged terrain can all interfere with GPS signals.
The important point is that it’s no longer useful to ask why GPS might fail. We must assume that at some point it will. So, the question becomes: how do we build systems that continue operating when it does? What happens operationally when positioning breaks down? What is the most resilient set of alternatives based on diverse scenarios?
Tory: That’s right. We see similar issues even in commercial environments. Inside buildings or in dense cities, GPS signals can bounce off glass and concrete. The result is multi-path error, where your position might drift tens of meters from reality. In a consumer setting that’s frustrating. In a mission critical operation context, it’s unacceptable.
Without a reliable positioning reference, systems lose the ability to coordinate. A drone, a ground vehicle, and a human operator may all be looking at the same environment but operating in completely different coordinate frames.
Q: How can visual positioning help solve this problem?
Tory: Visual positioning systems work by comparing a live camera feed to a known model of the world. When a device captures imagery, the system identifies visual features—buildings, terrain patterns, structural geometry—and matches those features against a reference dataset. Once the match is established, the system can determine precise position and orientation.
Our system can achieve centimeter-level localization under the right conditions, but historically these systems required a prior estimate of location, usually from GPS. The challenge in GPS-denied environments is that the prior may not exist. That’s where integrating with Vantor’s spatial foundation becomes powerful.
Q: Where does Vantor’s technology fit into the solution?
Josh: Vantor provides the global spatial reference layer that these visual systems can anchor to. Our global-scale 3D terrain models are built from more than two decades of satellite imagery using multi-view reconstruction techniques. That produces a detailed digital representation of the Earth’s surface that can serve as a consistent spatial foundation accurate down to 3 m in all dimensions.
When a drone captures full motion video and imagery from the air, Raptor, our visual positioning and georegistration software suite, can align those images against that terrain data to determine its position, even when GPS isn’t available.
By combining that aerial localization capability with Niantic Spatial’s ground-based Visual Positioning System (VPS), we create something new: a shared coordinate system that connects the air and the ground.
Q: What does “unified air-to-ground positioning” mean?
Tory: It means that every system operating in the environment – whether it’s a drone, a robot, a soldier-borne device, or an AR headset - can determine its location relative to the same spatial reference.
On the ground, our VPS aligns camera imagery against detailed spatial models. In the air, Vantor’s system aligns drone imagery against global 3D terrain data. Once both systems reference the same underlying model, air and ground assets can operate in a shared coordinate frame, enabling real-time coordination even without GPS.
The goal is not to swap out GPS or established backup solutions, but to augment them with a resilient visual positioning layer that keeps systems aligned when GPS performance breaks down.
Q: How accurate is the integrated system today?
Tory: GPS typically provides around five meters of accuracy under good conditions. In early testing that integrates Niantic Spatial’s VPS with Vantor’s terrain data and Raptor software, we’ve seen up to a 70% reduction in error and down to roughly 1.5 meters in many scenarios.
That level of precision becomes especially important when GPS is completely unavailable, and when operators would otherwise have little confidence in their location. In the room today, 59% have ranked accuracy as their primary decision factor for adopting new positioning technology.
Q: How easily can this technology integrate into existing systems?
Josh: One of the design principles behind both Raptor and VPS is flexibility. For example, Raptor software runs on existing hardware and typically uses the electro-optical camera already installed on a drone or similar aerial platform. Integration happens through an SDK, meaning manufacturers or integrators can incorporate the capability directly into their systems.
The same principle applies across the broader architecture. Whether the device is a drone, a robot, or a handheld device running ATAK, the goal is to plug into the existing ecosystem rather than forcing operators to adopt entirely new hardware.
Q: What are the current limitations of visual positioning systems?
Tory: Like any technology, visual positioning has edge cases. Dense tree canopy can obscure visual features. Interiors of buildings can be challenging if the system doesn’t have reference imagery. And extremely featureless environments like open ocean or deserts can make visual alignment more difficult.
But in environments where visual features are present, the technology performs well. And when multiple sensors are working together across domains, those edge cases become easier to manage.
Q: Looking ahead, what does this enable in the next three to five years?
Josh: The real breakthrough is interoperability. This is a great example of two companies coming together to make their systems interoperable for the good of the mission. When every sensor across space, air, and ground is accurately anchored to the same spatial foundation, you remove the friction that normally slows operations down. Instead of spending time aligning data sources, you can focus on decisions by leveraging fusion.
That shared reference system also unlocks new possibilities for AI-enabled analysis and autonomous systems that depend on precise spatial awareness.
–
Toward resilient autonomy and interoperability
As autonomy expands across defense, robotics, and mixed-reality systems, reliable positioning becomes essential for successful operations.
Systems must continue operating when GPS fails. They must coordinate across domains. And they must share a common understanding of the physical world.
The partnership between Niantic Spatial and Vantor is built around that vision: combining visual positioning technologies with a global spatial foundation to create a unified coordinate system that connects sensors across space, air, and ground.
To speak to the team and learn more about the integrated Vantor-Niantic Spatial solution, fill out this contact form.
Note: A version of this blog also appears on Vantor.com