Perfect connectivity is the wrong design condition.
A system that performs brilliantly with clean GPS, uninterrupted communications and unlimited operator attention may still fail exactly when it is needed most. Modern military autonomy has to start from the opposite assumption: positioning can become unreliable, links can degrade, bandwidth can collapse and the operator can be managing several problems at once.
NATO DIANA now frames autonomy in almost exactly those terms, emphasizing systems that can operate in complex, degraded and contested environments where communications are disrupted and GPS cannot simply be assumed.
That changes the definition of autonomy. It is not just the ability to follow waypoints without continuous stick input. It is the ability to preserve useful mission behavior as the information environment becomes incomplete.
Navigation resilience is more than 'GPS denied'.
GNSS failure is often discussed as a binary condition: available or denied. Real environments are messier. Signals can be jammed, spoofed, intermittent or locally unreliable. A system may have accurate navigation for part of a mission and then lose confidence precisely when it reaches the most contested area.
DARPA's current work on GPS-free precision timing is one example of how seriously militaries are treating the underlying position, navigation and timing problem. The point is not that every small drone needs an optical clock. The point is that defence architectures are being redesigned around the idea that GPS-dependent timing and navigation can be attacked.
For autonomous aircraft, resilience therefore comes from combining multiple sources of state, maintaining confidence estimates and knowing what behavior remains safe and useful when certainty falls.
The second bottleneck is the operator.
Even a perfectly resilient aircraft can overwhelm the people controlling it. If one operator has to manually pilot every platform, monitor every video feed, interpret every sensor alert and coordinate every task, adding more drones can reduce effectiveness instead of increasing it.
Mission autonomy should therefore remove repetitive coordination work while keeping important decisions visible and controllable. That means task-level commands instead of constant micromanagement, automated re-tasking where appropriate, and a common operational picture that lets one operator understand what the system is doing.
SPARTA is already performing this role across MATRIX's battle-tested FPV, fiber-optic and interceptor systems. It connects mission execution, aircraft behavior and operator control so the software layer is not something added after the hardware is finished; it is part of how the system is operated.
Collaborative autonomy matters because one aircraft rarely has the whole picture.
A single platform has limited sensing, endurance and perspective. Multiple aircraft can cover more space, approach from different directions and continue operating when one node is lost—but only if coordination does not create more work than it removes.
That is the logic behind HIVE: multiple high-speed FPV aircraft treated as a coordinated mission system rather than a collection of unrelated vehicles. The same architecture extends across RAVEN, BLACK WIDOW and SI, allowing different air vehicles to participate in a common mission layer while keeping their roles distinct.
The important word is not swarm. It is coordination. Useful collaborative autonomy is about distributing sensing, tasks and decisions so the network becomes more capable than the individual aircraft.
The standard should be mission continuity, not autonomy theatre.
Defence autonomy is easy to demonstrate in a clean environment. The harder question is what the system does after the demo conditions disappear.
Does it remain navigationally coherent when GNSS quality falls? Does it degrade gracefully when communications are intermittent? Can the operator understand what is happening without watching eight separate interfaces? Can a mission continue when individual nodes drop out? Can the software support a battle-tested FPV system today and a low-observable aircraft tomorrow without becoming a different product every time?
That is the standard MATRIX is building SPARTA around. The software is already embedded in battle-tested systems, while the same mission architecture connects the broader portfolio. The objective is not autonomy as a feature. It is an operational system that stays useful when the network becomes uncertain.
Sources and further reading
NATO DIANA — Autonomy and unmanned systems in degraded environments ↗DARPA — GPS-free operations and resilient timing ↗NATO ACT — Integrated counter-UAS and command-and-control ↗NCIA — Integrating counter-drone technologies into military operations ↗