The mother
Range, endurance and compute. It carries the children in, coordinates them, and holds the link back to the operator.
Research & Concepts
Ideas we're
working on.
A look at the research directions that shape where Rewind Dynamics is headed — the harder problems in autonomy we're exploring for defence and military applications. These are concepts and open questions, not finished products. We describe the idea; the engineering we keep for the people who need it.
How To Read This Page
Directions, not datasheets.
Everything below is early-stage research. We share the shape of each idea and why it matters for defence — deliberately at a high level. No specifications, no performance figures, no operational detail. Serious discussion happens directly with partners, evaluators and the armed forces, under appropriate terms.
01
Concept-level
We explain the intent and the value — not how it's built.
02
Defence-focused
Every direction is framed for defence and military use.
03
Human-in-command
A person sets the rules and stays responsible throughout.
04
In R&D
Nothing here is fielded. These are problems we're actively studying.
Flagship Concept
Two-tier autonomy:
a mother, and its children.
One of our most exciting research directions is a two-tier "mother–child" architecture for autonomous nano-drones. A larger carrier platform — the "mother" — travels to the area of interest and releases a group of very small, expendable "child" drones to do the close-in work.
The idea is a division of labour. The mother carries range, endurance, compute and communications; the children carry almost nothing but their sensors and their autonomy. Because the children are tiny, they can go where a larger aircraft can't — low, close, and quiet — while the mother stays back, coordinates them and relays what they find.
For defence, that opens up reconnaissance and observation in cluttered, confined or contested spaces without putting a person or an expensive platform forward. The children act as a forward layer of eyes; the mother is the brain and the link back to the people in command.
Mother · ChildrenArchitecture
The children
Tiny, low-cost and expendable. They go close and low to sense what a large platform never could — and are meant to be lost without loss.
Shared autonomy
The two tiers think as one system — the mother planning, the children executing, each adapting as the picture changes.
Why it matters
Reach into confined, contested or dangerous spaces without putting people or costly aircraft forward.
Concept
Autonomous surveillance & ISR.
A second direction is persistent, autonomous observation — intelligence, surveillance and reconnaissance that runs with minimal operator load. The aim is simple to state: watch an area reliably, understand what's normal, and flag what isn't — so a human can decide what it means.
We're deliberately keeping this one high-level. The value for defence is a forward set of patient, discreet eyes that reduce risk to personnel and extend how much a small team can keep watch over. A person always sets the tasking and stays in command of what's done with it.
Observe · Understand · AlertISR
PersistentWatch reliably, over long periods
DiscreetLow signature, minimal footprint
Pattern-awareLearn what's normal, surface what isn't
Human-decidedPeople judge meaning and response
More Directions
Other problems
we're exploring.
The hard, shared problems underneath everything above. Progress on any of these lifts every platform we build.
Swarming
Many small autonomous systems acting as one coordinated team — sharing what they see, splitting the task, and staying useful even if some are lost.
GPS-denied navigation
Knowing where you are and finding your way when satellite signals are jammed, spoofed or simply unavailable — using what the system can see and sense.
Onboard perception
Making sense of a cluttered, changing world in real time, on the platform itself — the difference between reacting and truly understanding.
Edge intelligence
Running capable autonomy on tiny, power-limited hardware, so a small system stays effective without a constant link home.
Trustworthy autonomy
Autonomy that stays predictable, safe and reviewable under stress — behaviour a commander can rely on and account for afterwards.
Human–machine teaming
The right hand-offs between operator and system, so autonomy amplifies a person's judgement instead of replacing it.
Ground Rules
How we research.
Ideas are cheap; disciplined research is not. A few principles hold across everything on this page.
01
Simulate first
We prove ideas in software before they ever touch hardware.
02
Rewind & assess
Every run is recorded and reviewed honestly — the idea behind our name.
03
Human in command
People set the rules and stay responsible; machines carry out tasks.
04
Restraint on detail
We share the idea, not the sensitive specifics. Serious talk happens privately.
Everything on this page describes research directions and concepts under active study at Rewind Dynamics. Nothing here is a product, specification, or claim of a fielded capability. All systems remain in research and development; our first real-world mission test is planned for December 2026.
Let's Talk
Working on something
adjacent?
If you're a partner, evaluator or engineer who cares about these problems, we'd like to hear from you. Detail is shared privately, under appropriate terms.