Autonomy
The ability of a machine to carry out a task without a person directing each step. Autonomy is a spectrum, not a switch: a system may plan its own route while still asking a human before doing anything consequential.
Reference · Glossary
Defence autonomy,
in plain English.
Autonomy has accumulated a lot of jargon, and much of it obscures more than it explains. This is our working glossary — the terms we use across this site, defined the way we would explain them to someone in the room. No marketing, no acronym soup.
The Vocabulary
The terms
that matter.
Grouped by how they are actually used rather than alphabetically — starting with the ideas everything else depends on.
Autonomous system
A platform that can sense its surroundings, decide what to do and act on that decision. At Rewind Dynamics this covers unmanned aerial, humanoid and tactical miniature platforms.
UAV / UAS
Unmanned aerial vehicle, or unmanned aerial system. UAV usually means the aircraft itself; UAS means the whole system — aircraft, ground control, communications and operators.
ISR
Intelligence, surveillance and reconnaissance. The work of gathering information about an area or situation so that commanders can make informed decisions.
Surveillance
Watching an area persistently over time to understand what is normal and notice what changes.
Reconnaissance
Going to look at a specific place to answer a specific question, usually once and quickly.
Force protection
Measures that keep personnel, equipment and sites safe — including early warning of an approaching threat.
Sense–decide–act
The core loop every autonomous system runs, many times a second: take in the world, choose an action, carry it out, and repeat. Making that loop fast, reliable and safe is the central engineering problem.
Perception
Turning raw camera and sensor data into a usable understanding of the environment — what is there, where it is, and whether it matters.
Sensor fusion
Combining several different sensors into a single picture that is more reliable than any one sensor could be alone.
Edge AI
Running artificial intelligence on the platform itself rather than in a remote data centre, so the system stays effective without a constant communications link.
GPS-denied navigation
Knowing your position and finding your way when satellite navigation is jammed, spoofed or simply unavailable — typically by using vision and onboard sensors instead.
GNSS
Global navigation satellite system: the general term for satellite positioning networks, of which GPS is one.
SLAM
Simultaneous localisation and mapping. Building a map of an unknown space while also working out where you are within it.
Swarm
A group of autonomous systems that coordinate as one team — sharing what they see, dividing the task, and remaining useful even if some are lost.
Mother–child architecture
A two-tier design in which a larger carrier platform transports and coordinates a group of much smaller platforms, which do the close-in work. One of our active research directions.
Nano-drone
A very small unmanned aircraft, typically light enough to be carried in quantity and cheap enough to be treated as expendable.
SWaP
Size, weight and power — the three constraints that dominate the design of any small autonomous system. Every capability added costs something in at least one of them.
Man-portable
Small and light enough to be carried and deployed by one person alongside their existing equipment.
Dismounted
Operating on foot, away from a vehicle. Dismounted teams have the tightest constraints on what they can carry.
Tactical edge
The forward, close-range environment where information is scarce, links are unreliable and decisions are made under pressure.
Degraded link
A communications connection that has become slow, intermittent or unusable. Well-designed autonomy continues behaving sensibly when this happens.
Human-in-the-loop
An arrangement where the system cannot take a significant action until a person approves it.
Human-on-the-loop
An arrangement where the system acts on its own but a person supervises continuously and can intervene or stop it at any time.
Human in command
Our own framing, and our operating principle: machines carry out tasks, but people set the rules, define the limits and remain responsible for the outcome.
Autonomy stack
The layered software architecture that produces autonomous behaviour — from low-level control, through perception and planning, up to mission and oversight.
Guidance, navigation and control
The three related jobs of deciding where to go, knowing where you are, and making the platform actually do it.
Simulation-first
Developing and proving behaviour in software before committing it to hardware, because software tests are faster, safer and repeatable.
After-action assessment
Reviewing a recorded run honestly to establish what went right, what went wrong and why — the discipline our name refers to.
Payload
Whatever a platform carries to do its job — usually sensors. Payload always competes with endurance for the same limited energy.
Endurance
How long a platform can operate before it must return, recharge or refuel.
Teleoperation
Direct remote control of a machine by a person. Distinct from autonomy, where the machine decides for itself within set limits.
These are working definitions written for clarity, not formal doctrine — usage varies between organisations and countries. If you think one of them is wrong or could be clearer, tell us and we'll fix it.
See It Applied
From definitions
to systems.
These ideas are what we actually build with. See how they come together in our technology and the systems we're developing.