Drone Technology

The frame the whole path hangs on. A multirotor, a fixed wing and a hybrid are three different answers to the same problem, and each buys hover capability, endurance or range at the direct expense of the others. Establishing that trade first means every later unit — propulsion, power, autonomy — is understood as serving a chosen compromise rather than as a set of universal facts about 'drones'.

The counterintuitive fact that shapes everything downstream: a multirotor has no control surfaces and no inherent stability. It steers entirely by varying the thrust and torque of fixed propellers, and it is aerodynamically unstable — it stays upright only because a control loop corrects it hundreds of times a second. Once that lands, the flight controller stops being an accessory and becomes the reason the aircraft exists in that form at all.

The three components that must be matched to each other and to the airframe, and the place where component-shopping goes wrong most expensively. Motor sizing and velocity constant, propeller diameter and pitch, and the electronic speed controller in between. The system-level number is thrust-to-weight ratio: it decides whether the aircraft is controllable, sluggish or violent, and it is a design output rather than a specification you can shop for.

🔴 The unit that constrains every drone design and the one with the most serious hazard content in the path. Endurance is a battery problem before it is anything else, and the relevant behaviour is not the printed capacity but discharge rating, voltage sag under load, usable versus nominal capacity, and the mass penalty of carrying more. The Risk elements are genuine and non-negotiable: lithium polymer thermal runaway, charging supervision, storage state of charge, physical damage, and transport rules. Hazard AND mitigation, which is not what a definition card tests.

The inputs to everything autonomous. Inertial measurement for rotation and acceleration, barometric pressure for altitude, magnetometer for heading, satellite navigation for position, and optical or ranging sensors for the near field. Taught with each sensor's specific weakness stated alongside it — drift, noise, vibration coupling, magnetic interference from the power wiring, multipath — because the weaknesses are the entire reason the next unit exists.

🔴 The conceptual centre of the path, and the unit most hobbyist material skips entirely. No single sensor gives usable state: accelerometers are noisy, gyroscopes drift, magnetometers are disturbed, satellite fixes are slow and occasionally wrong. Fusion combines them into an estimate better than any input, and understanding it is what separates 'the drone flew away' from a diagnosis. Nothing in autonomy, failsafes or log analysis can be reasoned about without it.

The hardware and the software that runs the loop from unit A2 using the estimate from unit A6. Covers the honest split in the ecosystem — firmware optimised for manually flown racing and acrobatics, versus firmware built around autonomous mission execution — because they have different architectures, different configuration models and different communities, and choosing by popularity rather than by intent is a common early mistake.

The cascade: an outer loop wanting a position, feeding an attitude loop, feeding a rate loop that talks to the motors. Each stage is a proportional-integral-derivative controller, and tuning is adjusting how aggressively each responds. This unit exists to replace the widespread practice of copying tuning values from a forum with an actual model of what each term does and what its failure looks like in flight.

Three separate links with three different requirements, frequently confused. Control needs low latency and extreme reliability at low data rate; video needs high throughput and tolerates loss; telemetry sits between. Covers the frequency bands in use, why range is a link-budget calculation rather than a marketing number, antenna polarisation and placement on a metal-and-carbon airframe, and interference between the links you are carrying.

How a human supervises an autonomous aircraft. The standard messaging protocol used across the open ecosystem, what a ground control station actually does, mission planning and upload, live telemetry, and the parameter interface that exposes essentially every setting in the firmware. This is also the practical entry point for anyone wanting to write software that talks to an aircraft.

Waypoint missions, return-to-home, geofencing, and the behaviour on lost link or low battery. Placed after fusion and control deliberately, because every one of these features is only as good as the state estimate underneath it — and the classic incidents are failsafes doing exactly what they were configured to do with bad position data. The Risk elements are about the interaction between failsafes, which is where the surprising outcomes come from.

Cameras, gimbals, and sensing payloads, and the fact that every gram and every watt is taken from endurance. This unit is really about systems engineering: a payload decision propagates back through the power budget, the propulsion sizing and the airframe, and treating it as a bolt-on is how a design ends up with three minutes of flight time.

A Skill element, closed by a built aircraft rather than an answer. Wiring, connector and current sizing, soldering the power distribution, mounting the flight controller with the correct orientation and vibration isolation, propeller balance, and centre-of-gravity placement. Vibration deserves particular attention: it corrupts inertial measurement and therefore the state estimate, so a mechanical build fault surfaces as a mystifying control problem.

The diagnostic skill, and the one that makes someone genuinely useful. Modern flight controllers record high-rate telemetry of everything, and almost every incident is fully explained in the log by someone who can read it. A Skill element by construction: the assessment is diagnosing real logged failures, which cannot be done by recall.

The seam, not the subject. Remote identification, weight thresholds, and the airworthiness expectations that constrain what may be built and flown — enough to know which decisions have regulatory consequences and when to go and read the actual rules. The operating and certification content lives in the FAA Part 107 path and is deliberately not duplicated here.