8.4 Multirotor Flight Controls, Stabilisation and GPS Hold

Every Input Is a Rotor Speed Change

A multirotor has no ailerons, no elevator, no rudder and no collective. Every control input is a change in the relative speed of the rotors, computed by the flight controller and delivered through the ESCs.

Two things follow from this table and both are examinable.

Translation and attitude are inseparable. A multirotor cannot move horizontally without tilting, because tilting the airframe is the only way to tilt the thrust vector. Any horizontal motion therefore reduces the vertical thrust component, which is why forward acceleration from a hover produces a slight sink unless power is added.

Yaw is torque, not thrust. Because a yaw input speeds one group up and slows the other by the same amount, total thrust is unchanged. This is why a multirotor can yaw on the spot without climbing or descending, and why a yaw input on an aircraft with a failed motor is the input most likely to lose control entirely.

Stabilisation Modes

Modern flight controllers offer a ladder of modes. Each rung removes pilot workload and adds a dependency.

Stabilisation in the general sense means the controller continuously compares the IMU's measured attitude against the commanded attitude and adjusts individual rotor speeds thousands of times per second to close the error. It is what makes a multirotor flyable at all — the airframe is inherently unstable and cannot be flown without it.

GPS Hold and Its Dependencies

GPS hold (position hold, P-mode, loiter) is the mode most commercial work is flown in. The controller uses the GNSS position solution to hold a fixed point over the ground, correcting automatically for wind. Release the sticks and the aircraft stops and stays put.

It has two hard dependencies that a remote pilot must understand:

1. A valid GNSS fix. A 3D fix with enough satellites and a low HDOP. Degrade the fix and the hold degrades with it — the aircraft will wander, oscillate, or drop out of the mode.
2. A healthy compass. Position hold works by computing a corrective vector — a direction and distance back to the held point — and then commanding a tilt in that direction. Commanding a direction requires knowing which way the aircraft is facing. A compass in error by 40 degrees produces corrections applied 40 degrees off, and the aircraft develops the characteristic circular wander pilots call toilet-bowling.

Toilet-bowling is therefore a diagnostic: it almost always means compass error or interference, not a GNSS problem. The correct response is to switch to attitude mode, fly the aircraft home manually, and recalibrate away from the interference source.

When GNSS Is Lost

If the fix degrades or disappears in flight, most flight controllers automatically revert to attitude mode. The aircraft still holds a level attitude and roughly holds altitude, but it no longer holds position — so it drifts with the wind, at whatever speed the wind is blowing.

The consequences to plan for:

• Return-to-home becomes unavailable or unreliable, because RTH navigates to a stored GNSS home point.
• The aircraft moves without pilot input and will keep moving until the pilot corrects it.
• The pilot must fly it manually, using visual reference, all the way to a landing.

This is precisely why the SOC require visual line of sight. A pilot who can see the aircraft can fly it home in attitude mode. A pilot flying on a camera feed with a failed GNSS has neither position information nor a visual reference.

The currency argument

The single most useful piece of practice a commercial remote pilot can do is fly deliberately in attitude mode in a safe, open area, regularly. Position hold makes a modern multirotor so easy that pilots can accumulate hundreds of hours without ever hand-flying one. The first time they do it should not be during an emergency, in wind, over a client's asset.

A sensible personal currency standard: every few weeks, in an open area with light wind, switch to attitude mode and fly a square circuit, a controlled descent and a landing. Note how much correction the wind demands. That is the skill that recovers a GNSS failure.

Other Controls Worth Knowing

Beyond the sticks, the controller carries switches whose behaviour must be known before they are needed:

• Flight-mode switch. Know which physical position selects which mode, by feel, without looking.
• Return-to-home. Know whether your RTH climbs to a set altitude first, and whether that altitude clears the obstacles between the aircraft and home. An RTH altitude set below a treeline is a scripted collision.
• Motor kill / emergency stop. Know whether yours is armed in flight. A mid-air motor cut is an appropriate response to an imminent strike on a person and a catastrophic error in almost every other circumstance.
• Gain or sensitivity settings. Adjusting control-loop gains changes how aggressively the aircraft corrects. Raising gains too far produces high-frequency oscillation; lowering them too far produces a sloppy, wallowing response. Change them only with a reason and test in a safe area.

Reading the aircraft in the hover

Finally, the first hover after take-off is a control check, and it is where problems announce themselves:

• Persistent drift with no stick input and no wind → compass or IMU problem.
• Slow circular wander → compass error (toilet-bowling).
• High-frequency twitch or oscillation → gains too high, or vibration reaching the IMU.
• Sluggish, delayed response → gains too low, an overweight aircraft, or an under-propped motor combination.
• A consistent tilt in one direction while holding position → wind, or a centre-of-gravity offset.

Land and investigate any of these before transiting. A fault found at the hover check is cheap; the same fault found 300 m downrange is not.