What is an aircraft autopilot, and how does it work?
An aircraft autopilot is a flight-control system that automatically moves the controls to hold or follow selected targets such as pitch, heading, altitude, speed or a navigation path. In real-world aviation, it compares sensor data with the commanded flight path, calculates corrections, and drives servos while the pilot selects modes and monitors performance.
How does an aircraft autopilot work?
An autopilot works as a closed-loop control system: it measures what the aircraft is doing, compares that with what has been commanded, then corrects any difference.
- The pilot selects a target and mode. This could be a heading, altitude, vertical speed, airspeed, localiser or route generated by the flight management system.
- Sensors measure the aircraft's state. Attitude and heading systems provide pitch, roll and direction; air-data systems supply altitude, airspeed and vertical speed; navigation receivers and the flight management system provide lateral or vertical guidance.
- The flight-guidance computer calculates an error. If the aircraft is below the selected altitude or drifting away from a route, the system determines the required correction.
- Servos or flight-control computers move the controls. Conventional aircraft may use electric or hydraulic servos connected to the control system. In fly-by-wire aircraft, the autopilot sends commands through the electronic flight-control computers.
- The cycle repeats continuously. New sensor readings show whether the correction worked, allowing the autopilot to adjust the command and maintain a stable path.
Many systems also operate trim to remove sustained control loads. The design, available modes and permitted operating limits vary considerably between a light aircraft, an airliner and a military aircraft.
What can an aircraft autopilot control?
An aircraft autopilot normally controls one or more rotational axes, but it does not necessarily manage the engines, route or every phase of flight.
- Roll: ailerons and sometimes spoilers control bank angle, heading, route tracking and localiser capture.
- Pitch: the elevator and stabiliser control pitch attitude, altitude, vertical speed, airspeed-based climbs and vertical navigation.
- Yaw: the rudder may provide turn coordination or directional control, although many aircraft use a separate yaw damper.
A basic single-axis autopilot may only keep the wings level or hold a heading. A two-axis system controls roll and pitch, while a three-axis installation also controls yaw. Capability depends on the aircraft rather than the presence of an AP button alone.
| System | What it does |
|---|---|
| Autopilot | Moves the flight controls to follow active guidance modes. |
| Flight director | Displays pitch and roll commands for the pilot or autopilot to follow. |
| Flight management system | Stores the route and performance data and can generate navigation guidance. |
| Autothrottle or autothrust | Controls engine thrust; it is usually separate from the autopilot but closely integrated with it. |
A speed mode does not always mean that thrust is automatic. Some autopilots maintain airspeed by changing pitch, which can alter the climb or descent rate. Without autothrottle or autothrust, the pilot may still need to set power manually.
The most important indication is the Flight Mode Annunciator, normally shown across the top of the primary flight display. It identifies which lateral, vertical and thrust modes are active or armed. A mistake we see constantly is watching the control-panel button rather than confirming the mode on the display. Our guide to how the A320 FCU, flight displays and mode annunciations interact illustrates this clearly.
Names and control logic differ by manufacturer. Airbus commonly distinguishes selected and managed guidance, while Boeing uses its own mode-control and flight-management conventions; our comparison of Boeing and Airbus flight-control philosophies explains the broader distinction. In every case, an armed mode is waiting for capture, while an active mode is controlling the aircraft now.
Can autopilot take off and land an aircraft?
Conventional airline autopilots normally do not perform the take-off, but suitably equipped and certified aircraft can conduct an automatic approach and landing under specific conditions.
After take-off, pilots engage the autopilot only when permitted by the aircraft's limitations and operating procedures. There is no universal engagement height: it depends on the type, system status and operation.
An autoland requires more than selecting approach mode. The aircraft, operator, crew and runway facilities must be qualified, the required navigation signals and redundant systems must be available, and wind and system limits must be respected. Depending on the installation, the system may control localiser tracking, glideslope, flare and runway rollout.
Approach mode by itself does not guarantee an automatic landing. Pilots still configure the aircraft, select or verify speeds, lower the landing gear and flaps, manage braking, monitor every mode transition and disconnect if performance becomes unsafe. Some specialist emergency landing systems can select an airport and land after activation, but that is not normal autopilot capability.
For a practical view of how heading, altitude and approach modes support rather than replace the pilot, see our guide to using automation during simulated IFR flying.
Why does an autopilot sometimes behave incorrectly?
When an autopilot behaves unexpectedly, the usual cause is the wrong active mode, target or navigation source rather than a computer randomly losing control.
- It will not engage: the aircraft may be badly out of trim, outside an engagement limit, receiving a manual control input, missing a valid sensor source or reporting a flight-control fault. Stabilise and trim first, then follow the aircraft checklist rather than repeatedly pressing the button.
- It turns away from the route: heading mode may still be active, the wrong navigation source may be selected, or the flight plan may contain a discontinuity. Check the active lateral mode and the displayed course.
- It levels off unexpectedly: the autopilot may have captured the selected altitude exactly as commanded. Verify the altitude window, vertical mode and any programmed altitude constraint.
- Airspeed rises or decays: the autopilot may be controlling pitch while thrust remains manual. Check power, configuration and the active pitch or speed mode before the aircraft approaches an unsafe energy state.
- It oscillates or disconnects: excessive control forces, turbulence, sensor disagreement or a system fault can trigger a disconnect. In a simulator, conflicting controller axes, noisy hardware inputs, accelerated simulation rates or poorly tuned add-on logic can produce similar symptoms.
Simulator users facing engagement problems can follow our focused Microsoft Flight Simulator autopilot engagement checks. In a real aircraft, the safe response to unexplained automation is to maintain control, disconnect and hand-fly if necessary, trim the aircraft, and use the approved checklist.
An autopilot follows commands; it does not judge whether those commands are sensible. Pilots must confirm the active modes, monitor flight path and airspeed, anticipate the next capture or transition, and intervene whenever the automation is not doing what was intended.