How Do Drones Hover? GPS, Barometers, and Optical Flow Explained

Lift Equals Weight

A drone hovers when the total lift generated by its four spinning propellers exactly equals the downward pull of gravity. The propellers accelerate air downward (Newton's third law: air pushed down, drone pushed up), and if the thrust matches the drone's weight, vertical position stays constant. This sounds simple, but maintaining this balance in real conditions requires constant adjustment. Wind, air density changes, battery voltage sag, and shifting center of gravity all disturb the equilibrium.

The flight controller maintains hover by running a Proportional-Integral-Derivative (PID) control loop hundreds of times per second. It reads sensor data, calculates how far the drone has deviated from the target position, and adjusts all four motor speeds simultaneously to correct the error. Different motors speed up or slow down by tiny amounts to correct roll, pitch, yaw, and altitude simultaneously.

Counter-Rotating Propellers: Why Quadcopters Use Four

A single spinning propeller creates thrust but also creates torque that would spin the drone body in the opposite direction. Quadcopters solve this with two pairs of counter-rotating propellers: motors 1 and 3 spin clockwise, motors 2 and 4 spin counter-clockwise. The torques cancel each other out, keeping the drone stable. Yaw control works by adjusting the speed balance between the two pairs: spin the CW motors faster and the drone yaws counter-clockwise, and vice versa.

Why Hovering Uses More Power Than Slow Forward Flight

When a drone hovers stationary, each propeller operates inside the downwash of its own previous blade passes. This turbulent recirculation, called blade vortex interaction, increases the induced power required to maintain thrust. Moving forward at 2-5 m/s takes the drone out of its own downwash and into undisturbed air, reducing the power required to stay airborne by roughly 15-20%.

Ground Effect: How Low Altitude Changes Hover Efficiency

Within approximately one rotor diameter of the ground (roughly 0.3-0.5 meters for most consumer drones), hovering actually becomes more efficient. The ground interrupts the downward airflow expelled by the propellers, redirecting it outward rather than allowing it to recirculate. This reduces blade vortex interaction and gives the motors a lift boost of around 5-10%. Pilots notice this as a tendency for the drone to rise slightly during the final descent before landing. It is not a malfunction, just air pressure building between the rotors and the surface. The effect disappears above one rotor diameter of altitude.

GPS Position Hold Accuracy

The drone's GPS module receives signals from multiple satellites and calculates its position using triangulation. Consumer GPS modules using a single frequency (L1) achieve horizontal accuracy of approximately 1.0 to 1.5 meters under good conditions. This means the drone may drift up to 1.5 meters in any direction from its locked position and still be within GPS accuracy. It will not appear to drift to the pilot's eye at range, but up close the drift is visible.

GLONASS dual-band GPS (available on mid-range and premium consumer drones) improves horizontal accuracy to approximately 0.3-0.5 meters by adding Russian satellite constellation data to the calculation. RTK GPS used in commercial mapping drones reaches centimeter-level accuracy, but requires a ground base station and is not found in consumer drones.

GPS Altitude Is Less Accurate Than a Barometer

A common misconception: GPS is not the primary altitude sensor. GPS triangulation calculates altitude (vertical position) much less accurately than horizontal position, with typical vertical accuracy of 3-5 meters under good conditions and potential drift of up to 15 meters under poor signal. The barometer handles altitude hold. GPS handles horizontal position. Both systems run simultaneously but serve different axes.

Hover Accuracy by Drone Price Tier

What Happens When GPS Drops Mid-Hover

When GPS signal drops during a hover, consumer drones (DJI, Autel) switch to ATTI (attitude) mode automatically. In ATTI mode, the barometer holds altitude but the drone has no horizontal position reference. It will drift with the wind. The pilot must take manual control and use visual reference to hold position. If the drone is far away or the wind is strong, maintaining position manually in ATTI mode is difficult. This is one reason flying near GPS-disrupting structures (metal rooftops, bridges) can be dangerous.

How the Barometer Holds Altitude

The barometer measures air pressure and converts it to altitude using the fact that pressure decreases predictably with height. At sea level, standard pressure is 1013.25 millibars. Every 8.5 meters of altitude gain corresponds to a pressure drop of approximately 1 millibar. The flight controller monitors this pressure continuously and adjusts motor speeds whenever the pressure reading indicates a change in altitude.

Barometric altitude accuracy under calm conditions is approximately 0.5-1 meter, which is significantly better than GPS vertical accuracy. However, barometers have a critical weakness: sudden wind gusts cause brief pressure spikes that the sensor misreads as an altitude drop. The flight controller overcorrects by climbing, then corrects back, producing the altitude oscillation visible in many drone hover videos shot in moderate wind.

Optical Flow: How Drones Hover Indoors

Indoors, GPS signals are too weak or unavailable. Optical flow sensors fill the gap. A downward-facing camera photographs the ground at high speed, and the flight controller analyzes the frame-to-frame shift in the ground texture to detect horizontal drift, similar to how a computer mouse tracks surface movement. When drift is detected, the motor corrections are applied to cancel it.

Optical flow has firm limitations:

• Effective range: approximately 6-9 meters AGL (about 20-30 feet). Beyond this, ground texture details are too small for the camera to track reliably.
• Fails on uniform surfaces: sand, snow, calm water, glass floors, and white tile all lack the texture features optical flow needs.
• Low-light failure: optical flow cameras need adequate lighting to resolve ground texture. Dark environments cause complete position hold failure.
• Wind: optical flow cannot compensate for wind that pushes the drone horizontally, because it only detects drift relative to the ground below, not airflow.

The Extended Kalman Filter: How All Sensors Work Together

No single sensor is accurate enough for stable hover on its own. The flight controller uses an Extended Kalman Filter (EKF) to combine all sensor inputs into a single best-estimate of the drone's position and orientation. The EKF weights each sensor based on its known accuracy and current reliability: if GPS is strong, it gets high weight for horizontal position; if GPS is weak but barometer is stable, altitude gets weighted more toward the barometer. The gyroscope and accelerometer feed into the EKF constantly as high-frequency attitude references that fill the gaps between slower GPS and barometer updates.

How Wind Affects Hovering

GPS position hold fights wind by detecting drift from the locked coordinates and applying corrective thrust. But the drone can only fight wind up to its rated wind resistance, which varies by model. The DJI Mini 4 Pro is rated for up to 38 kph (Level 5 Beaufort) in forward flight but noticeably struggles in hover at sustained 30 kph crosswinds. The DJI Air 3S and Mavic 4 Pro handle stronger winds in hover due to their larger motors and heavier airframes providing better stability.

A practical rule: fly at no more than two-thirds of the drone's rated wind resistance for comfortable hovering. If a DJI Mini 4 Pro is rated for 38 kph, plan hover sessions in conditions under 25 kph for reliable position hold.

Wind and Battery Drain During Hover

Every wind gust requires a motor correction. In steady light wind (under 15 kph), the corrections are small and battery impact is modest. In gusty conditions (wind speeds varying ±10 kph every few seconds), the motors spike their current draw repeatedly, increasing overall power consumption by 20-40% compared to calm-day hovering. A 34-minute battery drone on a calm day may deliver 20-22 minutes in gusty 25 kph wind while hovering.

How Long Can a Drone Hover?

Hover time is limited by battery capacity. Consumer drones are rated for 20-34 minutes of flight time, but this figure is measured under controlled conditions, typically in slow forward flight. Real hover time in calm conditions will be 2-4 minutes less than the spec, and significantly less in wind. The barometer and GPS continue to hold position as long as battery voltage stays above the drone's minimum threshold, at which point low battery warning triggers and Return to Home activates.

Why FPV Drones Cannot Hover Autonomously

FPV drones running Betaflight firmware in Acro (manual) mode have no GPS, no barometer-based altitude hold, and no optical flow. The flight controller handles only attitude stabilization (leveling) and motor mixing. Hovering requires the pilot to manually hold the throttle at exactly the right level to maintain altitude while simultaneously correcting for any horizontal drift with the other sticks. This is a real skill that takes weeks of practice to develop reliably.

The reason FPV pilots use simulators extensively is precisely because of hover: learning to hold position manually at a consistent altitude without GPS feedback is the fundamental skill that separates FPV pilots from GPS camera drone pilots. FPV pilots who practice this for 10-20 simulator hours before their first real flight crash significantly less than those who skip the simulator.

iNav: Adding GPS Hover to FPV Drones

FPV drones running iNav firmware instead of Betaflight can add GPS position hold. iNav supports GPS modules, barometers, magnetometers, and optical flow sensors, and provides a Position Hold flight mode functionally similar to GPS camera drones. The setup is more involved than plug-and-play consumer drones: the pilot must tune the iNav configuration and calibrate all sensors, but the result is a drone that can hover in place autonomously when switched into the appropriate mode.

The Skill Value of Manual Hover Practice

GPS camera drone pilots who practice manual hover on an FPV simulator or a toy drone without GPS develop better spatial awareness and stick control than those who rely exclusively on GPS position hold. Manual hover forces the pilot to read the drone's attitude, anticipate drift, and apply corrections proactively rather than waiting for GPS to correct automatically. This skill becomes directly useful in ATTI mode emergencies, in GPS-denied environments, and when obstacle avoidance is disabled in Sport mode.