# Flying a drone in a moving car/airplane

Why does a drone fly with respect to the air in a moving car/airplane, but not in an elevator as shown in this clip: What Happens If You Fly a Drone In An Elevator?

I understand that the drone will move at the same speed as a moving car/airplane. Therefore, you'd have no trouble controlling it. However, according to the clip, in the elevator the drone stays in the same position and not with the elevator. Thus, if you want the drone to fly inside the elevator, you have to fly it upward if the elevator is coming up and vice versa.

• Worth reading up on: Gallilean invariance: en.wikipedia.org/wiki/Galilean_invariance
– Gert
Commented Aug 31, 2020 at 19:25
• Warning : Don't do this experiment in a car at high speeds because if there is sudden braking the drone will be slammed against the windshield. Commented Sep 1, 2020 at 14:47
• It depends on how you drive the car, or fly the airplane. Try it (as a passenger) in a sports car on a twisty road, or a plane doing some aerobatics, and you'll understand :-) Commented Sep 1, 2020 at 15:58
• I expect that GPS is involved when the drone is holding a constant height above the floor. This means that it "falls" to the floor of the elevator when the elevator is going up, and it hits the ceiling when the elevator is going down, because the drone is holding a constant height and the elevator is moving around it. Commented Sep 1, 2020 at 18:53
• @DavidWhite: No, it's nothing to do with GPS. It is simply because the elevator is accelerating. Commented Sep 2, 2020 at 0:15

## 6 Answers

Very good question! The point is that when the elevator begins to move (either upwards or downwards), it's accelerating, while the drone -- having no force acting on it directly -- is still moving at a constant velocity.

Imagine you had a drone in a moving car or airplane, and it was "hovering" next to you. This means that its velocity with respect to you (and the car or airplane) is zero. If the car is moving at a constant velocity, the drone will continue to stay where it is, since no external force is acting on it.

On the other hand, in the elevator, the drone is at rest with respect to the "building". But when the elevator "starts", it begins to accelerate. The fact that the elevator has changed its velocity does not reach the drone as it isn't in contact with the elevator like we are. As a result, the drone continues to be at rest with respect to the building while the top of the elevator accelerates towards the drone, whacking it. Something similar occurs when you start from the ground floor and try to move up. Again the drone is at rest with respect to the ground floor, but when the elevator starts to accelerate upwards, the bottom of the elevator rams at the bottom of the drone.

I don't have a drone to test this, but I assume that something similar would happen in a moving vehicle if the vehicle suddenly accelerated (or turned). It's very similar to the feeling you experience when a car suddenly accelerates: you're slammed back because part of your body hasn't got the information yet that the velocity of the vehicle has changed.

If I had a drone (and a long enough elevator), I would try to show that once the elevator begins moving at a constant velocity (which happens after the initial jerk) I could control the motion of the drone without any trouble.

• Thank you! Regarding your last paragraph, if I want to be able to control the drone without trouble, does that mean I should probably wait a few seconds after the elevator has started for it to reach its constant velocity, then start up the drone? If it's reached constant velocity, like in a moving car/plane without interruptions from acceleration, I should be able to control it normally? Commented Aug 31, 2020 at 18:56
• Yep. In that case both you and the drone will be at rest relative to each other and to the walls of the elevator (until it reaches another floor and decelerates). I wish I could try it out :) Commented Aug 31, 2020 at 19:00
• ... until the elevator stops, anyway. Commented Sep 1, 2020 at 15:22
• @evildemonic: No, that is wrong. An elevator moving with constant speed has no effect on the weight of the objects inside it. It is only the acceleration and deceleration phases that affect weight. Commented Sep 2, 2020 at 0:17
• @evildemonic: No, you only feel lighter while the elevator is accelerating downwards. Once it reaches its constant downward speed, your weight returns to normal -- until it starts slowing down, when you briefly feel heavier than normal. (And by the way, the Vomit Comit is continuously accelerating or decelerating during the passengers' period of weightlessness.) Commented Sep 2, 2020 at 15:20

A drone does not always fly with respects to the air in a moving car/airplane. It does so only if they are not accelerating. The elevator is accelerating in your video.

The drone is responding to electronic accelerometer inputs on the device, and is simply a visual demonstration of the feedback loop in action. It is trying to hover in one spot, so when it suddenly moves up/down from that spot when the volume of air it is hovering in gets moved, its on-board controller tries to move it back to that spot.

It has nothing to do with airflow inside the elevator, something you could visually demonstrate by holding a child's toy windmill in the elevator. If the windmill stays still during ascent/descent, you know that the air is not moving (by anything your crude "measuring device" can detect).

• A typical drone has no position sensor. It detects only accelerations and rotations. So no, it is not trying to "hover in one spot" rather it is trying to null accelerations and rotations. Position holding requires an externally referenced input, such as a GPS. Commented Sep 1, 2020 at 17:06

The drone flies with respect to the air right there at the drone. The air meanwhile is in whatever state of motion is imposed on it by the forces acting on it. As a car or an elevator starts to move in some new direction, the air inside will not all immediately adjust. Rather, the air near the walls gets pushed at first, and this sends a pressure wave through the air, until eventually all the air is in motion, whether horizontally or vertically. The drone meanwhile 'rides' the air that is near to it and gets pushed around accordingly.

• But pressure waves in air are moving pretty fast, usually with the speed of sound.
– Uwe
Commented Sep 2, 2020 at 12:25
• @Uwe The pressure wave sets up the pressure disturbance; the bulk flow of the air then responds more slowly. Commented Sep 2, 2020 at 13:33
• Do you claim the air movement in the elevator is the leading effect, and inertia of the drone itself itself is less important? I would think the opposite. Commented Sep 2, 2020 at 20:35

Simply put, if the container (car, boat, plane, OR elevator) you are in is moving at a constant speed, then you won't feel any push and neither will the drone. It will fly nice and stable.

On the other hand, if the container is accelerating (a car speeds up, a boat is on rough seas, a plane does acrobatics, OR an elevator starts rising or falling) then you will feel that acceleration relative to the prior rest state. Likewise, your friendly drone will have a hard time staying in place and is more likely to crash into something.

Drones use a variety of sensors to measure their movement and change in movement relative to their environment. These sensors typically include inertial sensors (accelerometers, gyroscopes), GPS sensors, or magnetic sensors (compasses). They all evaluate the position of the drone with respect to a "fixed" earth reference. If you want/need to keep the drone stable in the elevator (or any of these other environments), it will need sensors that are not focused on stability with respect to earth, but instead with respect to that local environment.

Examples could include ultrasonic or laser rangefinders to measure distances to the walls and floor, video sensors to measure changes with respect to the local environment using some form of machine vision, or possibly transponders in different parts of the environment that would allow time/distance calculations for position determination. It is unlikely that you could keep a drone stable for any length of time in an elevator because they tend to be either slowing down or speeding up for a good part of their trip.

You have some good, correct answers here. So far none of them have addressed the point that when an elevator starts moving upwards you feel heavier, and when it starts moving downwards you feel lighter. This is because the force of its its acceleration is added to (upward acceleration), or subtracted from (downward acceleration) the force of gravity. When you feel heavier as the elevator starts upward, so does the drone. When you feel lighter as the elevator starts downward, so does the drone. So a suddenly "heavier" drone will lower, and a suddenly "lighter" drone will climb. Hope this will give you a better idea of the actions.

• I do not think this argument is correct, but perhaps I've misunderstood it. You are right about us feeling lighter and heavier as the elevator starts moving downwards or upwards respectively but that's because we're in contact with the elevator. The hovering drone does not sense this either until the air starts moving around it or the roof or floor of the elevator reaches the it. If your argument of "heavier" objects lowering were true, then under a constant acceleration the drone should attain a new equilibrium, and not slam against the floor of the elevator... Commented Aug 31, 2020 at 22:10
• Another reason this explanation seems wrong: if we took a drone which was carrying a 1 kg weight, accelerated the elevator to make the drone "heavier" by 1kg (ignoring for now that this wouldn't be the case as @Philip pointed out), and simultaneously removed the 1 kg weight, then by this answer's explanation we ought to expect the drone to remain still despite the acceleration. Commented Sep 1, 2020 at 13:01
• @JBentley Do you understand Einstein's postulates about frames of reference? If you have a test to prove it is acceleration and not increased gravity then you have proved him wrong. Removing weight from a drone suddenly will make a drone rise in most any situation. Commented Sep 1, 2020 at 15:07
• Or just look at it from a car example, where it's clearly not about the drone feeling heavier or lighter. Whether it's a elevator or a car is irrelevant to the drone, as @JBentley says it's just surfaces accelerating towards or away from the drone.
– eps
Commented Sep 1, 2020 at 20:33
• @AdrianHoward No, it is not inside the accelerating frame of reference. The accelerating frame of reference is the floor of the elevator. The drone is not on the floor, it is in a region of space above the floor. I have already given this analogy, but I will repeat it as you seem to have ignored it. If you remove the ceiling and walls of the elevator, you are left with a platform. According to you, a drone flying above a platform which accelerates up and down in an oscillating manner, will gain and lose weight accordingly. That is quite clearly not true. Commented Sep 2, 2020 at 13:20