3 ms·

Actually it is not easier at all, the complexity is obscured by an onboard computer, which constantly helps to keep position, not to flip upside down and lose c

by nuccy 6y ago

Actually it is not easier at all, the complexity is obscured by an onboard computer, which constantly helps to keep position, not to flip upside down and lose control completely. By applying differential thrust on different propellers it executes commands from operator to tilt and rotate. If the computer fails, operator, even with a direct control of the propellers, will not be able to maintain a stable flight. From aerodynamic standpoint planes are much easier to pilot (all but military, same story with computers for better maneuverability) since they are designed to be inherently stable. So if all power is lost and you still have control of aerodynamic surfaces, there is not problem to land a plane after planning for some tens or even hundreds of kilometers (depends on initial altitude). This is obviously not the case with quadropters, loss of a single propeller leads to complete loss of control.

mikepurvis 6y ago
It leads to loss of six-dimensional control, though interestingly you can still maintain five-dimensional control (losing yaw) with only three or even just two props, see demonstration from ETH Zurich: https://www.youtube.com/watch?v=t369aSInq-E https://www.youtube.com/watch?v=t369aSInq-E
- nuccy 6y ago
  Yes, indeed, though this seems to be an experimental drone. Most common commercially available DJI drones fall if single motor or propeller fails, though in future they may gain such functionality, which would be great for safety. A nice example of a different approach is a variable angle of attack propellers which are driven by a single motor [1]. See after 0:52 and 7:13. In case battery is dead since all the propellers are synchronized they auto-rotate and stabilize the drone (near the earth the pilot activates the motor just to land softly, but likely similar effect can be reached by properly adjusting the angle of propellers in proper time, like in case of auto-rotation landing for helicopters). [1] https://youtu.be/TnGhEInTXYc?t=52 https://youtu.be/TnGhEInTXYc?t=52 (this video was posted below by @jcims)
  - mikepurvis 6y ago
    Oh yes, it's for sure an experimental thing, and probably patented out the wazoo. My point is just that it's physically possible, as I don't think that is well known. (Note also that doing this outside of a mocap arena would probably require a pretty high-quality IMU since you need very accurate yaw and you're constantly integrating your gyro to get it, with only the magnetometer and maybe the camera providing any kind of absolute reference.)
asutekku 6y ago
For the end user it’s easier and that’s what matters. You can just hover the drone, but you need to constantly look after an RC plane.
giomasce 6y ago
I know, but the computer just has to do local corrections, i.e., varying a little bit the input on the propellers to keep the drone stable. On the other hand, there is no local correction that enables you to hover a winged plane (it always has to move at a minimum speed in the air to generate lift), so you can only drive them automatically if the computer can globally control your trip, which is much more inconvenient for a human driver. The aerodynamic stability is an interesting topic, but useless here. If you assume your computer is dead, you won't control neither a remote quadcopter nor a remote winged plane.

by nuccy 6y ago

mikepurvis 6y ago

It leads to loss of six-dimensional control, though interestingly you can still maintain five-dimensional control (losing yaw) with only three or even just two props, see demonstration from ETH Zurich: https://www.youtube.com/watch?v=t369aSInq-E https://www.youtube.com/watch?v=t369aSInq-E

nuccy 6y ago
Yes, indeed, though this seems to be an experimental drone. Most common commercially available DJI drones fall if single motor or propeller fails, though in future they may gain such functionality, which would be great for safety. A nice example of a different approach is a variable angle of attack propellers which are driven by a single motor [1]. See after 0:52 and 7:13. In case battery is dead since all the propellers are synchronized they auto-rotate and stabilize the drone (near the earth the pilot activates the motor just to land softly, but likely similar effect can be reached by properly adjusting the angle of propellers in proper time, like in case of auto-rotation landing for helicopters). [1] https://youtu.be/TnGhEInTXYc?t=52 https://youtu.be/TnGhEInTXYc?t=52 (this video was posted below by @jcims)
- mikepurvis 6y ago
  Oh yes, it's for sure an experimental thing, and probably patented out the wazoo. My point is just that it's physically possible, as I don't think that is well known. (Note also that doing this outside of a mocap arena would probably require a pretty high-quality IMU since you need very accurate yaw and you're constantly integrating your gyro to get it, with only the magnetometer and maybe the camera providing any kind of absolute reference.)

asutekku 6y ago

For the end user it’s easier and that’s what matters. You can just hover the drone, but you need to constantly look after an RC plane.

giomasce 6y ago

I know, but the computer just has to do local corrections, i.e., varying a little bit the input on the propellers to keep the drone stable. On the other hand, there is no local correction that enables you to hover a winged plane (it always has to move at a minimum speed in the air to generate lift), so you can only drive them automatically if the computer can globally control your trip, which is much more inconvenient for a human driver. The aerodynamic stability is an interesting topic, but useless here. If you assume your computer is dead, you won't control neither a remote quadcopter nor a remote winged plane.