Interpolation on waypoint mode - Litchi

If you use Focus POI first, and look at the heading angle and gimbal angle, then change to interpolate and set the heading and gimbal to the same angles as Focus POI automatically calculated for you, then when you are AT that waypoint the orientation of the camera will be exactly the same as it would have been if you left it as Focus POI. Further, the POI can be deleted from the mission, and THAT waypoint will still point right at the point in space where the POI was originally before you deleted it.

Perform this same task for a group of waypoints -- i.e. set a POI, set the 5 waypoints to "Focus POI", go to each individually, wriite down the heading and gimbal angles, change waypoint setting to Interpolate, set heading and gimbal to be the same as calculated by Litchi with Focus POI -- and camera behavior for those 5 waypoints will be exactly the same as if you'd simply left the POI on the map and the 5 waypoints as "Focus POI".

The firmware in the aircraft doesn't have a concept of POIs. Rather, it can be programmed to go to a particular lat/long at a particular altitude, point the heading in a particular direction and the gimbal at a particular angle.

That's all the aircraft knows about in terms of what gets uploaded to it. It doesn't have any clue how those angles were determined before they were uploaded to the aircraft.

The Focus POI and Interpolate settings function purely in the pre-planning phase. When the mission is flying, it's simply going from one location to another, and aiming the drone and gimbal at the specified angles that were uploaded.
Understood. My thinking was more around what happens between waypoints. That's the interpolation region. Not to bore everybody but for grins, here was my thinking: consider a point half way between waypoints. I see two ways to calculate the gimbal angle. The drone has an elevation, the POI has an elevation and distance. A simple trig calculation would yield the angle. Interpolation says just take half the difference between the angles at the two points. One could think that they might yield different results and that the former is the true angle. I'm not good at mathematical proofs but I'm sure there is one to show that the results are actually the same. Intuitively I guess that since it's a straight line and the altitude is increasing linearly then the angle decreases linearly then they must yield the same results. I know I'm making it more complicated than necessary. Just explaining why I thought the two modes might yield different results
 
From the book.

"Interpolate" allows you to set the tilt angle for a waypoint as a parameter.


"Focus POI" is just interpolate with Litchi handling setting the waypoint angles for you.
Actually I missed this in the manual. That could have saved me a lot of unnecessary thinking about 11th grade math. ;)
 
Understood. My thinking was more around what happens between waypoints. That's the interpolation region. Not to bore everybody but for grins, here was my thinking: consider a point half way between waypoints. I see two ways to calculate the gimbal angle. The drone has an elevation, the POI has an elevation and distance. A simple trig calculation would yield the angle. Interpolation says just take half the difference between the angles at the two points. One could think that they might yield different results and that the former is the true angle. I'm not good at mathematical proofs but I'm sure there is one to show that the results are actually the same. Intuitively I guess that since it's a straight line and the altitude is increasing linearly then the angle decreases linearly then they must yield the same results. I know I'm making it more complicated than necessary. Just explaining why I thought the two modes might yield different results
Gotcha. The hardware doesn't work that hard, doing all those calculations.

Rather, it knows the distance between the waypoints, its speed, so it knows how long it will take between the two points. It simply calculates a gimbal tilt rate based on those factors and the difference in the tilt angles, then moves the gimbal at that speed while it transitions between the waypoints. This will result in what you discussed above -- half-way between gimbal angles halfway between waypoints.

Think about the tilt dial on the left side of the controller... The more you push it, the faster it tilts. It's a proportional tilt rate control.

This is implemented as a speed value somewhere in the programming API for the drone, and that's the same way it's controlled under autonomous flight -- a rate is calculated, and the appropriate API function is called with the speed value.

You see here again there is no POI in the picture. The P4 doesn't know anything about POIs.

The "proof" of all of this is the geometry of the whole affair when you assume constant speed and climb rate between waypoints, which is how it works :)
 
Last edited:
Gotcha. The hardware doesn't work that hard, doing all those calculations.

Rather, it knows the distance between the waypoints, its speed, so it knows how long it will take between the two points. It simply calculates a gimbal tilt rate based on those factors and the difference in the tilt angles, then moves the gimbal at that speed while it transitions between the waypoints. This will result in what you discussed above -- half-way between gimbal angles halfway between waypoints.

Think about the tilt dial on the left side of the controller... The more you push it, the faster it tilts. It's a proportional tilt rate control.

This is implemented as a speed value somewhere in the programming API for the drone, and that's the same way it's controlled under autonomous flight -- a rate is calculated, and the appropriate API function is called with the speed value.

You see here again there is no POI in the picture. The P4 doesn't know anything about POIs.

The "proof" of all of this is the geometry of the whole affair when you assume constant speed and climb rate between waypoints, which is how it works :)

Excellent info! Thanks!
 

Recent Posts

Members online

No members online now.

Forum statistics

Threads
143,099
Messages
1,467,634
Members
104,985
Latest member
DonT