Hi; I have a waypoint mission with 2x POIs. It’s a video, and the drone flies backwards, up, and away. There’s a “jerk” at WP2, and also at waypoint 3 (where it transitions from focusing on poi1 to poi2).
How do I smooth those out? I’m uploading the raw video as a private youtube video so you can take a look at what I mean…I’ll post the link when it’s ready.
(How do I upload the csv and klm exported missions to this topic?)
Thanks for your help!
Search the forum for “jerky” and watch the video in this thread:
Here’s the link to the video: https://youtu.be/CNKYmIeczcA
Use this info to share your mission.
Thanks all; I had to run out and do a flight right after I posted but I’m back now.
Here’s a link to the mission, and I’ll watch the suggested video now.
Oh, and my bad: I’m flying a Mavic 2 Pro.
Unexpected stop-and-go motion that generates a “sawtooth” flight path in any drone equipped with obstacle avoidance sensors such as your Mavic 2 Pro, is very often the result of neglecting the need to calibrate those obstacle sensors, especially if the drone has been relocated some miles since its most recent flight.
If, like me, you prefer to disarm your M2P’s obstacle avoidance sensors to conserve flight battery power and prevent false triggering at low sun angles, you might wonder why one would need to laboriously calibrate those sensors in all directional axes using DJI Assistant 2 for Mavics checkerboard calibration screen, since they are going to be de-selected for the flight.
While I still cannot answer that legitimate and logical question as to why sensors whose use is optional still require calibration, I have learned by trial and error, after losing one Mavic 1 Pro for precisely this reason, that the obstacle avoidance sensors on ALL of DJI’s drones MUST be maintained in a state of perfect calibration at all times, regardless of whether those sensors are actuated for any flight or de-selected.
Even if my hypothesis turns out be wrong this time, there is no harm done by periodically recalibrating those obstacle avoidance sensors because drones do fly in a noticeably smoother fashion after that key calibration has been completed using DJI Assistant 2 for Mavics.
Interesting! I’ll have to try this when I get home from work today. Thanks for the heads up!
On a completely unrelated note, I recently and belatedly learned that the battery retention clips of the Mavic 1 Pro and Mavic 2 drones have a hidden design flaw that can result in the drone’s outright loss mid-flight unless the flight battery is secured in place with a zip tie.
Below is a picture of my trusty Mavic 2 Pro in formal attire complete with that all-important battery retention zip tie.
The risk of battery detachment increases when a Mavic 1 Pro or Mavic 2 Pro has been flown for a long duration or in hot weather, so affixing that battery retention zip tie would be mandated by drone law if I were ever to find myself enthroned as Global Drone Despot.
What. Is. That.
???
I want it!! Dual battery? How long can you fly on that thing? Probably not 2x as long, but maybe 1.75x as long?
Seems like velcro might be a good choice too, rather than having to fiddle with undoing a zip tie each time the batter (finally) needs to be changed?
BTW, Sunday I did a flight at a different site and needed to calibrate the compass there. Today I did the above flight, and while I haven’t looked at the actual video yet, it sure looked like it didn’t stumble / jerk at the different waypoints this time. I’ll know for sure later tonight.
16mph winds today, supposedly up to 21mph gusts. This mavic 2 is quite the machine, handled it with aplomb! Still not as stable as my old Autel XSP, but it’ll do. 
It is good to hear that sensor calibration may have made the uncommanded stops you’d seen during earlier flights of the M2 Pro less apparent, Bill. Remember that the Mavic 2 Pro variables normally calibrated outdoors are the compass and IMU, while the obstacle sensor calibration is done indoors using DJI Assistant 2 for Mavics whereby the drone with folded arms is oriented towards a checkerboard screen and pointed to match the on-screen sequence of squares as they change color on the checkerboard display.
The hump-back extra range battery in the picture above is from Ali Express, but I couldn’t find the direct link so I just copied the order page from my last purchase. These extra-range batteries are DIY creations most likely produced by a tinkerer, but that said, they have proved to be reliable improvements over the original OEM batteries from DJI.
Residing out here in the Third World boonies where FAA regulations are a distant curiosity, I launch my Mavic 1 Pro and Mavic 2 Pro on fully autonomous Litchi waypoint missions that traverse over 12 miles round trip with the cruise speed set at 27 mph, before landing with between 15% and 20% battery charge level remaining.
For these larger batteries, those zip ties are even more crucial due to the added weight of the battery and the stress that weight places on the drone’s stock battery retention clips. An alternative to velcro is re-useable zip ties that needn’t be cut off after each flight.
Thanks @Mad_Pup!
I just had a chance to look at the footage, and…it’s still jerking and slowing. So I looked into the calibration you mentioned and I’m certain that I’ve not done before. (I bought this bird a few months ago.) I’m looking forward to trying it out! Probably won’t have a chance again until next Tuesday or Wed, but I’ll updated this thread with the results when I do!
I didn’t bother learning this procedure until those random stops began to make my drone’s every flight a nail-biter. This sensor calibration is a tedious procedure that can be thwarted if a poor-quality USB cable is used to connect the drone to the laptop, but with repetition, it becomes an easier and faster process to complete.
Finally, there are multiple versions of DJI Assistant 2 and it is the version called DJI Assistant 2 For Mavics that is compatible with the Mavic 2 Pro.
Your “problem” has nothing to do with the vision sensors.
It’s caused by a high shutter speed.
Make sure your shutter speed is double that of the frame rate by increasing the aperture and/or using ND-filers.
The drone stops/slows down at WP4 because it’s curve size is set to zero.
Either increase the curve size or delete WP4 all together.
I don’t understand. You’re saying that camera settings are causing the hard/sudden yaw at :07 in the video, which is waypoint 2, and then again at :15 in the video, which is waypoint 3? How can camera aperture affect flight characteristics?
Or are you saying that the camera settings is just making the video “look” like it’s a flight characteristic issue?
I’ve changed wp4 to a curve, thanks for that.
Ha! I totally forgot to mention that I have used both versions of the assistant 2 (first one for my mavic pro, newer version for the mavic 2 pro), so I started it up and went thru the entire calibration process. Successfully.
So I’m looking forward to my next flight!
No.
In my experience with litchi, about 8 years now, that’s normal behaviour.
Perhaps that’s why I didn’t notice it.
What I see is a stuttering yaw movement BETWEEN WP1 & WP2 and between WP2 & WP3. That’s caused by a high shutter speed.
You could try to ignore the 2 POIs.
Enter the corresponding Heading and Gimbal Pitch values directly in the waypoint settings and set Gimbal Pitch to Interpolate and POI to None, at WP1, 2, 3 and 4.
All you have to do at each WP is:
-Change Gimbal Pitch to Interpolate
-Set POI to None.
Gotcha. It’ll be easy enough to copy that flight and change to interpolated with set gimbal pitch and aircraft yaw, I’ll give that a try.
Come to think of it, I do a perimeter flight of that entire property and I use the non-poi method and I get no harsh gimbal movements, so maybe that’s it.
Thanks for the ideas! I’ll update you all after next Tuesday’s flight.
Hi 
Nice shot. In my opinion, the linear speed of your drone is too high. In my opinion :0) in m/s it’s 11.8 too much at point 1, and then you go to a general 7.8 m/s. You make an arc from 1 to 2, that’s about 30 m, and from 2 to 3 it’s 60 m. I would fly from the beginning at up to 3 m/s and only the straight from point 3 would be faster. Best regards.
