Fly-to-Waypoint in Maven

[AZDave]

One of the benefits of the fly-to-waypoint feature in Maven (where you set waypoints for a mission by manually flying to each location and saving it as a waypoint for a later automatic flight) is the precision you can get. It's better than relying on a map view for location and elevation. It's a great feature and works perfectly ... unless you take off from a different location or don't wait for the drone to establish your takeoff point. I did manage to carelessly mess it up, though.

I manually flew to several different locations over terrain with a lot of elevation changes, added them as waypoints, and subsequently flew a complete mission without problems. However, I had neglected to set the drone orientation as "point to next waypoint" for each waypoint in my saved mission, which made the drone kind of slide sideways through some of the turns and I didn't like the resulting video. It was an easy fix to modify the mission to make the waypoints point to the next waypoint, and while I was at it I decided to increase the radius of curvature for the turns around some of the waypoints to make them smoother. When I flew the modified mission I almost plowed my Mavic Air 2 into a large tree, having neglected to consider that changing the radius of curvature would necessarily shift the path of the drone sideways. I had trashed the precision of the original path.

Luckily, the combination of me quickly reacting to the in-flight video feed to my iPad, as well as the drone's own collision avoidance system, averted disaster. I learned a good lesson without a lot of pain so I guess I should call that a win.

And if this post helps anyone else avoid a similar mistake, so much the better.

 

[GFields]

thanks for sharing!

 

[Claretcass]

thanks for sharing!

Thanks for that.
This coupled with the “above ground” waypoints option makes it a very powerful app. It just gets better.

 

[AZDave]

Thanks for that.
This coupled with the “above ground” waypoints option makes it a very powerful app. It just gets better.

Yes, I am super impressed with what Michelle has done with Maven and I really love the app.

One thing to be careful of, though, is that Maven can only be as accurate for elevation-above-ground as the data upon which it is based. In answer to my question on his YouTube video, Michelle told me that he uses Google Earth plus a couple of other sources for terrain elevations. I don't know about those "other sources", but Google Earth around my home has some limitations. There is a pretty deep ravine that runs across my property (I live on the side of a mountain range in southern Arizona) that is about 150 feet wide and probably 40 feet deep, but Google Earth tells me that is it only about 15 feet deep relative to the surrounding terrain.

There are other sharp ridges that I typically fly across that have the opposite problem. Google Earth understates their height ... understandably so. All of this is in spite of the fact that the camera resolution for Google Earth is excellent in my area. I can pick out really small objects with pretty good clarity. Other areas with worse resolution are probably even more uncertain for elevation accuracy.

Depending upon the terrain involved, when running a pre-programmed mission using height-above-ground it might be prudent to first run the mission with higher more conservative elevations and with the camera angled downward to evaluate the path. Then if all looks OK on the saved video, reduce the elevations for the next run. The same might even hold for fly-to-waypoint saved missions, since the mission is going to fly from waypoint to waypoint and any intervening structure that you manually avoided might not be avoided during the automatic mission.

Lots of stuff to keep in mind in addition to my experience with careless mission edits.

 

[Claretcass]

Yes, I am super impressed with what Michelle has done with Maven and I really love the app.

One thing to be careful of, though, is that Maven can only be as accurate for elevation-above-ground as the data upon which it is based. In answer to my question on his YouTube video, Michelle told me that he uses Google Earth plus a couple of other sources for terrain elevations. I don't know about those "other sources", but Google Earth around my home has some limitations. There is a pretty deep ravine that runs across my property (I live on the side of a mountain range in southern Arizona) that is about 150 feet wide and probably 40 feet deep, but Google Earth tells me that is it only about 15 feet deep relative to the surrounding terrain.

There are other sharp ridges that I typically fly across that have the opposite problem. Google Earth understates their height ... understandably so. All of this is in spite of the fact that the camera resolution for Google Earth is excellent in my area. I can pick out really small objects with pretty good clarity. Other areas with worse resolution are probably even more uncertain for elevation accuracy.

Depending upon the terrain involved, when running a pre-programmed mission using height-above-ground it might be prudent to first run the mission with higher more conservative elevations and with the camera angled downward to evaluate the path. Then if all looks OK on the saved video, reduce the elevations for the next run. The same might even hold for fly-to-waypoint saved missions, since the mission is going to fly from waypoint to waypoint and any intervening structure that you manually avoided might not be avoided during the automatic mission.

Lots of stuff to keep in mind in addition to my experience with careless mission edits.

thank you for your comment.
after flying ”above ground” many times with Litchi on uneven terrain and then importing via CSV into Maven and manually altering the altitudes to match the Litchi calculations then testing it became apparent that the altitude using Google earth could from 10’ up to 30’ out when flying the waypoints. I now always set the first waypoint 40’ higher and then home in to a lower altitude after flying the waypoints and manually observing, better safe than sorry.
For what I want as a recreational flyer the Maven app is perfect and meets my needs completely And I am sure that once (if) the MInit 2 SDK is released Michelle will be one of the first software developers to support it And if he ever gets the time to develop the Android version it would be absolutely brilliant as the mini 2 is now on the smart controller but I appreciate this would be a massive task for him and maybe simply too much to take on Because with this app he would enter the big league, we will see.