Active Track: If I want to go up or down a hill, should I be concerned of a crash or will the P4 adjust it altitude?
Active Track - Elevation change?
- Thread starter KykPhantom
- Start date
- Joined
- Sep 7, 2016
- Messages
- 312
- Reaction score
- 103
- Age
- 48
I don't know for sure for downhill, I seen videos and it did not seem to follow but for uphill it is a feature the P4 has at least as per release notes you can read below:
https://dl.djicdn.com/downloads/phantom_4/en/Phantom_4_Release_Notes_en_20160712.pdf
"Optimized altitude adjustment logic so that the aircraft automatically adjusts its altitude according to changes in terrain while using ActiveTrack (available when the aircraft is flying within 10 meters of the ground)."
Sent from my iPad using PhantomPilots
- Joined
- Mar 30, 2016
- Messages
- 720
- Reaction score
- 183
- Age
- 45
I just tested this out this evening. I stood on top of a small hill and set it to ActiveTrack on me, I walked down the hill and the Phantom followed me and lowered elevation to maintain it's AGL, I then turned around and walked up the hill and it adjusted going back up as well. This is when below 10m, but it is terrain aware and will adjust, both going downhill and up.
- Joined
- Sep 7, 2016
- Messages
- 312
- Reaction score
- 103
- Age
- 48
Brilliant! I always thought this and the rebadged terrain follow feature of the Mavic were the same!
Sent from my iPad using PhantomPilots
Last edited:
I was under the impression that on Active Track, the drone would also adjust altitude to keep the selected object roughly the same size. Did I misunderstand?
When I use active track and walk down a slope, the drone stays at the original altitude. It does not maintain agl. P4 Pro, Android, nVidia Shield K1.
- Joined
- Jan 2, 2017
- Messages
- 610
- Reaction score
- 303
- Age
- 62
You were probably over 10 meters AGL, where the ultrasonic ranging sensors are disabled, so it doesn't know how far above the ground it is, and can't follow elevation changes.
This has really been a feature since the P3. However, it's not design to maintain altitude. It uses the ultrasonic sensors to avoid crashing into the ground. It's far from perfect and, IMHO, you should not be relying on it to maintain altitude. It should be a last ditch effort to avoid crashing into the ground.
- Joined
- Jan 2, 2017
- Messages
- 610
- Reaction score
- 303
- Age
- 62
I disagree completely.
The ultrasonic ranging sensors are very accurate within their usable range. That's why they're there.
So long as you operate within the specified range, they do a very good job of maintaining AGL within inches.
Of course, they're not eyeballs, so there are materials that will screw it up, but this is no reason to be dismissive of the feature -- just know its limitations.
Don't try and use it walking through a foot of soap foam, for instance. Anything that will absorb or scatter ultrasonic frequencies will not work.
Thankfully, the vast majority of material you typically track over (dirt, rocks, asphalt, concrete, etc.) will not have this problem.
It's gotten better with the P4 in that the reaction speeds are slightly increased. However, I've done a lot of testing and it's far from something someone should rely on. I'm looking at this as a general recommendation in general. It's not something that DJI advertises as being a feature of Active Track. From prior posts in the DJI forum, it's not meant to be that either. This does not mean it does not work... it does a pretty good job. But I'd say that it's use is as a last dich effort to prevent a crash. Too each their own.
- Joined
- Jan 2, 2017
- Messages
- 610
- Reaction score
- 303
- Age
- 62
The barometer judges the Phantoms height above the take off point. The Ultrasonic sensor judges the Phantoms height above the ground directly below it at that particular moment. These are TWO DIFFERENT READINGS and don't have anything to do with each other. They server TWO DIFFERENT PURPOSES.
Edit: So you statement is correct but the barometer does not do what the ultrasonics do.They are _separate_ systems. Yes, if you don't understand this you can say it's incorrect. Please don't expect a response from if you want to make that a point.
Let me give you an example of this... If you take off and fly 100' up, the barometer ("H" in the Go app) will report that you are 100' above the take off point. If you lower the Phantom to 10 feet the barometer (again, "H") will read 10 feet. At this time the ultrasonic will also report the Phantom at 10 feet. Now, start (slowly) climbing a hill. The ultrasonic _will_ attempt to increase the altitude of the Phantom in order to maintain that 10 foot height _above the ground under it. If you continue to climb this hill the ultrasonic reading will continue to show 10 feet. If, at the end of this run you are now 50 feet higher then the take off point want to know what the barometer is going to report? Wait for it..... wait for it.... 50 feet! So now the ultrasonic reading will be showing 10 feet and the barometer (H) will be showing 50 feet!
But your statement is that, for some reason, the ultrasonic's over-ride the barometer. As stated and supported above, it simply does not. They are TWO DIFFERENT THINGS. the ultrasonics simply DO NOT take over.
With that said, I've never said that the ultrasonics are not more accurate then the barometer and also never said that it was not accurate (as some people would have you think). What I _DID_ state was that the ultrasonics were not designed (or to serve the purpose) to be relied upon to maintain height above ground in the Follow Me mode. For the average user there are just too many short comings to be used in this way. Want to prove me wrong? Place the Phantom in Follow Me mode and then start to drive, bike, run and walk up and down hills higher then the set altitude of the Phantom. If you don't manually control the Phantom it will crash pretty quickly.
If you want to see some testing of the ultrasonics maintain the height of a P3A (not a P4), I've uploaded a clip to YT:
There is no breakdown or explanation of the clip (I just threw it up now in case someone want to look at it themselves). I can tell you that in a few instances, the VPS simply stopped reporting correct height and stopped maintain the correct height. This was at 5 feet and 2 feet. You can especially see this toward the end of the clip. The ultrasonics go crazy off and on and report greatly incorrect information. Keep in mind, I _did not_ push this testing to the limit. The intent was not to show that they did not not work as needed... as this can only be shown perfectly with a crash.
Hover accuracy for the P3A and the P4 are the same. The only difference in this regard is range and speed at which they work. Obviously the more tilt the Phantom has, the less accurate the ultrasonics will be (to the point where they are useless).
Last edited:
- Joined
- Jan 2, 2017
- Messages
- 610
- Reaction score
- 303
- Age
- 62
"But I'd say that it's use is as a last dich effort to prevent a crash. Too each their own."
Just for the record, that's not part of F3's post. You said that I don't agree with something in his post. If I were of the opinion that the ultrasonics are a last ditch effort to prevent a crash, it does not disagree with anything in F3's post. Nothing F3 posted was incorrect, as far as I'm concerned, and I don't (and have not) disagreed with anything he stated. In my prior post above I do point out that the barometer altitude and ultrasonic height are two different things and attempt to support my statement that the ultrasonics are not reliable to adjust height in the Follow Me mode. I mentioned this before, I've done a _lot_ of testing this exact combo. It _can_ work... I just don't recommend it to the general pubic (the combination of use, as this is what this thread is about) as it has too many faults and can easily result in a crash.
I'm well aware of the differences between the two. My only point was that the ultrasonics are used for more than just crash avoidance. My example was that the drone with no stick input is designed to maintain position and current height. Above 30 feet the barometer tells the drone if it is falling or rising in order to maintain height. Below 30 feet the ultrasonic does this job instead.
Your controller shows the barometer height above takeoff point at all times. But internally the drone relies mostly on the ultrasonic to maintain height when below 30 feet instead of the barometer in the case of for example just hovering or low angle movements etc. This is because the ultrasonic is far more accurate than the barometer.
Your controller shows the barometer height above takeoff point at all times. But internally the drone relies mostly on the ultrasonic to maintain height when below 30 feet instead of the barometer in the case of for example just hovering or low angle movements etc. This is because the ultrasonic is far more accurate than the barometer.
Last edited:
I agree 100% with all of that, never said anything different. _However_, I went back and read my initial post. It's still correct, but only when read in context with what I was responding to. It was mentioned that while using Active Track (I think above I accidentally called it Follow Me.... two different modes but it still apples), that the Phantom maintain's it's altitude to the ground. This is only done with ultrasonics. The thought behind this was that the Phantom maintains this height. I stated that a person should not rely on the ultrasonics to maintain a height to where it won't crash. Ultrasonics will do a great job to simply maintain height above ground, for example when hovering. That is what you are saying and I agree 100% with. However, when it comes to Active Track (see, I got it right that time), a person should not rely on ultrasonics to maintain height above ground to avoid a collision. It's limit's on the P4 is 33 feet. The higher off the ground the worse the accuracy. But let's just discuss avoiding a crash. If you start up Active Track and start to move around (the faster the more predominant the result) and you move up and down (such as going up and down a hill), you will see that your height above ground will not maintain very well. Again, the faster you move, the more and faster this change will occur. Once the height is off, it's off. It will simply try to stick to that new height above ground. I'll give an example of this. Use my video and that hill as an example. If I start at the top at 5' and move the Phantom fairly fast over down that hill, it won't maintain that 5' height. It might end up at 20' above ground at the bottom. The ultrasonics will then maintain that 20' above ground altitude as best it can... but it's now maintaining at 20'. Point here is, there is no further automatic correction. Now, do this in reverse, going up a hill and you start to see the issue with relying on ultrasonics when using in Active Track mode. Now, add in any type of uneven terrain as well as objects within that terrain. The ultrasonics are not a laser beam, they are wide and get wider the further the distance. Software/Firmware programming is used to figure this is but there are _huge_ limitations on this. It's not easy to set up. But again, Any misreading will create a new height above ground that is not compensated for.
So , still... my recommendation is that a person not rely on ultrasonics to maintain height above ground when using Active tracking. It's not really designed for this type of use. I've supported my view on this with information and testing video. I've done my own testing. People can take this all as you wish.
I've done lots of testing on maintaining altitude and collision avoidance as these will continue to be the big issues with flying drones. As these two areas get better and better we will almost eliminate crashing into things and also open so many more ways to fly and use drones.
Similar threads
