Welcome DJI Spark Pilot!
Jump in and join our free Spark community today!
Sign up

My opinion about fly aways

Dodge DeBoulet

Well-Known Member
Join
Dec 27, 2018
Messages
146
Likes
115
Loc
Maine
#41
The Offline Maps option is NOT available in the latest version of the iOS app (currently 4.3.12), at least when installed on an OG iPad Air with iOS 12.1.3 ... that's the only iOS device I own.

It IS available on the current Android version (also 4.3.12).

I doubt this will sway anyone's platform preference, though ;)
 

Spark 317

Well-Known Member
Premium Pilot
Join
Jul 10, 2017
Messages
2,383
Likes
1,273
Loc
Indiana, USA
#42
I have a mini4 and the offline map was never there for me, it's a Droid thing.

I scroll around the map image while connected to the home WiFi, caching it in the Go4 app before flying at that location. I assume that's how all iOS users do it.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#43
I have a mini4 and the offline map was never there for me, it's a Droid thing.

I scroll around the map image while connected to the home WiFi, caching it in the Go4 app before flying at that location. I assume that's how all iOS users do it.
That makes perfect sense, I assumed it just worked without much fuss.
 

1uk3y

Well-Known Member
Join
Jan 4, 2019
Messages
51
Likes
54
Age
34
Loc
West Midlands, UK
#44
If GPS signal is lost, drone initiates ATTI mode, and CANNOT RTH.
It just hovers, drifting with the wind.
I think, in my case GPS was locked upon takeoff which provided the Spark sufficient time to record its location, I then flew it out and about and it switched into ATTI and also lost WiFi signal to my iPhone.. I can only assume that at some point it regained a GPS lock and switched to failsafe RTH.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#45
I think, in my case GPS was locked upon takeoff which provided the Spark sufficient time to record its location, I then flew it out and about and it switched into ATTI and also lost WiFi signal to my iPhone.. I can only assume that at some point it regained a GPS lock and switched to failsafe RTH.
I was thinking on this when GPS is lost, the Spark also has the compass so it should be able to estimate distance and speed using time as well as direction it has traveled. How close to your home point did it get to for landing?
 

lmel2005

Well-Known Member
Join
May 15, 2018
Messages
471
Likes
155
Age
61
#46
I was thinking on this when GPS is lost, the Spark also has the compass so it should be able to estimate distance and speed using time as well as direction it has traveled. How close to your home point did it get to for landing?
Spark cannot estimate distance, if GPS signal is lost.
Compass is used only for orientation and yaw measures.
Speed also, depends on stick movements, and on the wind. Speed isn't fixed.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#47
Spark cannot estimate distance, if GPS signal is lost.
Compass is used only for orientation and yaw measures.
Speed also, depends on stick movements, and on the wind. Speed isn't fixed.
I was pondering if the Spark utilizes ATTI mode in actual navigation when GPS is lost which should provide angle forward/aft/right/left along with how much power/RPM the props are turning to establish approximate speed over time and hence the distance traveled when GPS positioning is lost. Then the compass would provide direction back to the home point as last recorded. That was all I was thinking since ATTI mode seems fairly robust and functional from posts I have seen where GPS signals were lost, yet the Spark made it back home. It sounds like you have knowledge of what the Spark actually does with the data it has from it's onboard sensors when operating in ATTI mode. I was just guessing at how ATTI mode might work. Thanks!
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#48
On the non-GPS navigation scenario, I really do believe the Spark can actually navigate without GPS at some level. The IMU plays a crucial role in this (see below)l, along with the compass and time. I have also seen a post in this thread where a person lost GPS signal and yet the Spark returned to home. Anyway, I am just wondering and looking to learn, it's a great forum and discussions like this are very helpful. Here's info on what data the IMU provides which seems logical that this information could assist in non-GPS navigation routines when the Spark switches to ATTI mode:

The Essential Drone IMU
An inertial measurement unit works by detecting the current rate of acceleration using one or more accelerometers. The IMU detects changes in rotational attributes like pitch, roll and yaw using one or more gyroscopes. Some IMU on drones include a magnetometer, mostly to assist calibration against orientation drift.

On board processors continually calculate the drones current position. First, it integrates the sensed acceleration, together with an estimate of gravity, to calculate the current velocity. Then it integrates the velocity to calculate the current position.

To fly in any direction, the flight controller gathers the IMU data on present positioning, then sends new data to the motor electronic speed controllers (ESC). These electronic speed controllers signal to the motors the level of thrust and speed required for the quadcopter to fly or hover.

How a takes off and flies is fascinating technology. In another article entitled “How A Quadcopter Flies” we explain nice and simply how a drone can take off, hover, fly in any direction and land by adjusting its motor and propeller directions. The article includes some very informative videos.

Drone have many parts and you can read all about the various drone parts here. This covers all the physical components found in most quadcopters.
 

lmel2005

Well-Known Member
Join
May 15, 2018
Messages
471
Likes
155
Age
61
#49
On the non-GPS navigation scenario, I really do believe the Spark can actually navigate without GPS at some level. The IMU plays a crucial role in this (see below)l, along with the compass and time. I have also seen a post in this thread where a person lost GPS signal and yet the Spark returned to home. Anyway, I am just wondering and looking to learn, it's a great forum and discussions like this are very helpful. Here's info on what data the IMU provides which seems logical that this information could assist in non-GPS navigation routines when the Spark switches to ATTI mode:

The Essential Drone IMU
An inertial measurement unit works by detecting the current rate of acceleration using one or more accelerometers. The IMU detects changes in rotational attributes like pitch, roll and yaw using one or more gyroscopes. Some IMU on drones include a magnetometer, mostly to assist calibration against orientation drift.

On board processors continually calculate the drones current position. First, it integrates the sensed acceleration, together with an estimate of gravity, to calculate the current velocity. Then it integrates the velocity to calculate the current position.

To fly in any direction, the flight controller gathers the IMU data on present positioning, then sends new data to the motor electronic speed controllers (ESC). These electronic speed controllers signal to the motors the level of thrust and speed required for the quadcopter to fly or hover.

How a takes off and flies is fascinating technology. In another article entitled “How A Quadcopter Flies” we explain nice and simply how a drone can take off, hover, fly in any direction and land by adjusting its motor and propeller directions. The article includes some very informative videos.

Drone have many parts and you can read all about the various drone parts here. This covers all the physical components found in most quadcopters.
Spark cannot rth, if GPS signal is lost. Drones without a GPS module cannot rth, too.
Of course, without GPS, distance from home point cannot be calculated.

In ATTI mode, drone can hover, maintaining its height and it's orientation, if sensors (barometer, accelerometers, etc) are present and functional.
BUT, it will drift with the wind, and brake distance will be long.

There's no chance for a drone without GPS to rth, and if someone reported this, he is wrong.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#50
Spark cannot rth, if GPS signal is lost. Drones without a GPS module cannot rth, too.
Of course, without GPS, distance from home point cannot be calculated.

In ATTI mode, drone can hover, maintaining its height and it's orientation, if sensors (barometer, accelerometers, etc) are present and functional.
BUT, it will drift with the wind, and brake distance will be long.

There's no chance for a drone without GPS to rth, and if someone reported this, he is wrong.

I brought this up because I haven't seen any definitive info on this from DJI, other than users opinions which I value greatly and users posts similar to that shown below.

I fully understand mapping, GPS and geographic coordinates and how they are used for navigation. My career of 25-years was in civil engineering specializing in Geographic Information Systems (GIS), terrain mapping and computer 3D modeling/animation so I take your point and fully understand it.

However based on the vast amount of pertinent data that is constantly provided by the IMU, Barometer and Compass and used by the Spark it seems possible the Spark could calculate a vector from its current coordinate, when/where GPS was lost, to the RTH (point of take off) coordinate that is stored in the Spark and using the onboard tools/data perform an RTH.

It sure seems the numbers are there that are needed to fly home relatively safely and I would really like to know if the Spark actually uses them to RTH as I have seen people claim in posts.

Great discussion, thanks for your time!

This post was in this thread, I have seen others that made me wonder if the Spark actually uses that data to RTH without GPS support:
1549271159168.png
 

Spark 317

Well-Known Member
Premium Pilot
Join
Jul 10, 2017
Messages
2,383
Likes
1,273
Loc
Indiana, USA
#51
How about doing a test flight?

Turn on the Spark, establish GPS / Home Point lock, wrap aluminum foil around the back side of the Spark to force Atti-mode, spin up the motors and launch.

Fly to some distance away from you in Atti-mode and then hit the RTH button on the remote and see what happens.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#52
How about doing a test flight?

Turn on the Spark, establish GPS / Home Point lock, wrap aluminum foil around the back side of the Spark to force Atti-mode, spin up the motors and launch.

Fly to some distance away from you in Atti-mode and then hit the RTH button on the remote and see what happens.
It's a good idea but I think for the scenario to play out as others have posted you need to be able to disable GPS while flying once you get more than the 30m threshold for full RTH routine (ascend, travel to HP and land) to occur as programmed and then observe the Spark's behavior. If you simply disable GPS from the start as you suggest it's a disaster waiting to happen I suspect and no data/coords/location for the point of GPS signal loss would have been recorded. I don't think there is a GPS on/off toggle in the app, if there is, then that might work.

I don't know if DJI monitors this forum or not, I might send off an email because I am curious after reading multiple posts about this scenario and especially since the data is there from the onboard sensors to compute a return vector and altitude, etc. once both the home point and GPS failure point is set/recorded. I am also curious if that routine is built in, if it is it would provide some helpful insight when bad things happen and you are looking for a lost drone. Just info on behaviour that would be nice to know.
 
Last edited:

Spark 317

Well-Known Member
Premium Pilot
Join
Jul 10, 2017
Messages
2,383
Likes
1,273
Loc
Indiana, USA
#53
It's a good idea but I think for the scenario to play out as others have posted you need to be able to disable GPS while flying once you get more than the 30m threshold for full RTH routine (ascend, travel to HP and land) to occur as programmed and then observe the Spark's behavior.
Okay. How about launching with full telemetry and fly under a bridge and hover until GPS signal is lost, then initiate a RTH and see what happens then.

Make sure the RTH altitude is properly set so it won't fly up into the bridge or on full standby to cancel RTH.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#54
Okay. How about launching with full telemetry and fly under a bridge and hover until GPS signal is lost, then initiate a RTH and see what happens then.

Make sure the RTH altitude is properly set so it won't fly up into the bridge or on full standby to cancel RTH.
Again I like the idea but that is pretty much like initiating RTH indoors, probably smack up and into the bridge bottom :D as it obediently ascends to a thoughtfully set RTH altitude. Yes the scenario is good but it's the bottom of the hard deck that will give Spark a headache. Seriously though I think an email to DJI would be interesting and the reply should be shared here. I love this darn drone, it's much more capable than many people realize but along with that sophistication comes variables that may be unforeseen and cause problems, which is why knowing if Spark can autonomously navigate back home in ATTI mode would be helpful.

I do a lot of Active Track video following my Jeep off road and one day I almost lost my Spark when using dynamic home points during an Active Track session which is a pretty good story for another day, but in situations like that it would be really helpful to know better how Spark behaves in these situations.
 

Spark 317

Well-Known Member
Premium Pilot
Join
Jul 10, 2017
Messages
2,383
Likes
1,273
Loc
Indiana, USA
#55
Again I like the idea but that is pretty much like initiating RTH indoors, probably smack up and into the bridge bottom
Not really, inside a house you already have it in Atti-mode.

Set the RTH 10 feet lower than the bridge deck and in theory it should ascend up to that point and fly out from under the bridge.
The Spark would be high enough above VPS and be flying normal until you fly under the bridgeto test.

I agree an answer from DJI may solve this but for now someone could test it.

I don't have a bridge high enough around me to try.
I could probably find an overpass that crosses another road and test there some time, that way I won't have to go swimming for the Spark.
 

Mr Geek

Well-Known Member
Join
Jan 2, 2019
Messages
77
Likes
137
Age
65
#56
Not really, inside a house you already have it in Atti-mode.

Set the RTH 10 feet lower than the bridge deck and in theory it should ascend up to that point and fly out from under the bridge.
The Spark would be high enough above VPS and be flying normal until you fly under the bridgeto test.

I agree an answer from DJI may solve this but for now someone could test it.

I don't have a bridge high enough around me to try.
I could probably find an overpass that crosses another road and test there some time, that way I won't have to go swimming for the Spark.
I really like your idea about coming up with ways to test this, should be a fun Spark project that will benefit everyone with the results. OK on setting RTH altitude to 10-feet lower than underside of bridge/obstacle, but keeping in mind the actual elevation differential numbers to use for the RTH altitude are relative to the ground elevation = 0 feet where the Spark took off and is not elevation above the ground elsewhere, such as under that bridge.

I suspect the drone will reconnect to GPS satellites once it is moved out from under the bridge and ATTI mode will be cancelled, just a hunch. Actually I am thinking you will still have sufficient GPS reception even when under the bridge since I get "Ready to Go" green status even when I am inside my house most times.

As you know, the flight to the bridge/obstacle should be over 20m away from the point of take off to make sure the full RTH routine (ascend/fly to RTH/descend/land) will be carried out IF RTH is triggered when GPS is lost.
 
Last edited:

mha1

Member
Join
Aug 1, 2018
Messages
24
Likes
37
Age
40
#57
Spark cannot rth, if GPS signal is lost. Drones without a GPS module cannot rth, too.
Of course, without GPS, distance from home point cannot be calculated.

In ATTI mode, drone can hover, maintaining its height and it's orientation, if sensors (barometer, accelerometers, etc) are present and functional.
BUT, it will drift with the wind, and brake distance will be long.

There's no chance for a drone without GPS to rth, and if someone reported this, he is wrong.
Well, RTH without GPS could be done. Of course precision of present position estimates would degrade over time. Googling „position estimation using imu“ finds a lot of papers like this: http://iieng.org/images/proceedings_pdf/AE0516306.pdf

I doubt DJI implemented these kind of algorithms.
 

Dodge DeBoulet

Well-Known Member
Join
Dec 27, 2018
Messages
146
Likes
115
Loc
Maine
#58
Without IMU-based estimation, so much would depend on the accuracy of the compass and very favorable weather conditions. If, for example, there's a 1-2 degree variance on the compass heading and a winds aloft on the return leg that increase (or decrease) the ground speed (at the same wind speed) by 10-20%, the drone could end up landing from as little as 100 to potentially more than 200 meters away based on a RTH invocation at a 1km distance.

The compass heading could be spot-on, but a lateral drift due to crosswinds would also introduce significant uncertainty.

Factoring in IMU estimation, though, the best margin of error from the paper included in mha1's post is +/- 5% at 1 meter displacement. That's 50 meters at a 1km displacement, but the margin of error increases significantly at greater distances; +/- 20% at the relatively small displacement of 7 meters. So it would seem that IMU data does not significantly aid in improving accuracy at the sorts of distances at which we'd be flying our drones.
 

lmel2005

Well-Known Member
Join
May 15, 2018
Messages
471
Likes
155
Age
61
#59
Well, RTH without GPS could be done. Of course precision of present position estimates would degrade over time. Googling „position estimation using imu“ finds a lot of papers like this: http://iieng.org/images/proceedings_pdf/AE0516306.pdf

I doubt DJI implemented these kind of algorithms.
Reading this paper, we can see that error factor is high, even in laboratory conditions with no wind.

I don't know a single case of a drone being able to rth without GPS.

Maybe in the future, sensors can be accurate enough, to measure all the data needed for an IMU unit to calculate displacement accurately.
 

lmel2005

Well-Known Member
Join
May 15, 2018
Messages
471
Likes
155
Age
61
#60
I brought this up because I haven't seen any definitive info on this from DJI, other than users opinions which I value greatly and users posts similar to that shown below.

I fully understand mapping, GPS and geographic coordinates and how they are used for navigation. My career of 25-years was in civil engineering specializing in Geographic Information Systems (GIS), terrain mapping and computer 3D modeling/animation so I take your point and fully understand it.

However based on the vast amount of pertinent data that is constantly provided by the IMU, Barometer and Compass and used by the Spark it seems possible the Spark could calculate a vector from its current coordinate, when/where GPS was lost, to the RTH (point of take off) coordinate that is stored in the Spark and using the onboard tools/data perform an RTH.

It sure seems the numbers are there that are needed to fly home relatively safely and I would really like to know if the Spark actually uses them to RTH as I have seen people claim in posts.

Great discussion, thanks for your time!

This post was in this thread, I have seen others that made me wonder if the Spark actually uses that data to RTH without GPS support:
View attachment 8652
Spark cannot rth without GPS, because numbers are NOT there.

Air is a liquid, and is always moving. Winds can move a drone for miles, and IMU will never know.

Someday in the future, using other sensors, with other more powerful IMUs, maybe possible.
 

New Threads

Forum statistics

Threads
9,112
Messages
76,050
Members
13,240
Latest member
carmar555777