Please select Into the mobile phone version | Continue to access the computer ver.
It should be technically possible
492 30 2018-11-13
Uploading and Loding Picture ...(0/1)
o(^-^)o
Chuck101
Second Officer
Flight distance : 27815 ft
United States
Offline

With all the sophistication that the Spark has I would think it should be possible to have the AC fly toward the RC when it looses GPS data.
If programmed to do so it should be able to determine from what direction the RF from the RC is coming from and fly in that direction.
What do you think?

2018-11-13
Use props
Spark_Pilot_XxXx
Second Officer
Flight distance : 67841 ft
Germany
Offline

Well Sounds good ,but Not possible. The dji Spark does not know where the RC controls are from
2018-11-13
Use props
Chuck101
Second Officer
Flight distance : 27815 ft
United States
Offline

Spark_Pilot_XxXx Posted at 11-13 05:33
Well Sounds good ,but Not possible. The dji Spark does not know where the RC controls are from

Currently it does not know from what direction the RF signal from the remote is coming from. But it COULD be made to determine the direction and fly toward it.
2018-11-13
Use props
S-e-ven
Captain
Flight distance : 3787533 ft
Thailand
Offline

Mostly it looses/ignores GPS when the compass get "bad". So how can it manage to know the direction?
2018-11-13
Use props
Chuck101
Second Officer
Flight distance : 27815 ft
United States
Offline

S-e-ven Posted at 11-13 05:39
Mostly it looses/ignores GPS when the compass get "bad". So how can it manage to know the direction?

Signal strength change when rotating the Spark would provide the direction. Yes, a directional antenna would be required in the Spark.
The concept is simple.
2018-11-13
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

A GPS signal should be enough to get the AC back to its home point. To find the home point through radio signal strength is very complex and would require a lot of data processing.
2018-11-13
Use props
DJI Stephen
Super Moderator

Offline

Hello. Good day and thank you for sharing this information with us today. Rest assured that I will be forwarding this thread to our DJI R&D team for future development of the DJI Spark. Thank you for your suggestion and support.
2018-11-13
Use props
Chuck101
Second Officer
Flight distance : 27815 ft
United States
Offline

ssylca44 Posted at 11-13 06:23
A GPS signal should be enough to get the AC back to its home point. To find the home point through radio signal strength is very complex and would require a lot of data processing.

This is not intended to locate the home point. Only to fly in the direction of the RC RF signal. In this scenario GPS has failed and the Spark is in ATTI mode.

I wish I was articulate enough to convey this concept for everyone's understanding. However, I will continue to fine tune the idea for those that haven't quite grasped it yet.
2018-11-13
Use props
DeuceDriv3r
First Officer
Flight distance : 3933077 ft
United States
Offline

more than possible .. in the real world we call this flying 'pure inertial'  sure the accuracy of your position will degrade over the duration of the flight.. but before GPS  I would regularly fly across the atlantic pure inertial for 4-6 hours before getting the navigator to tighten up our position taking a celestial shot with the sextant .. of course we had ring laser gyros that cost a fortune but since the spark would have to stay airborne for about 3 minutes to come back the only thing holding it back would the the algorithms and math that DJI would have to place in the drone to keep position based off last known position updated by the gyro/accerlerometer inputs.. its not rocket science and should not even need that much processing power...  heading could also be updated again based off of last known good heading and the gyros / accelerometer .. it would do a far sight better navigating the drone than the wind does.. especially in situations where the bird goes atti and the same time dropping all video and RC links...
2018-11-13
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

Hi Chuck,

Your suggestion is very good, however, I believe that the GPS signal should be less prone to failure than compass and accelerometer. Returning to the take-off point would be a good start in getting the AC back.
2018-11-13
Use props
DeuceDriv3r
First Officer
Flight distance : 3933077 ft
United States
Offline

ssylca44 Posted at 11-13 08:29
Hi Chuck,

Your suggestion is very good, however, I believe that the GPS signal should be less prone to failure than compass and accelerometer. Returning to the take-off point would be a good start in getting the AC back.

what you are trying to say is what I have seen in my logs and those of other atti excursions on both the spark and mavic pro.. and that is the atti was preceded by compass issues where the FC purposely DROPPED GPS sometimes 20 sats and 5 bars.. to go into non nav atti mode...

pure inertial can theoretically still work again by using accelerometer  / gyro movements since last known good position and heading to at least estimate a course back to the homepoint  and estimate when it gets there .. it will be far from perfect but its better than what we have now where the FC makes no attempt to navigate ..at all.. the only thing working in atti is auto leveling ..
2018-11-13
Use props
DeuceDriv3r
First Officer
Flight distance : 3933077 ft
United States
Offline

rudimentary block diagram of inertial navigation ..notice no block for GPS position .. its calculated solely on the motions of the airborne craft after alignment and present position are input..

just like the good ole days...
Inertial-navigation-system-algorithm-block-diagram-Taken-from-Titterton-Weston-2004.png
2018-11-13
Use props
DeuceDriv3r
First Officer
Flight distance : 3933077 ft
United States
Offline

and its been out for awhile...

http://insideunmannedsystems.com ... stem-now-available/
2018-11-13
Use props
Chuck101
Second Officer
Flight distance : 27815 ft
United States
Offline

ssylca44 Posted at 11-13 08:29
Hi Chuck,

Your suggestion is very good, however, I believe that the GPS signal should be less prone to failure than compass and accelerometer. Returning to the take-off point would be a good start in getting the AC back.

You keep side-stepping a given here: GPS has failed, Spark is in ATTI mode.

AC is drifting with the wind. The object is determine from which direction the RC's RF signal is coming from and fly in that direction. When AC is close to you then you stick land it. This is far better than watching it drift away aimlessly or failure to manually fly it back home.

Inertia, compass heading, known position and all other factors are irrelevant.  Those considerations add unneeded complexity. The ONLY factor is determining from what relative direction the RF signal from the RC is coming from.
2018-11-13
Use props
DeuceDriv3r
First Officer
Flight distance : 3933077 ft
United States
Offline

Chuck101 Posted at 11-13 08:58
You keep side-stepping a given here: GPS has failed, Spark is in ATTI mode.

AC is drifting with the wind. The object is determine from which direction the RC's RF signal is coming from and fly in that direction. When AC is close to you then you stick land it. This is far better than watching it drift away aimlessly or failure to manually fly it back home.

another interesting concept for navigation..

just go off signal gain ... possibly direction if that is possible to determine azimuth using the 2 antennas as a built in, slow yaw routine to point the aircraft at the signal source..

can still be done, and it being done, using pure inertial though so that would not need the 'reinventing of the wheel'
2018-11-13
Use props
hallmark007
Captain
Flight distance : 4797602 ft
  • >>>
Ireland
Online

Chuck101 Posted at 11-13 08:58
You keep side-stepping a given here: GPS has failed, Spark is in ATTI mode.

AC is drifting with the wind. The object is determine from which direction the RC's RF signal is coming from and fly in that direction. When AC is close to you then you stick land it. This is far better than watching it drift away aimlessly or failure to manually fly it back home.

I think your problem might be what will give Aircraft it’s lock position, yes maybe with VPs which is effectively opti mode, barometer can hold aircrafts position horizontal but not vertical ,
Yes more antennas more tech, but you are no longer talking about a Spark at €350, everything comes at a price, but also would need a lot of research development and testing.
There are drones out there being tested for all sorts of things and I’m sure when tech becomes available at an affordable price we will see it implemented in consumer and pro drones.
I’m sure there will be plenty who would say it’s simple won’t cost anything and with a switch of a button we could have it in next FW update.
Landing without latterl movement of the drone is a great idea and I’m sure that future tech we will see this or even better as of now in all consumer drones we see very similar tech , with the exception of more redundancy on bigger drones, I’m certain if some redundancy hardware could be substituted for better SW we would or will see it in the future.
Good idea, no substitute for thinking.
2018-11-13
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

Hi Chuck,

Sorry, but I think of all the sensors available on the Spark, the GPS signals should be the very last to fail, as long as GPS was ok at taking off. If flying in congested areas such as cities, then it is a pilot error from the start causing loss of GPS signals.
The main issue with inertial navigation is the wind. This may be secondary to a commercial airplane, but it is a major factor causing a flyaway on the  Spark.
2018-11-13
Use props
HedgeTrimmer
Captain
United States
Offline

ssylca44 Posted at 11-13 06:23
A GPS signal should be enough to get the AC back to its home point. To find the home point through radio signal strength is very complex and would require a lot of data processing.

Not any more data processing and complexity than it required during WWii.    
2018-11-13
Use props
HedgeTrimmer
Captain
United States
Offline

ssylca44 Posted at 11-13 09:58
Hi Chuck,

Sorry, but I think of all the sensors available on the Spark, the GPS signals should be the very last to fail, as long as GPS was ok at taking off. If flying in congested areas such as cities, then it is a pilot error from the start causing loss of GPS signals.

With some low-tech modifications, any of DJI's future drones could find its way back to where R.C. is currently.

It would be possible to do what is proposed with two antennas, phase differential circuitry, and additional firmware to serve as interface to drone's flight system.  With two antennas, the drone would have to rotate till both antennas' signals were in phase, then make a test move to see if drone was moving closer or further away.   If drone had three (or four) antennas, it could determine direction to fly without an initial test move.

Because DJI's Drones and R.C.s function as both Transmitters and Receivers, it would be possible for Drone to accurately determine it's distance from R.C. by performing equivalent of a 'ping'.  Drone sends a 'ping' signal to R.C., and R.C. immediately responds by returning 'ping'.  Dividing 'ping' time (sent & returned) by 2, and subrtracting nanoseconds for R.C. internal response delay, would give number in meters - going with ~2.99 nanoseconds for radiowave to travel 1-meter.
2018-11-13
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

Hi HedgeTrimmer,

What's about reflexions and echoes from buildings, air masses, etc.
2018-11-13
Use props
DeuceDriv3r
First Officer
Flight distance : 3933077 ft
United States
Offline

HedgeTrimmer Posted at 11-13 11:44
With some low-tech modifications, any of DJI's future drones could find its way back to where R.C. is currently.

It would be possible to do what is proposed with two antennas, phase differential circuitry, and additional firmware to serve as interface to drone's flight system.  With two antennas, the drone would have to rotate till both antennas' signals were in phase, then make a test move to see if drone was moving closer or further away.   If drone had three (or four) antennas, it could determine direction to fly without an initial test move.

very similar to the logic a roomba uses to find the dock.. although it uses beams of infrared .. the concept is the same.. it finds the beam .. rotates and drives into the beam.. figures out divergence / convergence and drives right up into the charger..

pure inertial would only require the math be added to the FC ..
2018-11-13
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

What's about if the RC quits on you? I still think that GPS is the best approach as long as the home point has been recorded in the AC on taking off.
2018-11-13
Use props
HedgeTrimmer
Captain
United States
Offline

ssylca44 Posted at 11-13 15:18
What's about if the RC quits on you? I still think that GPS is the best approach as long as the home point has been recorded in the AC on taking off.

From OP's opening post: "With all the sophistication that the Spark has I would think it should be possible to have the AC fly toward the RC when it looses GPS data."

Thus GPS is not the best approach, because in OP's senario GPS's data is not available, GPS is not working, GPS is dead, dead, dead.

2018-11-13
Use props
HedgeTrimmer
Captain
United States
Offline

ssylca44 Posted at 11-13 11:59
Hi HedgeTrimmer,

What's about reflexions and echoes from buildings, air masses, etc.

Reflections / echoes should be weaker than signal coming directly from R.C.  System can be setup to go with strongest signal.  Along with using drone's IMU data as mentioned by DeuceDriv3r to return.

Thought for consideration: Something, even though it is not perfect, that keeps drone from going Beserker, (rapidly accelerating, while changing its flight path) is better than nothing we got now.
2018-11-13
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

Hello HedgeTrimmer,

Just one more question; why the AC losses GPS data? Is it because the GPS satellites stop working? Obviously NO. Bad weather? Maybe but one shouldn't be flying then. Obstacles such as tall buildings? Again one shouldn't be flying there. Or is it the Spark onboard processor that has decided to reject GPS data because of other sensors errors. We have no clue how the software running the AC operates. In my experience with computers, one shouldn't necessarily trust diagnostics reported by the OS.
2018-11-14
Use props
Chuck101
Second Officer
Flight distance : 27815 ft
United States
Offline

ssylca44 Posted at 11-14 06:34
Hello HedgeTrimmer,

Just one more question; why the AC losses GPS data? Is it because the GPS satellites stop working? Obviously NO. Bad weather? Maybe but one shouldn't be flying then. Obstacles such as tall buildings? Again one shouldn't be flying there. Or is it the Spark onboard processor that has decided to reject GPS data because of other sensors errors. We have no clue how the software running the AC operates. In my experience with computers, one shouldn't necessarily trust diagnostics reported by the OS.

You are missing the entire point of this proposal.

The "why" of GPS loss or compass errors is completely irrelevant.

FACT: The AC has entered ATTI mode. For whatever reason. It matters not.

You must manually fly it back home or let it drift with the wind and it becomes a lost drone.

With my suggestion at least there is a better chance of not losing the AC.

2018-11-14
Use props
hallmark007
Captain
Flight distance : 4797602 ft
  • >>>
Ireland
Online

ssylca44 Posted at 11-14 06:34
Hello HedgeTrimmer,

Just one more question; why the AC losses GPS data? Is it because the GPS satellites stop working? Obviously NO. Bad weather? Maybe but one shouldn't be flying then. Obstacles such as tall buildings? Again one shouldn't be flying there. Or is it the Spark onboard processor that has decided to reject GPS data because of other sensors errors. We have no clue how the software running the AC operates. In my experience with computers, one shouldn't necessarily trust diagnostics reported by the OS.

You are correct, it’s almost unheard of for these aircraft to lose gps, but it’s not loosing gps that’s the problem being discussed, it’s dropping gps in favour of compass because ofconflict with data, and proposed ideas to replace what is already there, I don’t believe it’s possible without Hardware, I don’t believe there is anything full stop, out there in any dji drone or any other drone that is better than what we already have, I believe if there were a tried and tested SW or HW solution to an aircraft going to Atti mode and one that could guarantee complete safe return of Aircraft and one that could be costed for both professional drones or small consumer drones then we would already have it, can something be developed I’m certain it can and I see many working ao other systems but none that have been used in mass produced drones, it’s all theory right now and if something is discovered to work on these drones we all fly , it’s most likely that it would be first used on bigger more professional drones before it trickles down to smaller consumer drones.
Every one has a theory some completely batty, some that make some sense, for those who believe it’s simple and know what to do, then there is a whole world of electronics out there for them to further their ideas.
2018-11-14
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

Hi hallmark007,

I fully agree with you, ' it’s almost unheard of for these aircraft to lose gps, but it’s not loosing gps that’s the problem being discussed, it’s dropping gps in favour of compass because ofconflict with data'
The question is why the software is not written to favor the GPS data over the rest, and why should the AC drop to ATTI mode right away without an attempt to at least hold position, or better return home on GPS data alone. Anyone with a handheld GPS now that either on land or on the water it is quite possible to return to your starting point using GPS tracking.

2018-11-14
Use props
hallmark007
Captain
Flight distance : 4797602 ft
  • >>>
Ireland
Online

ssylca44 Posted at 11-14 08:34
Hi hallmark007,

I fully agree with you, ' it’s almost unheard of for these aircraft to lose gps, but it’s not loosing gps that’s the problem being discussed, it’s dropping gps in favour of compass because ofconflict with data'

Unfortunately these craft and I’m talking about what we are working with now, cannot function with data conflict, it also cannot navigate without compass, it can fly in Atti mode can be steered, yes I know that sometimes it’s uncontrollable but there are plenty of times it can be controlled by controller. Yes many say go to gps drop compass then at least we could just land, but what would those flying over cliffs over water over people say when they have to sacrifice their drone or cause damage to people or property, in gps alone almost impossible to steer drone, in Atti mode it’s possible to land and steer drone to safety, I know for instance that Drivduece4 or whatever he calls himself says his drone went to Atti mode at 3000 metres and he had no problem flying it . Although I haven’t seen it I think if the drone had just gps it would at best just turn around in circles.

Are there better ways out there for mass produced drones consumer and professional, No, will we see better tech and hardware in the future, we’ll i believe that dji are so far ahead in the market that they will have the new tech and advancements first.For me I know how these drones work now , and that’s what I’m commenting on.

It’s also my opinion those who take the most care when flying are less likely to run into problems, not all but most.
2018-11-14
Use props
ssylca44
Captain
Flight distance : 117726 ft
Canada
Offline

Thank you for your comments hallmark007, well said!
One should practice Atti as much as possible, unfortunately, DJI doesn't allow manual switching between GPS and Atti. The only way (as seen on youtube) is to shield the GPS signal by covering part of the spark with aluminum foil.
2018-11-14
Use props
eYeSkYeYe
Captain

Croatia
Offline

There is also a software way to switch to atti mode using sports mode button. Software can be configured in a such a way that turning switch to the right would switch Spark to atti instead of sports mode.
2018-11-15
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules