DJI Mavic, Air and Mini Drones
Friendly, Helpful & Knowledgeable Community
Join Us Now

Go Home (RC Disconnected)

samysys

Well-Known Member
Joined
Apr 22, 2023
Messages
69
Reactions
40
Age
41
Location
hong kong
Hi,

I am helping to look into a drone accident where a M300 unexpectedly entered RTH mode and eventually landed into the sea. Having viewed the log, I think that this was due to DJI program error. Yet, I appeal here to see if anyone had experienced similar before and if there are any better way to response.

For some reason, I cannot fully disclose the log here, pls be understanding.

Incident

The drone first took off on a boat. The weather was fine. Obstacle sensors were switched off as the drone would be landing by hand-catch. RTH location was set to follow remote controller. The payload was camera lens by DJI. The pilot tested the condition of the drone after take off (observe the response of the drone against stick input). Everything was fine.

Upon took off, satellite reception was fine with home point updated. After about 6 minutes of uneventful flight, the radio connection suddenly lost for 10 seconds. The drone was within radio & visual line of sight and the estimated distance between RC and drone was only ard 800 meters.

Upon re-connection, the drone was stationary on the same flight height. It was 23 meter away from the point where connection was broken. No RTH was initiated. Instead, the drone was still in P-GPS mode and hovering. It stopped responding to the pilot’s command (full speed forward) and remain hovering.
No Change in Speed.jpg

About 12 second later, the drone suddenly flied back. It first ascended to the RTH height and head back to the home point (which was the take off point, HP01). When closing up to the take off point, it automatically upon the Home Point and changed direction towards the new Home Point (HP 02). It repeated the above twice and eventually auto landed on HP 04.
Return Path.jpg
During this period, pilot had no control over the drone (stick control/ payload operation/ any other setting). He can view the FPV camera only. As he had no control over the drone, it was considered too risky to land it on a floating boat (other ppl there). The drone thus landed on sea and sank as submarine.

Upon log review, it was found that the drone entered into “Go Home (RC Disconnected)” mode. The pilot did not operate/ initiate that function (RC.goHomeDepressed= false). According to pilot, he also did not update the home point during the flight.
Change Mode.jpg

It was also found that during this “Go Home (RC Disconnected)” condition, the downlink and uplink were ranged 80-100.
up& downlink.jpg
Assumption

Having considered the above, I incline to believe that the drone eerily entered into a strange condition and “thought” that the RC link was disconnected and thus disregarded all input from the pilot.

Has anyone seen something like this before? And more importantly, any better way to prevent and response to such situation?

Thank you.
 
  • Like
Reactions: RodPad
For some reason, I cannot fully disclose the log here, pls be understanding.
...
Has anyone seen something like this before? And more importantly, any better way to prevent and response to such situation?
Perhaps if you could post the data, someone might be able to find the cause.
But without being able to see the data which contains much, much more than you've explained, it's just not possible.
 
During this period, pilot had no control over the drone (stick control/ payload operation/ any other setting). He can view the FPV camera only. As he had no control over the drone, it was considered too risky to land it on a floating boat (other ppl there). The drone thus landed on sea and sank as submarine. And more importantly, any better way to prevent and response to such situation?

Why not navigate the boat to the hovering drone before it landed on the sea, and let it instead land on the boat, under its own controlled Autoland from a depleted battery? At that point, it was not flying unpredictably or wildly and was in a steady hover. Just clear the area directly below the drone of the people and wait for it to land.
 
  • Like
Reactions: RodPad and Meta4
Why not navigate the boat to the hovering drone before it landed on the sea, and let it instead land on the boat, under its own controlled Autoland from a depleted battery? At that point, it was not flying unpredictably or wildly and was in a steady hover. Just clear the area directly below the drone of the people and wait for it to land.
Hi,

I have the same thought too so asked the pilot who told me that there are other equipment on the boat way more expensive than the M300... He did not wish to take that risk.
 
Hi,

I have the same thought too so asked the pilot who told me that there are other equipment on the boat way more expensive than the M300... He did not wish to take that risk.
Not sure what risk there is when he gets to choose where the M300 gets to land, as it is in a stable hover until it descends straight down.

How was he planning on landing it on the boat otherwise, had there not been a control issue? It's the same thing either way, as the drone is not flying around wildly. It is in a stable hover, just waiting to descend after no input, while the battery runs down. Same as in a normal landing..
 
The drone first took off on a boat.
How come stories that start like this never end well?
The obvious question has already been asked and answered (why don't you post the actual logs)...so that's that with that.
 
  • Love
Reactions: RodPad
Not sure what risk there is when he gets to choose where the M300 gets to land, as it is in a stable hover until it descends straight down.

How was he planning on landing it on the boat otherwise, had there not been a control issue? It's the same thing either way, as the drone is not flying around wildly. It is in a stable hover, just waiting to descend after no input, while the battery runs down. Same as in a normal landing..

the normal drill would be hovering about 1.5 m above the deck and have an assistant to hand catch the lower frame then the pilot still the motor in the middle of the air.

the drone seems to be stable but the boat isn't.
 
How come stories that start like this never end well?
The obvious question has already been asked and answered (why don't you post the actual logs)...so that's that with that.
drone don't like boat~~ i suppose...


anyway, my question was if anyone has seen this "Go Home (RC disconnected)" status before? and if so, would it amount to disregard all control input by pilot.
 
drone don't like boat~~ i suppose...


anyway, my question was if anyone has seen this "Go Home (RC disconnected)" status before? and if so, would it amount to disregard all control input by pilot.
If the RC is disconnected, by definition, the pilot has no control input whatsoever. It will follow whatever the app setting is for "Loss of Signal" response, which appears to have been either RTH to where RTH was set to, or Hover. Either way, the drone was in a stationary stable hover in clear VLOS.
Navigate the boat to it, and wait for the drone to land on it, under a stable critical battery descent.

Still not getting what the difference is for recovery between launching from an unstable boat and landing on it, and navigating the boat to a hovering, stable drone to land it on the boat. If the former was already anticipated, the latter should pose a no greater challenge. Deliberately drowning the $10,000 drone because the pilot didn't know how to land is inexcusable! This was always a suicide mission!
 
If the RC is disconnected, by definition, the pilot has no control input whatsoever. It will follow whatever the app setting is for "Loss of Signal" response, which appears to have been either RTH to where RTH was set to, or Hover. Either way, the drone was in a stationary stable hover in clear VLOS.
Navigate the boat to it, and wait for the drone to land on it, under a stable critical battery descent.
yup. the situation is paradoxical...

It is like, I telling you "I am lying"....

So, on one hand, there is the log of "Go Home (RC Disconnected)" and on the other hand the log is still there (as far as I understand, all log are only live recorded, no retrospective recording). Everything in the log looks normal (even recorded the stick input and every button pressed), except the drone was not responding.

I don't suppose this status happens often, but may be someone in the forum saw it before and know how to recovery from this status.
 
yup. the situation is paradoxical...

It is like, I telling you "I am lying"....

So, on one hand, there is the log of "Go Home (RC Disconnected)" and on the other hand the log is still there (as far as I understand, all log are only live recorded, no retrospective recording). Everything in the log looks normal (even recorded the stick input and every button pressed), except the drone was not responding.

I don't suppose this status happens often, but may be someone in the forum saw it before and know how to recovery from this status.
"And more importantly, any better way to prevent and response to such situation?"

Nothing paradoxical about properly dealing with the RC being disconnected, whatever the reason, which results in the drone still being landed back on the boat, just like under pilot control, where you slowly descend from a hover, after maneuvering the boat to the hovering drone. Two entirely different issues. There is no excuse for deliberately drowning the $10,000+ Matrice M300 under these circumstances! You keep deliberately ignoring this fact.

Exactly why it disconnected is really academic. The answer lies in the log files which you refuse to share for proper analysis. The drone drowned due to pilot error. Every experienced pilot has encountered RC disconnections, for whatever reason, and knows how to still safely recover the aircraft from a hover. Clearly, the RTH was set to the location of the hover, and it finally descended at that location due to critical battery level, since the foolish pilot deliberately abandoned it, after the RC disconnection.
 
Still not getting what the difference is for recovery between launching from an unstable boat and landing on it, and navigating the boat to a hovering, stable drone to land it on the boat. If the former was already anticipated, the latter should pose a no greater challenge. Deliberately drowning the $10,000 drone because the pilot didn't know how to land is inexcusable! This was always a suicide mission!

I was told that adjusting the boat to meet the descending drone was difficult.

This was not a normal landing. The pilot had no control over the drone and the landing point was unknown (the drone seems to be updating the Home Point at irregular interval, the last update was during descend/auto landing phase but did not actually land there. These all were only revealed in post flight log review but not known/available during the flight).

And whilst I have not tried it myself, I do not think a drone can auto landing on a floating boat. Is there ever a successful case? if it is reliable, why every boat landing video in youtube advocate about hand-catch landing?
 
the landing point was unknown (the drone seems to be updating the Home Point at irregular interval, the last update was during descend/auto landing phase but did not actually land there.
It would be helpful to see the actual flight data to be able to work out what was actually happening, rather than rely on a brief second hand description.
These all were only revealed in post flight log review but not known/available during the flight).

And whilst I have not tried it myself, I do not think a drone can auto landing on a floating boat. Is there ever a successful case? if it is reliable, why every boat landing video in youtube advocate about hand-catch landing?
If a drone was autolanding above a boat, hand catching shouldn't be a problem.
How big was the boat and how rough was the water?
 
the normal drill would be hovering about 1.5 m above the deck and have an assistant to hand catch the lower frame then the pilot still the motor in the middle of the air.

the drone seems to be stable but the boat isn't.
How was this any different than the "normal drill?"The drone will stop the motors after the hand catch, which it perceives as a landing, when it cannot descend any further. The battery could also have been ejected by someone else immediately upon hand-catching, while the hand-catcher waited for the motors to stop, if it was taking too long.
 
  • Like
Reactions: Meta4
I was told that adjusting the boat to meet the descending drone was difficult.

This was not a normal landing. The pilot had no control over the drone and the landing point was unknown (the drone seems to be updating the Home Point at irregular interval, the last update was during descend/auto landing phase but did not actually land there. These all were only revealed in post flight log review but not known/available during the flight).

And whilst I have not tried it myself, I do not think a drone can auto landing on a floating boat. Is there ever a successful case? if it is reliable, why every boat landing video in youtube advocate about hand-catch landing?
How long did the drone hover in place, before descending, during which time the pilot decided to abandon it, while it was still hovering?

Must not have been his Matrice M300!

Autoland and Forced Landing are different. Forced landing happens due to Critical battery. It will land straight down over whatever is there, whether into the ocean, or into the open hands of a hand catcher, or onto the surface of a boat.

From your prior description, the drone was not Autolanding, it was hovering without any RC control. Eventually, the battery reached Critical Battery, leading to a Forced landing in place.

Plenty of time between the lengthy final hover, and the eventual inevitable Forced Landing in place, leaving opportunity to navigate the boat into position, without control of the drone.

However, instead, the pilot just abandoned the $10k+ drone, because of his inexperience and lack of contingency planning.
 
How was this any different than the "normal drill?"The drone will stop the motors after the hand catch, which it perceives as a landing, when it cannot descend any further. The battery could also have been ejected by someone else immediately upon hand-catching, while the hand-catcher waited for the motors to stop, if it was taking too long.
Normally, for a drone in that size, the pilot would hover it 1.2-1.6 meter above the deck and another guy would be half kneeing and catch both legs with two hands. The Pilot would do CSC instead of waiting the drone to perceives as a landing.

ejecting battery.....well, it is a M300 which can operate with just one battery and you have to unlock the battery lock and pull hard horizontally to eject the battery. With the carbon fibre rotor blade still turning, doing that on a unstable platform twice.... I don't think that is really pratical.
 
How long did the drone hover in place, before descending, during which time the pilot decided to abandon it, while it was still hovering?

Must not have been his Matrice M300!

Autoland and Forced Landing are different. Forced landing happens due to Critical battery. It will land straight down over whatever is there, whether into the ocean, or into the open hands of a hand catcher, or onto the surface of a boat.

From your prior description, the drone was not Autolanding, it was hovering without any RC control. Eventually, the battery reached Critical Battery, leading to a Forced landing in place.

Plenty of time between the lengthy final hover, and the eventual inevitable Forced Landing in place, leaving opportunity to navigate the boat into position, without control of the drone.

However, instead, the pilot just abandoned the $10k+ drone, because of his inexperience and lack of contingency planning.
Here is the sequence of the incident.

1) There was a brief RC disconnection (T=0)
2) RC reconnected (T=10s). Drone was hovering and in P-GPS mode.
3) Between T=10s and T=22s, Drone remain hoovering regardless of the pilot's stick control
4) T=22s, drone changed mode into Go Home (OSD.flycState = Go Home, OSD.flightAction=Go Home (RC Disconnected))
5) Between T=22 and T=ard 2 min, drone flied as if it was returning home (first ascend to RTH height and then fly back). However, it continued to ignore any input of the pilot and auto updated the Home Point location several times.
6) Between T=2min and T=3min, it began Auto Landing (OSD.flycState = Auto Landing, OSD.flightAction=Go Home finished). It still continued to ignore any input of the pilot. During landing, the home point was auto updated again but the drone did not fly to that last home point.
7) Eventually, the drone landed and dived into water.

Between 2) and 7), the pilot can see the FPV feedback but was not able to do anything on the controller to change its course/ stop the RTH action. His action was recorded in the flight log too. Both collaborate with each other.

The last record of the log show it still have more than 70% battery.
 
If a drone was autolanding above a boat, hand catching shouldn't be a problem.
How big was the boat and how rough was the water?

But hand-catching an uncommanded M300 on a boat still have substantial risk.

I do not know how to best describe water condition. Let's just say the drone measures level 4 wind in Beaufort Scale (20-30km per hr) during the flight.
 
6) Between T=2min and T=3min, it began Auto Landing (OSD.flycState = Auto Landing, OSD.flightAction=Go Home finished). It still continued to ignore any input of the pilot. During landing, the home point was auto updated again but the drone did not fly to that last home point.
I'd really like to see the recorded data to understand what's happening here.
But hand-catching an uncommanded M300 on a boat still have substantial risk.
Before the problem happened, was it planned to hand catch on the boat?
What was the plan to end the flight if there were no problems?

How big was the boat?
 
Wow!
REALLY!
Still, no actual flight log?
I looked at the your first post again.
I see some graphical computer skills, I know I would not be able to do.

"I am helping to look into a drone accident where a M300 unexpectedly entered RTH mode and eventually landed into the sea. Having viewed the log, I think that this was due to DJI program error. Yet, I appeal here to see if anyone had experienced similar before and if there are any better way to response.

For some reason, I cannot fully disclose the log here, pls be understanding.
"

I am helping to look into a drone accident?
For some reason, I cannot fully disclose the log?
I'm done, adjourned.👌


Rod ..
 

DJI Drone Deals

New Threads

Forum statistics

Threads
135,779
Messages
1,609,881
Members
164,239
Latest member
sytandy
Want to Remove this Ad? Simply login or create a free account