Hi , I’m experiencing an unintended landing detection during the RTL phase of an auto mission and would appreciate any insights from the community.
Flight Summary :
Mission type: Auto Mission with RTL as final waypoint
RTL initiated at: ~80m AGL
Descent rate: 1.3 m/s vertical speed
Unexpected landing detected at: ~35m AGL
As per Log Observations: During RTL descent, the flight controller triggered a landing detection at 35m AGL. Reviewing the logs, I found that the collective pitch held at -1° blade angle continuously for 3–4 seconds prior to the detection event. It appears the land detector interpreted this sustained minimum collective as ground contact.
Should i put more than -2 or -0.7 in collective blade pitch angle minimum when landed or Is there a parameter to increase the minimum duration or add altitude gating to the landing detection logic during RTL?
I would check that configuration and reality match with regard to blade angles. AFAIK you shouldn’t be getting negative blade angles during the descent.
I looked over the log. During the descent, the collective is at the lower limit. That along with the slow descent rate (< 1m/s) caused the landing detector to meet the criteria to declare the aircraft landed. It is odd that you have a collective value of -1 deg that produced a descent rate of only 1 m/s. I agree with @LupusTheCanine that you should check your collective values to ensure they are accurate. Are you using asymmetric airfoil blades? Based on your hover collective setting, you are hovering with only 2.5 deg of collective. so you either have a very light heli or you are using a high rotor speed, or both.
Sorry again about the crash. please adjust your lower collective limit and the land collective min value to lower values so you don’t get into this situation again.
Hi, thank you for the detailed analysis and feedback.
Yes, I’m using cambered (asymmetric airfoil) blades, which is likely why I’m able to hover at around 2.5° collective with the current rotor speed.
I’ll review and recalibrate my collective pitch range to ensure the values are accurate. As you pointed out, the low descent rate combined with the lower collective limit seems to have triggered the landing detector prematurely.
Could you please suggest suitable values for:
Lower collective angle limit
H_COL_MIN (or equivalent land collective minimum)
so I can tune them appropriately and avoid this issue going forward?
Appreciate your support, and thanks again for the guidance.
@Karan_S you probably want to set the H_COL_MIN to have enough negative collective to achieve at least 3 m/s descent rate which is what the auto mode is probably set to. I would probably set it so I get at least 5 m/s descent rate. As long as you don’t have a lot of vibration when on the ground with rotors turning, I would set the H_LAND_COL_MIN to the H_COL_MIN_ANG. As for the zero thrust collective, it is probably just a guess if you don’t have any data on the airfoil or manufacturer data. Setting it to -2 deg might be fine.
Hi , @bnsgeyer@LupusTheCanine Hope you’re doing great , I am seeking assistance regarding an unexpected landing detection issue that I am repeatedly encountering during RTL (Return-to-Launch) on my ArduPilot helicopter.
the Changes Made :
Based on the previous incident, I modified the following parameters to prevent false landing detection:
H_COL_LAND_MIN = -0.8
H_COL_ANG_MIN = -1.1
Unfortunately, despite these changes, the issue occurred again during today’s test flight. while testing i again faced the unintended landing detection during the RTL at 98m ALT and recovered the craft by again arming the craft in mid air at 8m from the Ground. the specification of the helicopter is as mentioned above in the same thread.
Unexpected landing detected at: ~98m AGL
As per Log Observations: During RTL descent, the flight controller triggered a landing detection at 98m AGL. Reviewing the logs, It appears land detector count was established after approximately 1 second while the vertical descent rate was around 0.7 m/s.. kindly review the log and give me a suggestion do not encouter this problem again.
collective angles are still not correctly set up. I strongly agree with @bnsgeyer suggestion on how to set those. I see also a major flight safety hazard in your procedure right after the FC disarmed: you switched to Loiter flight mode. In that situation (which should not happen once the setup is rectified) you should switch to Stabilize mode to regain manual control over the collective and cyclic.
Hi @bnsgeyer and @lupusthecanine@Ferrosan As suggested in the previous replies, we thoroughly reviewed our swashplate setup and verified that it is configured correctly. We have attached the swash setup images below for your reference.
We have also attached the flight log from our latest test flight. From the log, it can be observed that the swash collective demand during the RTL descent never exceeds the configured landing collective limit of -0.8, indicating that the flight controller is not commanding additional negative collective beyond this threshold.
Based on our discussions and repeated flight testing, we suspect that the helicopter is entering turbulent or recirculating airflow during certain descent conditions. In these scenarios, even -1° of collective pitch does not generate sufficient descent, resulting in a low vertical descent rate.
From the logs, it appears that because the descent rate remains around the landing detection threshold for a short duration, the landing detector counter is satisfied, causing the flight controller to falsely declare a landing.
We have attached the relevant flight logs, swash setup screenshots, and other supporting information for your review. We would greatly appreciate your analysis and recommendations to help us resolve this issue.
Additionally, we would like to highlight an important constraint regarding our helicopter configuration.
When we configure the minimum swash collective to -2°, the helicopter experiences severe ground resonance during landing, making this setting unsuitable for our airframe. Based on extensive flight testing, we have found that a minimum collective range of -1.0° to -1.5° provides the best compromise between flight performance and landing stability, while avoiding ground resonance.
Because of this limitation, increasing the negative collective further is unfortunately not a practical solution for our helicopter ( 2.2 m blade wingspan)
We would appreciate any recommendations or guidance from the developers on how best to address this scenario.
Hi @Karan_S , for this issue I can suggest using ATC_LAND_P/Y/R_MULT. This feature will reduce gains upon landing detection and should prevent ground resonance, thus allowing wider collective range on the low side. Possibly update to 4.7, there was a bug in 4.6 on this feature.
Currently, our helicopter is running ArduPilot 4.6.2.
Does the ATC_LAND_P/Y/R_MULT feature function correctly in the 4.5 firmware, or is the fix only available in ArduPilot 4.7 (beta version) ?
At the moment, upgrading to 4.7 would require us to repeat our complete flight validation and tuning process. If the feature works reliably in 4.5, we would prefer to evaluate it on our current firmware before considering an upgrade.
Could you please advise whether this solution is applicable to 4.5, or if upgrading to 4.7 is necessary?
