Success! I was at last able to fly my EasyStar with the Paparazzi open source autopilot. A few days ago I made the first flight in MANUAL and AUTO1 mode to get everything tested and tuned. Today I tried for the first time AUTO2 mode and worked beautifully! Some tuning is obviously required, especially for the roll response but it worked after only a few test flights. I will make a separate post detailing my setup, wiring (very important) and configuration files.
Following is a picture of my EasyStar, I had to move components around to minimize interference from the various receivers/transmitter. I currently have the following equipment:
- EasyStar airframe with added ailerons
- Paparazzi TWOG autopilot (from PPZ UAV)
- GPS receiver
- IR, horizontal and vertical
- PPM encoder (worked great and with excellent support from Chris)
- Xbee Pro 2.4GHz modem
- KX191 camera
- 900MHz, 500mW video transmitter
- Pan/tilt video pod
- Spektrum AR9000 receiver with a remote receiver (total of three receivers)
The following two screenshots are from the Paparazzi ground station, two flight plans were tested an oval and a figure eight. Both worked very well considering that I have not spent too much time on tuning. Altitude was alse held within about 5 meters of the target.
Finally a video from the flight follows.
I am very impressed with the Paparazzi autopilot so far, it is open source, it has the best ground station software and a large and helpful community, what else can I ask for.
We’ve seen the demostration on the stablization of the Quad but could not replicate it despite our best efforts. The problem is that the quad will always converge to a level that is slightly off the targeted setting.
For example, we target the Quad to maintain level. But the Quad will converge to a value that is near the targeted value (example :4 degrees pitch) [This remains the same even if external disturbances were induced such as someone touches the tethered Quad].
If the weight distribution of the Quad is changed, the converg value will change again in the next test. [Example -3degrees pitch].
For the PID loop, we added a lot of Derivative as suggested but the integrals were kept to a minimum as too much of the integral component will cause the response time to be slow.
Yaw alignment is also a problem too. You did not mention exactly how motor alignment affects yaw.
In short, it is not possible to hover our Quad even though we could “fly” it. In fact it acts like a drunk flying saucer, since it spins and with the PID algorithm unable to reach the 0 degree setpoint, the Quad will just move in a spiral form.
Thus we would appreciate if you could spare some time to help us [like 10 mins?] and email us. Tks.
Chris
AWESOME !!!!!
Chris,
There may be a number of reasons for the behavior you are describing:
- IMU board is slightly tilted and the offsets are not correctly set in the source code.
- Motors are not positioned correctly. It is important that the motors are pointing straight up and are equally spaced from the center. Try to modify the tilt and spacing.
- ESCs are not configured with the same min and max throttle values. This will result in different RPM for the same commended value, if you have an RPM meter try to measure.
- Last but not least, is mechanical vibration. In my case I had to replace one of the motors because it was causing way too much vibration. The cheap motors I used are not well balanced. Vibration messes with the accels causing an offset on the absolute reference used in the Kalman filters. You might also want to adjust the Kalman filter and averaging filter parameters. The problem is that the vibration is usually well above the sampling frequency of the accels making it impossible to filter. So my advise is do whatever you can to eliminate vibration (change motors, different mounting of the IMU board etc).
Hope this helps.
Vassilis
Hi Vassilis,
I have been following you IMU and Paparazzi developments. I have code a DCM version IMU from DIYDRONES into LPC2138 and intend to use the PPZ TWOG board to fly my IMU Gyro, Accel, Magnetometer, SCP1000 etc. But before I do I am trying the Paparazzi UAV system and I am sure impressed. I have a question.
My airframe is a flying wing, but setup based on standard Funjet configuration file. I have flown it successfully in AUTO1 with stabilization straight and level flight, but I don’t have much manual RC control stick control while in AUTO1. I am using XY plus vertical IR sensors. Just wondered if you have been able to have much manual control in AUTO1 mode. I thought I could fly it normally in AUTO1 mode and when I release the RC sticks it stabilizes…. much like the FMA CoPilot….. But I really don’t have enough elevon movement in AUTO1 mode from the RC sticks to really fly it… the elevons barely move. I have more than enough elevon control surface movement in MANUAL mode. In AUTO1 and AUTO2 I have the same max movements in elevons when controlled by PPZ and IR sensors… but very little when I command it from RC control sticks.
When I disable the vertical IR sensor ADC, I have more elevon surface movement in AUTO1 mode…
Any ideas?
Could this be a calibration issue with the vertical IR sensor being more sensitive than the XY sensors?
Would I need to change the line
in the airframe config file section
so that the vertical sensor gain is changed to matched XY sensor?
I could email you my XML files if that would help.
Thanks
Sean
I can’t get smooth flight in AUTO2 with my Easystar using a TWOG.
What tuning parameters did you use?
Can you show me the airframe file?
Mark,
You can download the airframe configuration for my EasyStar from: http://vrhome.net/vassilis/paparazzi/
Sean,
You can download the airframe configuration for my EasyStar from: http://vrhome.net/vassilis/paparazzi/
In AUTO1 the autopilot should behave as well (if not better) as the FMA copilot. Assuming that you have your IR sensors setup correctly, it may be that your roll and pitch gains are too low. You may want to adjust the following parameters: PITCH_PGAIN and ROLL_ATTITUDE_GAIN, both of them should have negative values.