Developer
paul riseborough
  • Victoria
  • Australia
Share

Paul riseborough's Friends

  • Dries Hulens
  • Brian Schwartz
  • Stuart Wishart
  • Basil
  • Dominik Strnad
  • Nikola Rabchevsky
  • Michael Day
  • Sam Tabor
  • Johnny Ace
  • bahloul
  • iskess
  • Faraz Ahmad
  • Szilard
  • David Anders
  • einma1to

paul riseborough's Groups

paul riseborough's Discussions

Repeated mid-air APM reboot - only happens at one flying site - possible weather radar interference

Started this discussion. Last reply by paul riseborough Apr 21, 2013. 13 Replies

I have three flying sites that I use depending on weather and availability. My primary flying site is located only 800 meters from a weather radar and I have been having repeated in-air reboots when…Continue

Please help - 3DR USB radio will upload firmware but won't enter command mode

Started this discussion. Last reply by Marcin Krawczyk May 31, 2013. 7 Replies

I have a 3DR 915 MHz v1.1 USB interface radio that will not enter command mode using the 3DR radio config tool and does not respond to any commands. The unit was communicating whenfirst received, but…Continue

Tags: of, out, now, I'm

 

paul riseborough's Page

Latest Activity


Developer
paul riseborough replied to John Oberholtzer's discussion Pilot Erorr/Solo Error? in the group 3DR SOLO
"If you rename remove the .2 extension on the solo.tlog.2 file it can be replayed and plotted using mission planner. Doing this shows that the GPS thought it was moving away from the house to the NE, which caused the copter to pitch back and fly SW…"
Friday

Developer
paul riseborough left a comment for Ravi Pinnamaneni
"The Simulink model is a very early version and and the ArduPilot code has diverged significantly. I will not be doing further EKF implementations in Simulink given I have acquired the skills to be able to code in C++ and very few people in the open…"
Jun 4
Ravi Pinnamaneni left a comment for paul riseborough
"Also, I am using your 24-state EKF model and I understand that there are differences in the states between the model and the actual EKF implementation in the ardupilot code. -Ravi"
Jun 4
Ravi Pinnamaneni left a comment for paul riseborough
"Hi Paul, I am working with your Simulink model to simulate the EKF output using data from a GPS-based flight I did a couple of weeks back. I am successfully able to run the model and generate the output. However, the model output (particularly NED…"
Jun 4

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"Given the gyro has 10x the change in zero rate offset with temperature that the gyro manufacturer says is 'typical' which results in large heading errors and marginal flight characteristics with standard software and you have what is…"
May 26

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"Forgot to mention you could also increase the value of EKF_GBIAS_PNOISE to make the EKF learn the gyro bias change faster."
May 25

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"Andy, The magnetometer data looks good. The underlying cause is that the second rate gyro on your board (the L3GD20) has excessive temperature sensitivity on the Z axis. It is showing about a 3.5 deg/second change in bias for a temperature change…"
May 25

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"Andy, At your latitude keeping your magnetometer level relative to the Pixhawk board is important, not just the yaw angle that most people concentrate on. The IMU can handle yaw rates much higher than what you have. What happens with mag errors…"
May 12

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"I wouldn't enable in-flight learning with AC3.2.1 as the method used has not worked well for copters in the past. When 3.3 is released it will do compass learning using the EKF and save the offsets. The declination being used in your log…"
May 10

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"Yes, you should investigate further as these are large yaw errors. The large and sudden increases in EKF innovations and DCM ErrYaw when it yaws are most likely caused by bad compass offsets. The solution is to repeat the mag cal (compass dance).…"
May 9

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"Have a look towards the bottom of the EKF wiki http://plane.ardupilot.com/wiki/common-apm-navigation-extended-kalman-filter-overview/#ekf4 I agree ErrYaw (and ErrRP) are confusing because they are in the same group as the attitude control…"
May 8

Developer
paul riseborough replied to Gonzalo Garat's discussion APM 3.2 QUADCOPTER Err: EKF_CHECK-2 / Err: FAILSAFE_ EKF-1
"The  primary cause is bad compass errors. Look at the EKF4.SM, EKF4.SMY which are EKF measures of X and Y compass errors. Anything over 1 is  red light fail, anything over 0.5 is a yellow light caution. When the Copter is hovering, the…"
May 7

Developer
paul riseborough replied to Randy's discussion Copter-3.3 beta testing
"No I haven't flown that mount, but I have heard good reports of its performance in copters provided suitable isolators are fitted. For planes it has a potential problem that g forces from hand or catapult launches cause the mount to rotate due…"
May 6

Developer
paul riseborough replied to Randy's discussion Copter-3.3 beta testing
"Ravi, It is a common misconception that vibration mounts need to damp vibration and that a more damped material is required. The job of the mount is to isolate the autopilot from the vibration of the airframe by not transmitting…"
May 6

Developer
paul riseborough replied to Randy's discussion Copter-3.3 beta testing
"The relevant data is logged in the UBX1, UBX2 and UBX3 packets The definition of the proprietary UBlox messages in these packets (eg noisePerMS) is found in the following UBlox manual (see page 161 for…"
May 4

Developer
paul riseborough replied to Randy's discussion Copter-3.3 beta testing
"Didn't you get black foam pads with the Pixhawk?"
May 4

Profile Information

Please tell us a bit about your UAV interest
Fixed wing
Hometown:
hoppers crossing

Comment Wall (12 comments)

At 5:07am on September 24, 2014, liuqiang said…

dear sir

          I download apm:arduplane, I want to use AHRS_NavEKF on APM2,BUT I found ahrs update in AHRS_NavEKF.cpp not run,only run update in AHRS_DCM,I hope your help,thank you.
          

At 7:31pm on September 24, 2014, liuqiang said…
I define ap_ahrs_navekf ahrs(ins,baro,gps),but the IDE compile tell me ap_ahrs_navekf not a name type and compilication terminated,why?
At 11:58pm on September 24, 2014,
Developer
paul riseborough
said…

You won't get EKF working for APM2 (the microprocessor is too slow). Follow the instructions on https://github.com/priseborough/ardupilot to build for px4.

At 5:00am on September 25, 2014, liuqiang said…

I run AHRS_TEST,and use AP_AHRS_NAVEKF instead of AP_AHRS_DCM,but  compile tell me "AHRS_Test.pde:56:1: error: 'AP_AHRS_NavEKF' does not name a type". Could you tell me where should be set about AOP_AHRS_NavEKF?

At 6:16pm on September 25, 2014, liuqiang said…
I do not run the code on apm2,just build it and find CLASS AP_AHRS_NAVEKF can not define a variables,tell me AP_AHRS_NAVEKF not a name type,why?
At 12:31am on October 16, 2014, liuqiang said…
I buy one pixhawk board,and fly on fixwing ,I find the output of AZ in EKF2 (z accelerometer bias of IMU1) is too big,it is gradually become larger ,max value about 30,AX an AY max value about 4,I feel attitude angle measurement have a bit problem,how to solve this situation!
At 10:39am on April 15, 2015, Aghilas said…

Dear Paul, we are two master thesis students currently working with a PX4 and arducopter 3.2.1. Our project is about offboard camera tracking and fault diagnosis on the quadcopter. We have looked at the arducopter code and set the AHRS_EKF_USE to 1 but are unsure about which EKF is really running as there are a couple of them, (AP_NavEKF, AP_AHRS_NavEKF), could you possibly give us a pointer?

We are also very much interested in your simulink model, but we see that it needs airspeed sensor data (it was probably implemented with fixed wing aircraft in mind?) which we do not have, is there any way to replace those measurements with barometer data?

At 4:30pm on April 15, 2015,
Developer
paul riseborough
said…

It is running the code in AP_NavEKF. AP_AHRS_NavEKF is jsut there so that the EKF can be accessed by the rest of the code via the AHRS object.

The simulink model can be used without airspeed data by removing the airspeed fusion block (or just setting the useAirspeed input to false).

Fusion of  bnon GPS height measurement  is already available  in the simulink model.

At 6:24am on April 21, 2015, Aghilas said…

Dear Paul,

Thanks for your time and reply.

Sorry for many stupid questions, but I have one regarding the simulink model "InertialNav", previously you said that the fusion of baro GPS height measurement is available for the model. In the Read Measurement block you use RelAlt for height measurement, what is that measurement exactly representing? Is that an estimated value from barometer and accelerometer ?

Thanks in advance

At 8:21pm on April 21, 2015,
Developer
paul riseborough
said…

It is height measured relative to the origin of the local NED  axis system. it could come from baro, GPS, or any other sensor capable of height measurement.

You need to be a member of DIY Drones to add comments!

Join DIY Drones

 
 
 

DIY Drones Monthly
Newsletter

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Pylon Race, is here

© 2015   Created by Chris Anderson.

Badges  |  Report an Issue  |  Terms of Service