USC-ACTLab / crazyswarm

A Large Quadcopter Swarm
MIT License
320 stars 315 forks source link

crazyswarm LPS issue #113

Closed podnerogato closed 4 years ago

podnerogato commented 5 years ago

Hi, I am using ROS kinetic on Ubuntu 16.04, 5 crazyflies, LPS and latest CFClient and Crazyswarm repositories. I have followed your crazyswarm instructions and everything went well. Updated all the firmwares (both CF firmwares, CF radio and LPS) with no problems. I had the addresses set up, and scripts modified as suggested. Chooser.py is functioning as well. I can't get TDoA3 to work. At first i could change mode to TDoA3 using CFClient but the position of the crazyflies would have drifted into hard-coded limits(100,100,100). Then i have updated LPS nodes firmware once again and had setup redone. Now I can't even switch to TDoA3 because all the anchors are red. TWR and TDoA2 are functioning and returning the right position. I execute source ros_ws/devel/setup.bash roslaunch crazyswarm hover_swarm.launch with no errors and rviz starts but I get empty space without crazyflies. Rviz closes with no errors. When i execute "start" on xbox360 controller crazyflies just flip uncontrollably. All the crazyflie were facing +X direction, as suggested. I hope I'm at the right address with this question. Thanks!

EDIT: Using kalman estimator (instead of KalmanUSC), TDoA2 and modifying initial position I have managed to achieve take off and position hold in X and Y, but the Z coordinate from allCrazyflies.yaml is not achieved. When I use KalmanUSC and TDoA2 everything just goes crazy and crashes.

whoenig commented 5 years ago

For LPS, you should use kalman and not kalmanUSC. @ataffanel, can you comment on what else to do for TDoA3?

podnerogato commented 5 years ago

Thank you for your answer. I'll give you a bit more info about the problem. All anchors are green in TWR and TDoA2 (as shown on images below). When I select current mode and click Change mode to TDoA3 all anchors remain green in current mode except for anchor 2 that blinks red for a second but turns to green immediately. In some other cases anchors in current mode turn red. Anchors in TDoA3 never turn green and position always drifts to hardcoded limits. Positioning and switching (between) in TWR and TDoA2 works fine. Thanks! tdoa2_crop twr_crop

EDIT: Another Issue accrued while trying to do figure8 test. Here is the console output: Traceback (most recent call last): File "figure8_csv.py", line 5, in <module> from pycrazyswarm import * File "/home/vice/crazyswarm/ros_ws/src/crazyswarm/scripts/pycrazyswarm/piecewise.py", line 1, in <module> from .cfsim import cffirmware as firm ImportError: cannot import name cffirmware I get the same message for python figure8_csv.py --sim and python figure8_csv.py. Is the figure8 supported when using LPS? Thank you!

whoenig commented 5 years ago

Yes, figure8 should work with LPS. The error message indicates that there was an issue during installation. Does ./build.sh run through without errors?

podnerogato commented 5 years ago

Hi, there are no errors.

whoenig commented 5 years ago

Errors might be hard to spot, in particular, if these lines fail, that would explain your issue. Also, make sure that "python" actually refers to python2 and not python3 (which should be the default for Ubuntu 16.04).