You can open this file in chrome for a better viewing experience.
Point the browser to
file:///home/robotlearning2/infants/notes.md
Windows machine log in info: username: "ut austin" (you will need to use "" cause of the spacebar) password: 1234 To ssh: ssh "ut austin"@192.168.253.101
Linux machine log in info:
username: robotlearning2
password: robotlearning2
- Ensure time-sync is good using Verifying NTP time sync
- Open a terminal on linux machine.
- Check if NAS (synology) is mounted:
run:
mountpoint -q ~/synology-tuli && echo "Mounted" || echo "Not mounted"If output is "Not mounted": Run:sudo mount -t nfs 192.168.253.1:/volume1/tuli ~/synology-tuli(password: robotlearning2 ) - Perform calibration by following the calibration guide
- Note: change settings on Qualisys as follows:
- Resolution: 1080p
- Aspect Ratio: 16:9
- FPS:
25 Hz
cd ~/infants/- [Optional]
arecord -lCheck what card # does "USB Audio" shows and ensure rs_cam.launch has that. - Launch the cameras:
./infants/scripts/start_all.sh - Verify that they start with
./infants/scripts/check_cams.shYou should see all 6 camera topics (color image raw and aligned depth image raw for cameras L, M, R) and audio!
topic rate min_delta max_delta std_dev window
===========================================================================================
/cam_L/color/image_raw 29.93 0.02236 0.04497 0.003658 141
/cam_L/aligned_depth_to_color/image_raw 29.93 0.02354 0.04423 0.003293 141
/cam_M/color/image_raw 29.94 0.02401 0.04426 0.003387 142
/cam_M/aligned_depth_to_color/image_raw 29.93 0.02458 0.04402 0.003249 142
/cam_R/color/image_raw 29.99 0.02847 0.03942 0.001995 141
/cam_R/aligned_depth_to_color/image_raw 29.99 0.02894 0.039 0.001779 141
/audio/audio 38.28 0.01948 0.0305 0.003658 141 if audio miss, run: pulseaudio -k and then rerun step 4 and 5.
6. Activate the virtualenv, source ~/envs/infants/bin/activate
7. Run the experiment script: python infants/experiment/experiment_driver.py
8. It will prompt for subject ID, task name, and condition ID.
Make sure the subject ID 3 digits. Example (1) write 001 for (2) write 002
Subject ID should be an integer. Task should be in [bang, slide, hammer].
- Press SPACEBAR to start recording that trial.
- Start recording on Qualisys side. Choose appropriate name
- When done, stop Qualisys recording
- Press SPACEBAR to stop recording on linux
- Say
[y/n]to keep trial or delete.
- Transfer Qualisys trial files (C3D, TSV, Miqus AVIs) into each
trial_*folder:
python infants/scripts/organize_trial_session.py \
--session 2026-08-18_16-10-10 \
--infant 050- Put marker data into rosbag
python infants/scripts/process_marker_c3d.py \
data/2026-08-18_16-10-10/trial_001
# optional: --num-markers 500- Visualize on Mocap RGB video (markers overlaid)
python infants/scripts/viz/overlay_miqus_markers.py \
--trial-dir data/2026-08-18_16-10-10/trial_001 \
--calibration-dir data/calibration_data/26_08_18_050- Visualize on RealSense RGB video (markers overlaid)
Basic (no audio, OpenCV window only):
python infants/scripts/viz/overlay_realsense_markers.py \
--trial-dir data/2026-08-18_16-10-10/trial_001 \
--calibration-dir data/calibration_data/26_08_18_050 \
--camera L \
--save-mp4Optional flags:
--audio— play bag/audio/audiowhile showing frames (paced to bag time)--save-mp4— writevisualizations/realsense/realsense_marker_overlay_<L|M|R>.mp4(includes audio if present)--output /path/to/out.mp4— custom MP4 path (same as--save-mp4but choose the name)--no-display— skip the OpenCV window (useful with--save-mp4only)
Example with audio + MP4 export:
python infants/scripts/viz/overlay_realsense_markers.py \
--trial-dir data/2026-06-29_15-03-28/trial_001 \
--calibration-dir data/calibration_data/26_06_29_infant_017 \
--camera L \
--audio \
--save-mp4- Visualize in RViz (point clouds ± markers)
Basic:
python infants/scripts/viz/run_trial_viz.py \
--bag /home/robotlearning2/infants/data/2026-06-29_15-03-28/trial_001/trial_ros_combined.bag \
--cameras L \
--markers \
--calib-config data/calibration_data/26_06_29_infant_017/calibration_markers.yamlAdd --audio to also start audio_play so bag audio plays with RViz (same bag must contain /audio/audio).
Add --record to capture the screen (RViz + your view moves) via record_rviz_screen.sh:
python infants/scripts/viz/run_trial_viz.py \
--bag /home/robotlearning2/infants/data/2026-06-29_15-03-28/trial_001/trial_ros_combined.bag \
--cameras L \
--markers \
--calib-config data/calibration_data/26_06_29_infant_017/calibration_markers.yaml \
--audio \
--recordOptional recording flags:
--record-output /path/to/out.mp4— custom output path (default:recordings/rviz_<bag>_<timestamp>.mp4)--record-delay 2— seconds to wait for RViz before starting capture (default: 2)
Note: --record grabs the screen, then muxes bag /audio/audio into the MP4 (aligned using --record-delay vs --bag-delay). Live speaker playback still needs --audio.
Manual audio-only test (without the viz scripts):
# Terminal 1
roslaunch audio_play play.launch
# Terminal 2
rosbag play --clock /path/to/trial_ros.bagOld
- RViz + rosbag playback (restores wall time on exit / Ctrl+C):
./infants/scripts/viz/roslaunch_restore_wall_time.sh launch/visualize_data.launch bag_file:=/absolute/path/to/trial_ros.bag - Record RViz screen:
./infants/scripts/viz/record_rviz_screen.sh /home/robotlearning2/infants/recordings/rviz_$(date +%Y%m%d_%H%M%S).mp4 "$DISPLAY" - Stop recording with
Ctrl+C.
rqt_bag <path to bag> and open a bagfile with the gui. This will show images but audio will not play properly.
rqt_image_view first and then rosbag play
You can replay the audio with
c
roslaunch audio_play play.launch
rosbag play <path to bag>
To rsync data:
sudo rsync -avh --progress --no-owner --no-group --ignore-existing 2026-07-17_10-57-17 /home/robotlearning2/synology-tuli/
- Make sure it is switched on (blue light should be visible on the big black box)
- http://192.168.253.1:5000/#/signin
- https://finds.synology.com/#
- http://169.254.68.74:5000/
Troubleshooting:
If you see "no new messgaes"
rosnode kill /play_* # or kill the trial_viz / visualize launch
rosparam set /use_sim_time false