Changes between Version 31 and Version 32 of Other/Summer/2025/LLM-Based


Ignore:
Timestamp:
Jul 24, 2025, 2:34:36 PM (3 days ago)
Author:
jerryh
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Other/Summer/2025/LLM-Based

    v31 v32  
     1
    12====== LLM-Based VR Activity Recognition ======
    23
     
    5657 - Measures angular velocity around 3 axes: pitch, roll, and yaw
    5758
     59Additionally, we tried graphing one of our experimental trials in a 3d graph.
    5860
     61[[Image(3d-head-spin-graph.png, height=350, width=700)]]
     62
     63Drifting was a major issue when graphing and we still haven't fully fixed it.
    5964=== Week 5: [[https://docs.google.com/presentation/d/1e0RMR1sGx5zReXR4hQsTyly3uDP29hHNLIj4iPaGZzA/edit?slide=id.g36334886556_0_343#slide=id.g36334886556_0_343|Slides]]
    6065
     
    6671 - Stand up
    6772 - Walk
    68 [[Image(stand.gif, width=250, height=400)]][[Image(walk.gif, width=250, height=400)]][[Image(nod.gif, width=250, height=400)]][[Image(turnright.gif, width=250, height=400)]][[Image(turnleft.gif, width=250, height=400)]]
     73[[Image(stand.gif, width=250, height=400)]][[Image(walk.gif, width=250, height=400)]][[Image(nod.gif, width=250, height=400)]]
     74[[Image(turnright.gif, width=250, height=400)]][[Image(turnleft.gif, width=250, height=400)]]
    6975----
    7076
    71 The data we collected all had a 2-second interval with faulty and incomplete data. We were unable to determine what caused this error, but every sample had this error. All data after this 2-second interval was accurate, so we decided to trim out the 2 seconds of faulty data and only take the last 5000 (effectively 5 seconds) samples.  After that, we used the movmean function in MATLAB to smooth out the data.
     77The data we collected all had a 2-second interval with faulty and incomplete data. We were unable to determine what caused this error, but every sample had this error. All data after this 2-second interval was accurate, so we decided to trim out the 2 seconds of faulty data and only take the last 5000 (effectively 5 seconds) samples.
    7278
    73 Before segmentation:
     79Before
    7480
    7581[[Image(Screenshot 2025-07-23 105427.png​, width=500, height=250)]]
     
    7985[[Image(Screenshot 2025-07-23 103349.png​​, width=500, height=127)]]
    8086
    81 After segmentation:
     87----
     88After
    8289
    8390[[Image(Screenshot 2025-07-23 110334.png, width=500, height=250)]]
    84 
    85 After segmentation and smoothing:
    86 
    87 [[Image(Screenshot 2025-07-24 102937.png​, width=500, height=250)]]
    88 
    8991
    9092Questions or contributions? Please contact us.