R11: indoor track 2026-10-08 (walk/jog/run intervals); sessions.csv + README table to 11 sessions (part 2)
Browse files- README.md +3 -2
- data/S01/R11/level/imu-right_foot-4444.csv +0 -0
- data/S01/R11/level/imu-wrist-1111.csv +0 -0
- data/S01/R11/level/meta-summary.json +91 -0
- figures/S01/R11/analysis-foottilt.png +3 -0
- figures/S01/R11/analysis-gait.png +3 -0
- figures/S01/R11/analysis-groundtruth.png +3 -0
- figures/S01/R11/data-fit3.png +3 -0
- figures/S01/R11/data-imu.png +3 -0
- figures/S01/R11/derived-spectrum.png +3 -0
- figures/S01/R11/overview.png +3 -0
- sessions.csv +1 -0
README.md
CHANGED
|
@@ -69,7 +69,7 @@ this dataset's public data.
|
|
| 69 |
## The sessions
|
| 70 |
|
| 71 |
<!-- glance:start (generated by scripts/make_catalog.py) -->
|
| 72 |
-
**
|
| 73 |
<!-- glance:end -->
|
| 74 |
|
| 75 |
<!-- sessions:start (generated by scripts/make_catalog.py from sessions.csv -- do not edit by hand) -->
|
|
@@ -85,8 +85,9 @@ this dataset's public data.
|
|
| 85 |
| 2026-09-30 | [R08](figures/S01/R08/overview.png) | outdoor | 45 | 45 | | 4 | | 0.7 | 99.2 | 0.31 |
|
| 86 |
| 2026-10-03 | [R09](figures/S01/R09/overview.png) | outdoor | 20 | 41 | 5.00 | 4 | watch | 0.0 | 99.9 | 0.01 |
|
| 87 |
| 2026-10-05 | [R10](figures/S01/R10/overview.png) | outdoor | 33 | 33 | 4.46 | 4 | watch | 1.9 | 98.0 | 0.36 |
|
|
|
|
| 88 |
|
| 89 |
-
*
|
| 90 |
<!-- sessions:end -->
|
| 91 |
|
| 92 |
Every run's figures sit in `figures/<subject>/<session>/`: `overview.png` (start there), `data-*.png` for
|
|
|
|
| 69 |
## The sessions
|
| 70 |
|
| 71 |
<!-- glance:start (generated by scripts/make_catalog.py) -->
|
| 72 |
+
**11 sessions, 1 participant, 5.6 h of recording, 34 km outdoors, 2026-09-07 to 2026-10-08.**
|
| 73 |
<!-- glance:end -->
|
| 74 |
|
| 75 |
<!-- sessions:start (generated by scripts/make_catalog.py from sessions.csv -- do not edit by hand) -->
|
|
|
|
| 85 |
| 2026-09-30 | [R08](figures/S01/R08/overview.png) | outdoor | 45 | 45 | | 4 | | 0.7 | 99.2 | 0.31 |
|
| 86 |
| 2026-10-03 | [R09](figures/S01/R09/overview.png) | outdoor | 20 | 41 | 5.00 | 4 | watch | 0.0 | 99.9 | 0.01 |
|
| 87 |
| 2026-10-05 | [R10](figures/S01/R10/overview.png) | outdoor | 33 | 33 | 4.46 | 4 | watch | 1.9 | 98.0 | 0.36 |
|
| 88 |
+
| 2026-10-08 | [R11](figures/S01/R11/overview.png) | indoor | 14 | 14 | 2.03 | 4 | watch | 0.0 | 99.9 | 0.06 |
|
| 89 |
|
| 90 |
+
*11 sessions. Click a session for its overview figure; every column (setup, versions, heart rate, capture quality) is in `sessions.csv`.*
|
| 91 |
<!-- sessions:end -->
|
| 92 |
|
| 93 |
Every run's figures sit in `figures/<subject>/<session>/`: `overview.png` (start there), `data-*.png` for
|
data/S01/R11/level/imu-right_foot-4444.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
data/S01/R11/level/imu-wrist-1111.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
data/S01/R11/level/meta-summary.json
ADDED
|
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"schema_version": 2,
|
| 3 |
+
"app": {
|
| 4 |
+
"versionName": "0.1.13-exp"
|
| 5 |
+
},
|
| 6 |
+
"session": {
|
| 7 |
+
"activityCode": "BTIMU_RAW",
|
| 8 |
+
"sensorType": "BT_IMU",
|
| 9 |
+
"startedAtIso": "1970-01-01T00:00:00.497000+00:00",
|
| 10 |
+
"startedAtMs": 497,
|
| 11 |
+
"endedAtMs": 807983,
|
| 12 |
+
"durationMs": 807486,
|
| 13 |
+
"endReason": "clean"
|
| 14 |
+
},
|
| 15 |
+
"sensors": [
|
| 16 |
+
{
|
| 17 |
+
"shortMac": "4444",
|
| 18 |
+
"sensorLabel": "right_foot",
|
| 19 |
+
"transport": "DONGLE",
|
| 20 |
+
"firmwareVersion": "0.102.0"
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"shortMac": "8888",
|
| 24 |
+
"sensorLabel": "left_foot",
|
| 25 |
+
"transport": "DONGLE",
|
| 26 |
+
"firmwareVersion": "0.102.0"
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"shortMac": "0000",
|
| 30 |
+
"sensorLabel": "back",
|
| 31 |
+
"transport": "DONGLE",
|
| 32 |
+
"firmwareVersion": "0.102.0"
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"shortMac": "1111",
|
| 36 |
+
"sensorLabel": "wrist",
|
| 37 |
+
"transport": "DONGLE",
|
| 38 |
+
"firmwareVersion": "0.102.0"
|
| 39 |
+
}
|
| 40 |
+
],
|
| 41 |
+
"transports": [
|
| 42 |
+
{
|
| 43 |
+
"type": "DONGLE",
|
| 44 |
+
"shortMac": null,
|
| 45 |
+
"firmwareVersion": "0.102.0"
|
| 46 |
+
}
|
| 47 |
+
],
|
| 48 |
+
"settings": {
|
| 49 |
+
"sampling_rate_hz": "100.0",
|
| 50 |
+
"sensor_count": "4",
|
| 51 |
+
"time_sync": "true"
|
| 52 |
+
},
|
| 53 |
+
"files": [
|
| 54 |
+
{
|
| 55 |
+
"name": "imu-wrist-1111.csv",
|
| 56 |
+
"modality": "ACCEL_GYRO",
|
| 57 |
+
"shortMac": "1111",
|
| 58 |
+
"rotationCount": 1,
|
| 59 |
+
"units": "us, m/s^2, rad/s"
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"name": "imu-left_foot-8888.csv",
|
| 63 |
+
"modality": "ACCEL_GYRO",
|
| 64 |
+
"shortMac": "8888",
|
| 65 |
+
"rotationCount": 1,
|
| 66 |
+
"units": "us, m/s^2, rad/s"
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"name": "imu-back-0000.csv",
|
| 70 |
+
"modality": "ACCEL_GYRO",
|
| 71 |
+
"shortMac": "0000",
|
| 72 |
+
"rotationCount": 1,
|
| 73 |
+
"units": "us, m/s^2, rad/s"
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"name": "imu-right_foot-4444.csv",
|
| 77 |
+
"modality": "ACCEL_GYRO",
|
| 78 |
+
"shortMac": "4444",
|
| 79 |
+
"rotationCount": 1,
|
| 80 |
+
"units": "us, m/s^2, rad/s"
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"name": "meta-summary.json",
|
| 84 |
+
"modality": "META",
|
| 85 |
+
"shortMac": null,
|
| 86 |
+
"rotationCount": 1,
|
| 87 |
+
"units": null
|
| 88 |
+
}
|
| 89 |
+
],
|
| 90 |
+
"notes": null
|
| 91 |
+
}
|
figures/S01/R11/analysis-foottilt.png
ADDED
|
Git LFS Details
|
figures/S01/R11/analysis-gait.png
ADDED
|
Git LFS Details
|
figures/S01/R11/analysis-groundtruth.png
ADDED
|
Git LFS Details
|
figures/S01/R11/data-fit3.png
ADDED
|
Git LFS Details
|
figures/S01/R11/data-imu.png
ADDED
|
Git LFS Details
|
figures/S01/R11/derived-spectrum.png
ADDED
|
Git LFS Details
|
figures/S01/R11/overview.png
ADDED
|
Git LFS Details
|
sessions.csv
CHANGED
|
@@ -9,3 +9,4 @@ S01,2026-09-28,R07,outdoor,pavement,,,4,Back;Left Foot;Left Wrist;Right Foot,acc
|
|
| 9 |
S01,2026-09-30,R08,outdoor,pavement,,,4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,hand,post,False,True,True,False,False,False,False,44.8,44.8,recording,,,,,,,151.0,phone_steps,0.71,99.21,0.31,40.2,7.4,False,ALIGNED,back,100.0,98.0,24.5,19.5,4.14,4.08,28.4,25.7,7,92.2,"Phone carried in the hand. GPS off, so no pace profile; the barometer and step counter were still logged. The back sensor was loose and re-seated several times: trunk measures such as vertical oscillation are unreliable, the feet are unaffected. The pre-run calibration walk did not register on the feet, so the post-run calibration is used"
|
| 10 |
S01,2026-10-03,R09,outdoor,pavement,parkrun,"Victoria Park parkrun, Kitchener",4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,pack,pre,False,False,False,False,False,False,True,20.2,41.0,reported,reported,5.0,8.2,160.0,178.0,fit3,151.0,fit3,0.0,99.91,0.01,0.0,108.6,True,ALIGNED,,90.0,88.0,23.8,23.5,4.08,4.05,28.6,26.2,5,37.4,"Dongle on a short USB cable; the second phone was in the same pack, powered off. The recording covers the first ~18 min of a ~41 min event: the sensor data then stopped with no error while sensors and dongle still reported connected; The watch was in airplane mode with no GPS, so its distance is an estimate and reads well above the 5.00 km course"
|
| 11 |
S01,2026-10-05,R10,outdoor,pavement,,,4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,pack,pre,False,False,False,False,False,False,True,33.1,33.0,fit3,fit3,4.46,7.39,157.0,171.0,fit3,149.0,fit3,1.87,98.05,0.36,103.6,8.9,False,ALIGNED,,81.0,80.0,25.0,22.1,4.06,4.02,25.9,28.6,5,66.5,"Phone carried in the waist pack, with a second phone in the pack powered off. GPS off, so no pace profile. IMU data loss was 1.7-2.1 % per sensor (effective rate 97.8-98.2 Hz of 100 Hz): 428-539 gaps per sensor, every one a whole number of 50 ms packets; 58-67 % of gaps were a single packet (50 ms), 98-99 % were 4 packets (200 ms) or fewer, and the longest was 0.3-0.4 s. The gaps cluster in about seven short episodes (rate dipping to ~70 Hz, each lasting up to ~1.5 min) that hit all four sensors at once, with full rate in between. Three of the four IMU files end 5-9 s before the session end."
|
|
|
|
|
|
| 9 |
S01,2026-09-30,R08,outdoor,pavement,,,4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,hand,post,False,True,True,False,False,False,False,44.8,44.8,recording,,,,,,,151.0,phone_steps,0.71,99.21,0.31,40.2,7.4,False,ALIGNED,back,100.0,98.0,24.5,19.5,4.14,4.08,28.4,25.7,7,92.2,"Phone carried in the hand. GPS off, so no pace profile; the barometer and step counter were still logged. The back sensor was loose and re-seated several times: trunk measures such as vertical oscillation are unreliable, the feet are unaffected. The pre-run calibration walk did not register on the feet, so the post-run calibration is used"
|
| 10 |
S01,2026-10-03,R09,outdoor,pavement,parkrun,"Victoria Park parkrun, Kitchener",4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,pack,pre,False,False,False,False,False,False,True,20.2,41.0,reported,reported,5.0,8.2,160.0,178.0,fit3,151.0,fit3,0.0,99.91,0.01,0.0,108.6,True,ALIGNED,,90.0,88.0,23.8,23.5,4.08,4.05,28.6,26.2,5,37.4,"Dongle on a short USB cable; the second phone was in the same pack, powered off. The recording covers the first ~18 min of a ~41 min event: the sensor data then stopped with no error while sensors and dongle still reported connected; The watch was in airplane mode with no GPS, so its distance is an estimate and reads well above the 5.00 km course"
|
| 11 |
S01,2026-10-05,R10,outdoor,pavement,,,4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,pack,pre,False,False,False,False,False,False,True,33.1,33.0,fit3,fit3,4.46,7.39,157.0,171.0,fit3,149.0,fit3,1.87,98.05,0.36,103.6,8.9,False,ALIGNED,,81.0,80.0,25.0,22.1,4.06,4.02,25.9,28.6,5,66.5,"Phone carried in the waist pack, with a second phone in the pack powered off. GPS off, so no pace profile. IMU data loss was 1.7-2.1 % per sensor (effective rate 97.8-98.2 Hz of 100 Hz): 428-539 gaps per sensor, every one a whole number of 50 ms packets; 58-67 % of gaps were a single packet (50 ms), 98-99 % were 4 packets (200 ms) or fewer, and the longest was 0.3-0.4 s. The gaps cluster in about seven short episodes (rate dipping to ~70 Hz, each lasting up to ~1.5 min) that hit all four sensors at once, with full rate in between. Three of the four IMU files end 5-9 s before the session end."
|
| 12 |
+
S01,2026-10-08,R11,indoor,track,,,4,Back;Left Foot;Right Foot;Wrist,accel+gyro,100.0,5,DONGLE,0.102.0,0.102.0,0.1.13-exp,samsung SM-S926W,none,BEACON_TIMESYNC,hand,pre,False,False,False,False,False,False,True,13.5,14.0,fit3,fit3,2.03,6.9,,,,,,0.01,99.9,0.06,0.0,1.3,,MISALIGNED,,91.0,90.0,25.0,25.4,3.67,3.67,25.8,28.0,5,27.3,"Indoor track, training intervals alternating walking, jogging and running; phone carried in the hand. Phone GPS off, so no pace profile. IMU data loss was negligible (0.00-0.02 % per sensor). The beacon time sync did not take on three of the four sensors (stable offsets of roughly 0.2, 0.3 and 0.8 s against the wrist beacon for the whole session), so cross-sensor and bilateral timing metrics are withheld; per-sensor metrics are unaffected. The watch workout was started about half a minute after the recording and is a cross-check only."
|