File size: 3,010 Bytes
793890c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
---
license: mit
tags:
- LeRobot
- robotics
- tactile
- xarm
task_categories:
- robotics
---

# xarm_tube_binmarker9

xArm7 phone-teleop dataset for **tube** (`"pick up the tube and put it in the nearest slot"`) with the **binmarker9**
tactile overlay burned into the camera frames: **9 arrows per gripper finger**
(one at each of the 9 forward-kinematics-projected sensor pad cells) with **fixed length AND fixed direction**
(the sensor-frame pad normal - a rigid "flag" on the finger), drawn **iff that
finger's binary contact state is ON**.

The overlay carries **only the per-finger contact bit** - NO force magnitude
and NO force direction.

## Why this variant exists

`binmarker9` is the **9-arrow arm** of a matched pair that holds information
content constant and varies only presentation:

| Variant | Arrows/finger | Contact bit | Direction | Magnitude |
|---|---|---|---|---|
| `binmarker` | 1 | YES | no | no |
| `binmarker9` | 9 | YES | no | no |

Because all 9 arrows carry the identical fixed direction and length, the
rendering is **provably independent of the tactile values**: 0 max pixel change
across 300 randomized force readings at a fixed pose. The 9-arrow arm paints
~3.1x the arrow ink of the 1-arrow arm, so `binmarker` vs `binmarker9` measures
sensitivity to overlay presentation / salience rather than to force information.

Note on appearance: the 9 arrows share one direction and the pad cells sit ~1 cm
apart, so they render as a thick overlapping marker rather than 9 individually
countable arrows. That is forced by the ablation - 9 arrows carrying no direction
or magnitude information must be 9 identical arrows.

## Contact detection (how the threshold was tuned)

Per-finger binary contact from the 3x3x3-axis tactile array:

- **Normalization offset** = median over known-idle frames (first 10 frames of
  each episode, post-homing, gripper open). The naive whole-dataset median is
  contact-polluted on tasks that spend most frames in contact (charger: 60%
  closed frames made idle read as a fake magnitude of ~1.2).
- **Metric** = max over the 9 cells of the normalized cell-vector L2 norm
  ("max-cell"). Idle p99.5 <= 0.24 on every task, real contacts >= ~0.3.
- **Hysteresis + debounce** (Schmitt trigger): ON after 2 consecutive frames
  >= 0.25, OFF after 3 consecutive frames < 0.18, state reset at episode
  boundaries.

Identical detector and thresholds to `binmarker`, so the two arms gate on the
exact same contact signal.

## Contents

- 97 episodes / 18408 frames @ 10 Hz, LeRobot v2.1 format
- `image` (224x224x3, agent view) and `wrist_image` (224x224x3) - overlay burned in
- `state` (8-dim: ee_pos_m, ee_axis_angle_rad, grasp, grasp)
- `actions` (7-dim world-frame deltas: dxyz_m, daxis_angle_rad, grasp in {-1,+1})

Same source recordings as every other overlay variant of this task - only the
burned-in pixels differ.

## Paired model

[EdwardoSunny/pi05_xarm_tube_binmarker9_lora](https://huggingface.co/EdwardoSunny/pi05_xarm_tube_binmarker9_lora)