Tactile Learning

Tactile Datasets

Supplementary material and dataset for the "Towards Effective Tactile Identification of Textures using a Hybrid Touch Approach", T. Taunyazov et al.

Dataset description

This dataset contains tactile sensor data 23 textures. The data are gathered from iCub's forearm tactile sensor for touch and slide. Each object has 3 folders:

1. slide_raw
2. touch_raw (only 10 samples)
3. touch_raw2

Each folder there contains 3 folders:

1. right_arm_encoders: encoder values of 15 dof of iCub's right arm
2. right_forearm_comp: tactile sensor values
3. skin_events: force values

These are objects' ids used in the paper:

• BT -> bathTowel
• CB -> cardBoard
• PM -> carpetNet
• CC -> cork
• CH -> cotton
• FM -> cuisionFoam
• CS -> denim
• ES -> eva
• LC -> fakeLeather
• FL -> felt
• FB -> fiberBoard
• MS -> metal
• RS -> paper1
• LM -> paper2
• PS -> polypropileno
• CT -> polypropilenoSmooth
• BM -> softMaterial1
• CP -> softMaterial2
• SS -> spongeWhiteSmall
• PF -> styrofoam
• WP -> thinPolypropylene
• BD -> woodHard
• YM -> yogaMat