Purdue researchers introduce a flexible touch sensor for robots

The team says a gel cartridge lets the sensor estimate force along three axes at each sensing point. In a robot test, the researchers found that adding touch to vision raised successful peg insertions from 6 of 15 trials to 13 of 15.

Numbers in brackets link to the sources listed below.

Purdue’s MARS Lab introduced LuxTac, an optical touch sensor for robots, on its project website[5LuxTac projectluxtac/luxtac.github.ioOpen sourceSources [5]]. In a preprint submitted to arXiv on 17 November, the researchers describe a flexible pad about 3.5 mm thick made with off-the-shelf parts[1arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [1]].

The researchers say a plain gel cartridge measures force pressing into the pad[1arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [1]]. As an object presses deeper into the silicone, less light from an LED returns to the detector beneath it[1arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [1]]. With a second cartridge, colored walls change the color of the returning light in response to sideways force, letting the team estimate force along three axes at each detector[3arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [3]].

In tests with three indenter shapes, the authors reported a median error of 4.8 degrees in the estimated direction of sideways force[1arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [1]]. For the robot trials, they used pressure as the sole touch input[3arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [3]].

In cube sorting, successes rose from 5 to 6 in 15 trials, a change the authors found statistically insignificant[3arXivLuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array with Distributed Three-Axis Force Sensing for Tactile Robot LearningOpen sourceSources [3]].

Sources

  1. 1
  2. 2
    LuxTac: A Flexible, Scalable, and Open-Source Optical Taxel Array

    PrimaryLuxTac projectAccessed

  3. 3
  4. 4
    AdeeshDesai/LuxTac-Sim

    PrimaryLuxTac projectAccessed

  5. 5
    luxtac/luxtac.github.io

    PrimaryLuxTac projectAccessed

  6. 6
    Types of internal forces

    IndependentUniversity of AlbertaAccessed

ITANU’s newsroom system produced this article with AI and checked it against the sources listed above. How we work

ITANU Magazine

AI, computing and science news