Our project addresses the problem of slip detection in prosthetic hands, making a glove based prototype with the long-term goal of integrating the solution into an affordable prosthetic hand for upper-limb amputees made by Haifa 3D Association.
Current advanced prosthetic hands are extremely expensive (often tens of thousands of dollars) and therefore inaccessible to most users. They also involve high technological complexity, including multiple motors, sophisticated algorithms and neural or myoelectric sensing, which increases cost, operating complexity and maintenance, which can discourage users and reduce long-term adoption.
We propose the development of a slip-feedback system that is simple, low-cost, low-maintenance and accessible, providing practical and intuitive slip indication without ‘false positive alerts’.
The project began with a literature review of existing slip detection approaches, including mechanical rollers, force and pressure sensors, and lastly navigation sensors (optical, laser, IR) that were originally developed for computer mice. We derived a set of design requirements relevant to prosthetic use: the sensor system should be low cost, physically compact, low-power, robust to everyday handling, and sufficiently sensitive to detect early slip on common household surfaces without ‘false positives’.
On the basis of these criteria, and thorough testing, we selected two complementary sensing modalities for detailed investigation:
1.
High-resolution laser motion sensor capable of measuring micro-motion at the contact interface (ADNS 9800).
2.
Thin force/pressure sensors (FSRs) for estimation of normal grip force (FlexiForce 11kg).
The prototype platform consists of a glove with sensors mounted on its fingertips and palm, controlled by an ESP32 WROOM microcontroller. The microcontroller communicates with the laser sensors via SPI to acquire high-rate motion data (Δx, Δy), while also receiving samples of the grasp force from the FSRs. We designed a simple and easy-to-use system that allows the user to grasp objects of different shapes, masses, and surface textures (e.g., plastic bottles, thin pens, paper cups, smartphones) while we systematically vary the grip force and the imposed motion to identify whether the object is held stably, vibrating a little, or starting to slip.
When incipient slip is detected, the system activates a small vibration motor on the glove, providing immediate haptic feedback to the user, making it fast and discreet.
Experimental results demonstrate that our system enables reliable detection of incipient slip in a variety of scenarios, with short response times and good robustness to small shaking of the object held.