Researchers retrofitted a commercial TASKA prosthetic hand with multimodal fingertip proximity and pressure sensors and implemented a novel shared-control algorithm that continuously blends user myoelectric intent (via a modified Kalman filter) with an autonomous machine goal (MLP-based distance-to-contact), enabling the machine to conform each digit to objects while the user controls overall grasp; tested in nine intact-limb participants and four transradial amputees, shared control increased fragile-object transfer success from ~59% to ~89%, reduced average grip force (e.g., amputees 14.9 N to 9.7 N), markedly extended hold times (amputee median hold time 7.8 s to 51.6 s), reduced drops in power-grasp tests, and lowered cognitive load as measured by faster DRT response times (up to a 120 ms, ~24% improvement) without increasing muscular effort. The system generalizes across multiple grips and real-world tasks, is designed to be retrofittable and lower-power than camera-based solutions, and—if validated in longer at-home trials—could materially improve functional outcomes and commercial attractiveness of next-generation myoelectric prostheses, with implications for adoption and competitive differentiation in the prosthetics market.
Researchers retrofitted a commercial TASKA prosthetic hand with multimodal fingertip sensors (VCNL4010 infrared proximity, MS5637 barometric pressure) and implemented a shared-control architecture blending an MLP-based machine goal with a modified Kalman-filter myoelectric decode. In nine intact-limb participants and four transradial amputees, shared control increased fragile-object transfer success from 59 ± 23% to 89 ± 10% (p < 0.01), reduced average amputee grip force from 14.86 ± 4.54 N to 9.74 ± 4.27 N (p < 0.001), extended amputee hold time on a fragile object from 7.81 ± 12.33 s to 51.56 ± 45.0 s (p < 0.01), and shortened detection-response task (DRT) times by ~60 ms for intact participants and ~120 ms for amputees (24–29% cognitive-load reductions, p ≤ 0.05). Machine-only autonomous control also materially reduced contact force and failure rates on a fragile transfer task (human 23 ± 21% success vs. machine 99 ± 3%; p < 0.001), illustrating the value of proximity feedforward plus pressure feedback for near-zero contact. The technology is designed to be retrofittable, lower-power than camera-based approaches, and generalizes across multiple grips and ADLs in lab testing; shared control preserved user agency while reducing cognitive burden without a consistent increase in physical exertion. Validation with four amputees and diverse tasks strengthens translational prospects, but the study is short-term and focused on controlled tasks rather than long-duration, at-home use. Key risks that would affect commercialization and investment timing include small sample size and short follow-up, the need for longer at-home clinical trials and regulatory/clinical validation (AM-ULA or similar), potential failure modes such as sensor occlusion or machine lag/precede that can transiently increase user effort, and the dependence on OEM adoption of embedded multimodal sensors and low-power integration.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request a DemoOverall Sentiment
moderately positive
Sentiment Score
0.45
Ticker Sentiment
