1) Background
Myoelectric prosthetic hands decode muscle activity from the user’s residual limb to control the hand movement. Yet despite notable technological advances, abandonment remains high—reported up to 44% in some countries—because current devices often fail to meet user needs and expectations [1]. A major gap in commercially available prosthetic hands, is the sense of touch: users cannot reliably sense contact, force, slip, or texture. This lack of tactile information leads to “blind” grasping, force overshoot, object slip, limited fine manipulation, and constrained social touch (e.g., hugging or petting). The result is a higher cognitive load for the user and reduces their confidence, independence, and connection to others.
2) Novelty & Importance
This project equips myoelectric hands with soft, low-profile tactile skin that conforms to the palm and fingers. The aim is for the bio-inspired skins to capture distributed contact, incipient slip, and surface texture during everyday use. The skins will encode each of these events in real time, as intuitive vibrotactile/skin-stretch feedback to the user. This can be on the residual limb, or potentially another location on the body. In parallel, a personalised, data-driven controller fusing EMG and tactile signals to stabilise grasping will be formulated. A key challenge will be to ensure the control approach can adapt to the user’s preferences with minimal calibration. Emphasis on efficient, on-device processing (including spiking/neuromorphic methods where beneficial) supports all-day wear. Grounded in user-centred co-design with prosthesis users and clinicians, this work targets a practical pathway from benchtop prototypes to deployable devices that reduce grip errors and restore meaningful social touch.
3) Aims & Objectives
Aim: Develop a myoelectric prosthetic hand with tactile sensing skin that delivers intuitive, useful haptic feedback for social touch and dexterous manipulation.
Objectives:
O1: Design and integrate soft tactile skins to estimate normal force (0–50 N), detect incipient slip (<150 ms), and classify coarse texture. O2: Fuse EMG with tactile signals for adaptive grip control and map tactile events to residual-limb haptics with <100 ms end-to-end latency. O3: Demonstrate higher task success, reduced force overshoot/slip, and lower subjective effort in controlled lab studies and real-world trials with end users.