Constraint-induced movement therapy (CIMT) is an effective recovery protocol for patients with stroke. However, not all the stroke patients can participate in CIMT. Devices that assist the movement of the affected side could help increase the number of stroke patients who can participate in CIMT like protocol. In this letter, a soft cable-driven wrist-wearable robot called ExoWrist, which assists the wrist of a paretic arm in performing the dart-throwing motion (DTM) with an active forearm anchor, is proposed. To increase the force transmission efficiency while preventing medical issues due to long-term pressure, a corset active anchor, which compresses the body only when anchoring is needed, is developed. A moving pulley mechanism and a three-layered structure increase the functionality of the anchor. Considering the natural movement of wrist extension and the frequently-used wrist motions in daily life, the wrist is assisted along the DTM direction, and the method to find a personal DTM is introduced. A 3D-printed custom wrist armlet is developed to hold the tendon path and exert a torque along the DTM direction. Together with a hand assistive device, Exo-Wrist could be used as a tool to help stroke patients with difficulty in wrist extension to participate in task-oriented rehabilitation protocols such as CIMT.
Soft wearable hand robots with tendon-sheath mechanisms are being actively developed to assist people with lost hand mobility. For these robots, accurately estimating fingertip forces leads to successful object grasping. An approach can utilize information from actuators assuming quasi-static environments. However, non-linearity and hysteresis with regards to the dynamic changes of the tendon-sheath mechanism hinder accurate fingertip force estimation. This paper proposes a learning-based method to estimate fingertip forces by integrating dynamic information of motor encoders, wire tension, and sheath bending angles. The model is modified from Long Short-Term Memory by incorporating a residual term that governs the dynamic changes in sheath bending angles. Using a tendon-driven soft wearable hand robot, the proposed model obtained RMSE less than 0.44 N. It was further evaluated under criteria ranging from different object sizes, bending angle ranges, and forces. Finally, a repeatability test (0.46 N in RMSE), real-time applicability (125 Hz), and force control (12.7% in MAPE) were performed to verify the feasibility of the proposed method.