To comply with the large displacements and forces required by the work-out range of motion, a suitable actuator was designed as a cartridge of approximately 150mmx20mmx15mm, containing 2.5m of 0.25mm-diameter NiTi wire. During tests, this actuator was activated by Joule’s effect employing a 7s current input at 0.7A. This provided 10N through 76mm displacement. Cooling and reset by natural convection took 30s.

 A prototype of SHADE was assembled with two thermoplastic shells hinged together at the ankle and strapped on the frontal aspect of the shin and on the foot. Two actuators were fixed on the upper shell while an inextensible thread connected each NiTi wire to the foot shell. A dedicated software switched on and off a 30Vdc-generator providing sufficient power to get shape recovery in 7s. Martensite detwinning, caused by foot weight, occurred in 30s.

 The passive ankle motion (PROM) generated by SHADE was evaluated optoelectronically on three flaccid patients (58±5y/o); acceptability was assessed by a questionnaire presented to further three flaccid patients (44±11.5y/o) who used SHADE for 1 week, 30min a day. 

 SHADE was well accepted by all patients, produced good PROM and caused no pain. Actuator displacement self-adapted to account for patients’ changing ankle rigidity. The rate of dorsiflexion was comparable with self-imposed movement in patient with residual active control, while the slow plantarflexion speed (driven by wire cooling) ensured muscle stretching for around 70% of the cycle.

 The results prove that suitable limb mobilisation can be produced by SMA actuators.