Abstract
Cerebral Palsy refers to a group of incurable motor disorders affecting 0.22% of the global population. Symptoms are managed by physiotherapists, often using rehabilitation robotics. Exoskeletons, offering advantages over conventional therapies, are evolving to be more wearable and biomimetic, requiring new flexible actuators that mimic human tissue. The main objective behind this article is the design of a flexible exosuit based on shape-memory-alloy-based artificial muscles for pediatric patients that replicate the walking cycle pattern in the ankle joint. Thus, four shape-memory-alloy-based actuators were sewn to an exosuit at the desired actuation points and controlled by a two-level controller. The loop is closed through six inertial sensors that estimate the real angular position of both ankles. Different frequencies of actuation have been tested, along with the response of the actuators to different walking cycle patterns. These tests have been performed over long periods of time, comparing the reference created by a reference generator based on pediatric walking patterns and the response measured by the inertial sensors. The results provide important measurements concerning errors, working frequencies and cooling times, proving that this technology could be used in this and similar applications and highlighting its limitations.