Enhancing Gait Assistance Control Robustness of a Hip Exosuit by means of Machine Learning

Abstract

Optimally synchronising the assistance provided by wearable devices with the human voluntary motion is still an open challenge in robotics. In order to provide accurate and robust assistance, this paper presents a novel approach that combines a layered implementation of a controller for an underactuated exosuit assisting hip flexion during human locomotion: the first layer is based on Adaptive Oscillators (AOs layer), while the second one uses Machine Learning (ML layer). The latter has been introduced to enhance the robustness of the AOs-based controller in abrupt changes of the gait frequency, with the final goal to achieve higher synchronisation and symbiosis between the user and assistive devices in presence of variable and unpredictable locomotion patterns. The effectiveness of the layered controller has been tested on six healthy subjects. Preliminary results suggested that the additional ML layer provided improvement to the overall performances during overground walking. In addition, we found a reduction of metabolic rates when receiving assistance from the device: 7.4% on average on treadmill evaluations and 10% overground including the extra ML layer, without alteration of the physiological human motion.