Real-Time Assistive Control via IMU Locomotion Mode Detection in a Soft Exosuit: An Effective Approach to Enhance Walking Metabolic Efficiency

In order to substantially improve human's walking endurance and energy economy, wearable assistive devices need to accurately recognize and timely adapt to different locomotion modes, such as ascending/descending stairs or level ground walking. In this work, we developed a control strategy for a soft hip exosuit entirely based on inertial measurement units (IMUs), able to online distinguish among three different walking patterns and optimally assist the user's gait phase. A time-delay compensation strategy was incorporated in the controller to promote high human-device synchronicity. The effectiveness of this control strategy was tested on healthy participants during overground walking consisting of a combination of staircases and level grounds. We found that the overall accuracy of the IMUs classification strategy based on human kinematics exceeded 90% for the three locomotion modes. Preliminary results showed that our assistive exosuit reduced the wearer's metabolic rate by 13.4% during walking when compared with an unpowered condition, and by 8.5% with respect to not wearing the exosuit at all. This work contributes to the development of compact high-performance lower-limb assistive technologies and their exploitation in real-world applications.