Subject-Independent Continuous Locomotion Mode Classification for Robotic Hip Exoskeleton Applications

Estimating human joint moments unifies exoskeleton control, reducing user effort

“…Confirming hypotheses H3 and H4, the TCN significantly outperformed the Baseline method in RMSE and R 2 , even though the Baseline method in this study assumed a perfectly accurate ambulation mode classifier and gait phase estimator. In practice, mode classifiers and gait phase estimators also incur error (13)(14)(15)(16)(17)(26)(27)(28), further increasing the differences between the TCN and Baseline method. Nevertheless, this indicates that the TCN not only captured changes in hip moments as ambulation mode and gait phase varied but also modeled changes in hip moments across participants, across intensities, and/or across strides.…”

“…Generally, exoskeleton controllers are divided into three layers: high-level, mid-level, and low-level (12). The high-level layer estimates user and environmental states, such as ambulation mode (9,(13)(14)(15)(16)(17) or ground slope (18,19), used to modulate assistance with changes in user joint demands. The state estimates are passed to the mid-level layer, which computes desired assistance on the basis of predefined control laws, such as spline-based assistance trajectories (8,20,21).…”

Exoskeletal solutions to enable mobility with a lower leg fracture in austere environments

“…This requires a control system capable of shaping the exoskeleton assistance appropriately across a diverse set of activities. This has been a concentrated research area over the last decade which include direct estimation of environmental state through techniques including heuristic state machine logic (Stolyarov et al, 2021 ) to machine learning activity patterns (Laschowski et al, 2021 ; Kang et al, 2022 ; Wang et al, 2022 ). Alternatively, control strategies have also been formulated to provide task-invariant capability such as through myoelectric control (Nasr et al, 2021 ) or energy-shaping techniques (Lin et al, 2021 ).…”

Opportunities and challenges in the development of exoskeletons for locomotor assistance

“…Therefore, exoskeletons often rely on sensor input to measure intent of movement and to predict the wearer's motion before it occurs, to appropriately adjust assistance. However, detecting intent of movement is distinct from activity classification, in which heuristic methods or machine-learning techniques are used to identify activities beyond level-ground walking (such as walk-to-run transitions; or walking on terrains such as ramps or stairs 204,205 ). Similar to detecting the gait phase, the intent of stepping can be detected using IMUs that measure joint angles and angular velocities 206 and that estimate the centre-of-mass position 64 .…”