Next‐generation control strategies for lower‐extremity exoskeletons: State‐of‐the‐art, limitations, and future directions

Abstract

Wearable powered exoskeletons are emerging rehabilitation tools for restoring gait in individuals with neurologic or motor impairments, including stroke, spinal cord injury, myopathy, and cerebral palsy. Lower‐extremity exoskeleton‐assisted walking can improve motor function, gait parameters, balance, and independence. Exoskeleton use has also been associated with secondary health benefits, including reduced spasticity and improvements in autonomic functions such as bladder and bowel control. However, despite these clinical benefits, exoskeleton control systems remain a central bottleneck limiting functional gains and widespread clinical adoption. The majority of clinically approved lower‐extremity devices rely on weight shift triggers, predefined gait phase controllers, impedance‐based assistance, or simple electromyographic signal activation. These approaches often feel unintuitive and require substantial user training. As a result, persistent challenges remain in detecting user intent, reproducing a user's expected dynamic behaviors, adapting to variable gait patterns and changing environmental conditions such as uneven terrain or inclines, and minimizing cognitive burden during everyday use. This review focuses on control strategies for lower‐extremity exoskeletons used in gait rehabilitation and restoration of motor function. It summarizes current approaches in commercial and research systems and outlines why existing methods remain insufficient for seamless, user‐driven assistance. We also examine emerging approaches for determining user intent, including brain–computer interfaces, peripheral nerve interfaces, novel biosignals such as ultrasound and mechanomyography, and regenerative surgical approaches. Together, these advances may help bridge the gap between human intent and robotic assistance, enabling more natural and autonomous mobility.