A hip–knee–ankle exoskeleton emulator for studying gait assistance

Gwendolyn M. Bryan, Patrick W. Franks, Stefan C. Klein, Robert J. Peuchen, Steven H. Collins

Research output: Contribution to journalArticleScientificpeer-review

1 Citation (Scopus)


Lower-limb exoskeletons could improve the mobility of people with disabilities, older adults, workers, first responders, and military personnel. Despite recent advances, few products are commercially available and exoskeleton research is still often limited by hardware constraints. Many promising multi-joint assistance strategies, especially those with high-torque and high-power components, have yet to be tested because they are beyond the capabilities of current devices. To study these untested assistance strategies, we present a hip–knee–ankle exoskeleton emulator that can apply high torques and powers that match or exceed those observed in uphill running. The system has powerful off-board motors that actuate a 13.5 kg exoskeleton end effector worn by the user. It can apply up to 200 Nm of torque in hip flexion, hip extension, and ankle plantarflexion, 250 Nm of torque in knee extension, and 140 Nm of torque in knee flexion, with over 4.5 kW of power at each joint and a closed-loop torque bandwidth of at least 18 Hz in each direction of actuation. The exoskeleton is compliant in unactuated directions, adjustable for a wide range of users and comfortable during walking and running. When paired with human-in-the-loop optimization, we expect that this system will identify new assistance strategies to improve human mobility. A complete computer-aided design (CAD) model of the exoskeleton and a bill of materials are included and available for download.

Original languageEnglish
Number of pages25
JournalInternational Journal of Robotics Research
Publication statusPublished - 2020
Externally publishedYes


  • Exoskeleton design
  • human augmentation
  • locomotion
  • wearable robotics


Dive into the research topics of 'A hip–knee–ankle exoskeleton emulator for studying gait assistance'. Together they form a unique fingerprint.

Cite this