Compensating elastic faults in a torque-assisted knee exoskeleton: functional evaluation and user perception study.

The study involved only 10 healthy young adults, so the results may not apply to patients with neurological or musculoskeletal conditions who might have different sensory thresholds. The walking speed was fixed at a slow 0.58 m/s to avoid motor limitations, meaning the device's performance at faster, more functional speeds is unknown. The 'faults' were simulated by adjusting spring pretension, which may not perfectly replicate the complex mechanical failures that occur in real-world use. Residual torque during the swing phase of walking influenced stiffness perception, potentially confounding the results for high-severity faults. Comfort and trust were not affected, possibly because the short duration of exposure and the low impact of the faults on actual walking performance meant these higher-level constructs were not challenged.

Funded by the Deutsche Forschungsgemeinschaft and Friedrich-Alexander-Universität Erlangen-Nürnberg. No commercial conflicts of interest are reported.

Do not assume that fault-tolerant control algorithms will guarantee patient safety or satisfaction in clinical populations. The study showed that while healthy users could not feel compensated faults, the device's performance at faster speeds or with patients who have altered sensation was not tested. Also, the fact that trust did not change does not mean faults are harmless; it may simply mean the faults were too subtle to impact the user's overall confidence in this specific short-term scenario.