Robust Friction Estimation for an Active Upper-Limb Exoskeleton Via SOSML Observer
Hamidreza Heidari, Paolino De Risi, Fanny Ficuciello
AI summary
Problem
Unmodeled joint friction in compact, geared exoskeleton actuators degrades motion transparency and control precision, especially near zero velocity and during reversals, while existing estimation methods struggle with noise sensitivity, hardware complexity, or low-speed accuracy.
Approach
The authors estimate joint friction as a lumped disturbance using only standard proprioceptive sensors by comparing a first-order momentum observer against a second-order sliding-mode momentum observer (SOSML), validated through simulation and hardware experiments on a two-degree-of-freedom upper-limb exoskeleton.
Key results
- SOSML adheres more closely to Stribeck friction laws near zero velocity than first-order observers
- SOSML maintains tighter error bounds and lower phase lag under encoder noise and model mismatch
- SOSML recovers friction parameters with reduced drift and higher symmetry across direction reversals
- Hardware tests confirm superior low-speed estimation accuracy without modifying the sensor suite or excitation signals
Why it matters
This sensor-less friction estimation approach enables more transparent and precise control of wearable exoskeletons, directly benefiting rehabilitation robotics and human-machine interaction design.
Abstract
Friction in compact, geared actuators remains a primary barrier to transparency in upper-limb exoskeletons, especially near zero velocity and during frequent reversals. A momentum-based estimation framework is developed and evaluated on a two-DoF active device (modified EDUExo), where joint friction is recovered from on-board joint measurements and fitted to Coulomb–viscous and Stribeck laws. Two estimators are compared under identical conditions: a first-order momentum observer (FO) and a second-order sliding-mode momentum observer (SOSML). Three velocity trajectories are designed to probe complementary behaviors. In simulation, SOSML adheres more closely to the S-shaped friction law, and preserves loop symmetry under encoder noise; parameter variance and robustness under structured model mismatch are likewise improved relative to FO. The results indicate that SOSML delivers lower lag, cleaner noise profiles, and reduced parameter drift without changing the signal set or adding sensors, thereby strengthening friction identification and compensation on compact, gear-reduced actuators.