A Hierarchical FSM–BT Control Architecture for a Passive/Active Rehabilitation Exoskeleton

Rehabilitation exoskeletons require supervisory control that can preserve globally safe operating modes while executing local therapeutic behaviors reactively. Pure Finite-State Machine (FSM) implementations provide explicit mode supervision but become increasingly coupled as behaviors and recovery transitions are added, whereas Behavior Trees (BTs) provide modular reactive execution but do not inherently encode a unique global operating mode. This work presents an application-specific hierarchical FSM–BT architecture for the SARA passive–active lower-limb rehabilitation platform. The novelty is not the generic combination of FSMs and BTs; rather, it lies in the rehabilitation-oriented separation of global therapeutic-mode supervision and event-priority safety arbitration from parameterizable BT execution and local recovery before escalation. The architecture was evaluated exclusively in a custom Python Software-in-the-Loop (SIL) environment using an eight-degree-of-freedom second-order joint model, a 100 Hz plant integration rate, a 10 Hz supervisory FSM, and 50 Hz BT execution. Thirty Monte Carlo trials were performed for each of four scenarios and for three functionally equivalent architectures (FSM-only, BT-only, and FSM–BT), for 360 comparative trials with a fixed random seed. The proposed FSM–BT architecture achieved a 98.3% overall trial-success rate, compared with 89.2% for FSM-only and 92.5% for BT-only. Its mean software-level unsafe-condition-to- Emergency Stop response was 71.9 ms, significantly lower than the 159.7 ms FSM-only response (p<10−8, paired Wilcoxon test), while BT-only produced the same safety timing because it used the same 50 Hz local evaluation cadence. The active-assistance model yielded a mean eight-DOF angular RMSE of 2.42∘. A structural extension study further required 54 edit operations to add six therapeutic routines to FSM–BT, versus 60 for BT-only, 96 for a hierarchical state-machine baseline, and 138 for FSM-only.

Authors

Institutions

Publication Details

Journal
Sensors
Published
2026-09-09
DOI
https://doi.org/10.3390/s26185731
Primary Topic
Prosthetics and Rehabilitation Robotics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

A Hierarchical FSM–BT Control Architecture for a Passive/Active Rehabilitation Exoskeleton

Raquel E. Patiño-Escarcina, Dennis Barrios-Aranibar, Regina B. M. Chávez-Reynoso
Sensors
Prosthetics and Rehabilitation Robotics
article

A Hierarchical FSM–BT Control Architecture for a Passive/Active Rehabilitation Exoskeleton

Raquel E. Patiño-Escarcina, Dennis Barrios-Aranibar, Regina B. M. Chávez-Reynoso
article en

Abstract

Rehabilitation exoskeletons require supervisory control that can preserve globally safe operating modes while executing local therapeutic behaviors reactively. Pure Finite-State Machine (FSM) implementations provide explicit mode supervision but become increasingly coupled as behaviors and recovery transitions are added, whereas Behavior Trees (BTs) provide modular reactive execution but do not inherently encode a unique global operating mode. This work presents an application-specific hierarchical FSM–BT architecture for the SARA passive–active lower-limb rehabilitation platform. The novelty is not the generic combination of FSMs and BTs; rather, it lies in the rehabilitation-oriented separation of global therapeutic-mode supervision and event-priority safety arbitration from parameterizable BT execution and local recovery before escalation. The architecture was evaluated exclusively in a custom Python Software-in-the-Loop (SIL) environment using an eight-degree-of-freedom second-order joint model, a 100 Hz plant integration rate, a 10 Hz supervisory FSM, and 50 Hz BT execution. Thirty Monte Carlo trials were performed for each of four scenarios and for three functionally equivalent architectures (FSM-only, BT-only, and FSM–BT), for 360 comparative trials with a fixed random seed. The proposed FSM–BT architecture achieved a 98.3% overall trial-success rate, compared with 89.2% for FSM-only and 92.5% for BT-only. Its mean software-level unsafe-condition-to- Emergency Stop response was 71.9 ms, significantly lower than the 159.7 ms FSM-only response (p<10−8, paired Wilcoxon test), while BT-only produced the same safety timing because it used the same 50 Hz local evaluation cadence. The active-assistance model yielded a mean eight-DOF angular RMSE of 2.42∘. A structural extension study further required 54 edit operations to add six therapeutic routines to FSM–BT, versus 60 for BT-only, 96 for a hierarchical state-machine baseline, and 138 for FSM-only.

SensorsVol. 26(18)
Universidad Católica San Pablo (PE)
Openalex Percentile: Top 20%
Prosthetics and Rehabilitation Robotics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.