Lower-limb exoskeletons represent an emerging technology in clinical and rehabilitative settings, aimed at supporting gait in individuals with neuromotor impairments and reducing biomechanical load. A crucial element for the effectiveness of such devices is the accurate detection of foot–ground contact, which is necessary to synchronize assistance with the different phases of the gait cycle. Conventional methods for identifying contact events frequently rely on sensors dedicated to the foot or the ground, such as force platforms or pressure-sensing systems, or on inertial units specifically employed for gait event inference. Although these solutions provide high accuracy, they introduce increased hardware complexity and dependence on additional components. This thesis proposes an approach for the indirect determination of foot–ground contact based exclusively on the embedded sensor already integrated within the actuators of a unilateral lowerlimb exoskeleton. Through the analysis of joint torque measured by the motors and kinematic parameters provided by encoders and an Inertial Measurement Unit, an algorithm was developed to estimate the vertical component of the ground reaction force and to discriminate in real time between stance and swing phases. The method was experimentally evaluated through synchronized acquisitions with force platforms, assumed as ground truth, on a sample of experimental data. Preliminary results show good consistency between the estimates obtained from the embedded sensors and the direct measurements of ground reaction forces, suggesting the possibility of identifying Heel Strike and Toe Off events without the use of additional sensors. The work is presented as a preliminary descriptive analysis aimed at exploring the feasibility of the proposed approach. The obtained evidence indicates the potential of relying exclusively on internal actuator sensors to reduce system complexity and bulk, opening perspectives for further validation studies on larger samples and in clinical contexts.
Gli esoscheletri per arto inferiore rappresentano una tecnologia emergente con applicazioni cliniche e riabilitative, finalizzate a supportare il cammino in soggetti con deficit neuromotori e a ridurre il carico biomeccanico in ambito industriale. Un aspetto cruciale per l’efficacia di questi dispositivi è il rilevamento accurato del contatto piede–suolo, necessario per sincronizzare l’assistenza con le diverse fasi del passo. I metodi convenzionali, basati su sensori esterni quali pedane di forza, solette pressometriche o unità inerziali, garantiscono elevata precisione ma presentano limiti pratici legati a costo, complessità e scarsa portabilità. La presente tesi propone un approccio innovativo per la determinazione indiretta del contatto piede–suolo sfruttando esclusivamente i sensori già integrati nei motori di un esoscheletro unilaterale (Robopedics AWAKE). Attraverso l’analisi della coppia articolare misurata dai motori e dei parametri cinematici forniti da encoder e Inertial Measurement Unit (IMU), è stato sviluppato un algoritmo capace di stimare la componente verticale della Ground Reaction Force (GRF) e di discriminare in tempo reale le fasi di stance e swing. Il metodo è stato validato sperimentalmente mediante acquisizioni sincronizzate con piattaforme dinamometriche IBV, assunte come ground truth. I risultati mostrano una buona corrispondenza tra le stime derivate dai sensori embedded e le misure dirette delle forze di reazione, confermando la possibilità di identificare in modo robusto gli eventi di Heel Strike (Initial Contact, IC) e Toe Off (TO) senza ricorrere a sensori aggiuntivi. Questi risultati evidenziano come l’impiego dei soli sensori integrati nei motori possa ridurre complessità, costi e ingombro degli esoscheletri, mantenendo prestazioni comparabili ai sistemi tradizionali. Tra gli sviluppi futuri si prevede l’ottimizzazione dell’algoritmo per diverse velocità e tipologie di cammino e la sua estensione a contesti clinici, in particolare per applicazioni nella riabilitazione post-ictus.
Determinazione indiretta dell’appoggio del piede al suolo tramite sensori nei motori di un esoscheletro
MOZZICATO, ANTONINO
2025/2026
Abstract
Lower-limb exoskeletons represent an emerging technology in clinical and rehabilitative settings, aimed at supporting gait in individuals with neuromotor impairments and reducing biomechanical load. A crucial element for the effectiveness of such devices is the accurate detection of foot–ground contact, which is necessary to synchronize assistance with the different phases of the gait cycle. Conventional methods for identifying contact events frequently rely on sensors dedicated to the foot or the ground, such as force platforms or pressure-sensing systems, or on inertial units specifically employed for gait event inference. Although these solutions provide high accuracy, they introduce increased hardware complexity and dependence on additional components. This thesis proposes an approach for the indirect determination of foot–ground contact based exclusively on the embedded sensor already integrated within the actuators of a unilateral lowerlimb exoskeleton. Through the analysis of joint torque measured by the motors and kinematic parameters provided by encoders and an Inertial Measurement Unit, an algorithm was developed to estimate the vertical component of the ground reaction force and to discriminate in real time between stance and swing phases. The method was experimentally evaluated through synchronized acquisitions with force platforms, assumed as ground truth, on a sample of experimental data. Preliminary results show good consistency between the estimates obtained from the embedded sensors and the direct measurements of ground reaction forces, suggesting the possibility of identifying Heel Strike and Toe Off events without the use of additional sensors. The work is presented as a preliminary descriptive analysis aimed at exploring the feasibility of the proposed approach. The obtained evidence indicates the potential of relying exclusively on internal actuator sensors to reduce system complexity and bulk, opening perspectives for further validation studies on larger samples and in clinical contexts.| File | Dimensione | Formato | |
|---|---|---|---|
|
Mozzicato_Antonino.pdf
embargo fino al 31/03/2029
Dimensione
1.44 MB
Formato
Adobe PDF
|
1.44 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/106029