Algoritmo de segmentación para las fases de la prueba Timed Up and Go (TUG) considerando variables físicas de movimiento medidas a partir de sensores inerciales especializados.
Loading...
Date
2025
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
El presente estudio aborda la evaluación de la movilidad mediante el desarrollo y la validación de un algoritmo para la segmentación automática de las fases de la prueba Timed Up and Go (TUG). El objetivo general fue diseñar, implementar y validar un sistema basado en un único sensor de medición inercial (IMU) inalámbrico, ubicado en la región lumbar, capaz de cuantificar y analizar de manera precisa las sub-tareas de la prueba, superando las limitaciones del TUG tradicional basado en cronómetro.
La metodología se basó en un estudio con una cohorte de 60 adultos jóvenes sanos (edad media 23.03±2.57 años). Se empleó un sensor IMU en la región L3-L4 para registrar la aceleración lineal, la velocidad angular y la orientación (mediante cuaterniones) a 100 Hz. Las señales crudas fueron sometidas a un preprocesamiento que incluyó filtrado de paso bajo y conversión de cuaterniones a ángulos de Euler para su análisis. El algoritmo de segmentación propuesto se fundamenta en la detección de características cinemáticas clave en las señales de inclinación (Pitch) y rotación (Yaw) del tronco para identificar con alta precisión las transiciones entre las fases de sentado a de pie, caminata, giro y de pie a sentado. A partir de la segmentación, se extrajo un conjunto de parámetros biomecánicos para cada fase. El análisis estadístico de estos parámetros permitió establecer rangos normativos (media ±2 DE) para la cohorte, proporcionando una base cuantitativa para la caracterización de la "normalidad" del movimiento en adultos jóvenes.
La conclusión principal de este trabajo es que la metodología de TUG instrumentado (iTUG) aquí presentada constituye una herramienta válida, precisa y objetiva. Supera las limitaciones del TUG tradicional al ofrecer un análisis granular y detallado del rendimiento motor, lo que representa un avance significativo con un alto potencial para enriquecer la práctica clínica en la evaluación de la movilidad, el diseño de rehabilitaciones personalizadas y la prevención de caídas.
This study addresses mobility assessment through the development and validation of an algorithm for the automatic segmentation of the Timed Up and Go (TUG) test phases. The general objective was to design, implement, and validate a system based on a single wireless inertial measurement unit (IMU), placed on the lumbar region, capable of accurately quantifying and analyzing the test's sub-tasks, overcoming the limitations of the traditional stopwatch-based TUG. The methodology was based on a study with a cohort of 60 healthy young adults (mean age 23.03±2.57 years). An IMU sensor was used at the L3-L4 region to record linear acceleration, angular velocity, and orientation (via quaternions) at 100 Hz. The raw signals underwent a preprocessing stage that included low-pass filtering and conversion from quaternions to Euler angles for analysis. The proposed segmentation algorithm is based on detecting key kinematic features in the trunk's pitch and yaw signals to identify with high precision the transitions between the sit-to stand, walking, turning, and stand-to-sit phases. Following segmentation, a set of biomechanical parameters was extracted for each phase. The statistical analysis of these parameters allowed for the establishment of normative ranges (mean ±2 SD) for the cohort, providing a quantitative basis for characterizing "normal" movement in young adults. The main conclusion of this work is that the instrumented TUG (iTUG) methodology presented herein is a valid, accurate, and objective tool. It surpasses the limitations of the traditional TUG by offering a granular and detailed analysis of motor performance, representing a significant advancement with high potential to enrich clinical practice in mobility assessment, personalized rehabilitation design, and fall prevention.
This study addresses mobility assessment through the development and validation of an algorithm for the automatic segmentation of the Timed Up and Go (TUG) test phases. The general objective was to design, implement, and validate a system based on a single wireless inertial measurement unit (IMU), placed on the lumbar region, capable of accurately quantifying and analyzing the test's sub-tasks, overcoming the limitations of the traditional stopwatch-based TUG. The methodology was based on a study with a cohort of 60 healthy young adults (mean age 23.03±2.57 years). An IMU sensor was used at the L3-L4 region to record linear acceleration, angular velocity, and orientation (via quaternions) at 100 Hz. The raw signals underwent a preprocessing stage that included low-pass filtering and conversion from quaternions to Euler angles for analysis. The proposed segmentation algorithm is based on detecting key kinematic features in the trunk's pitch and yaw signals to identify with high precision the transitions between the sit-to stand, walking, turning, and stand-to-sit phases. Following segmentation, a set of biomechanical parameters was extracted for each phase. The statistical analysis of these parameters allowed for the establishment of normative ranges (mean ±2 SD) for the cohort, providing a quantitative basis for characterizing "normal" movement in young adults. The main conclusion of this work is that the instrumented TUG (iTUG) methodology presented herein is a valid, accurate, and objective tool. It surpasses the limitations of the traditional TUG by offering a granular and detailed analysis of motor performance, representing a significant advancement with high potential to enrich clinical practice in mobility assessment, personalized rehabilitation design, and fall prevention.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Biomédico.
Keywords
Monitoreo fisiológico, Ingeniería biomédica, Algoritmos Procesamiento de datos