Human Activity Monitoring Based On Hidden Markov Models Using a Smartphone

Rubén San Segundo; Julian  David Echeverry Correa; Christian Raul Salamea Palacios; José M. Pardoa

doi:10.1109/MIM.2016.7777649

Human Activity Monitoring Based On Hidden Markov Models Using a Smartphone

Título traducido de la contribución: Monitoreo de actividad humana basado en modelos ocultos de Markov utilizando un teléfono inteligente

Rubén San Segundo, Julian David Echeverry Correa, Christian Raul Salamea Palacios, José M. Pardoa

Grupo de Investigación en Interacción, Robótica y Automática (GIIRA)

Producción científica: Contribución a una revista › Artículo

Resumen

Este trabajo presenta un sistema de detección humana (HS) basado en Modelos de Markov Ocultos (HMMs) para clasificar actividades físicas: caminar, subir escaleras, bajar escaleras, sentarse, estar de pie y acostarse. El sistema incluye un extractor de características (desarrollado por los autores y presentado en un trabajo anterior), un módulo de entrenamiento de HMMs y un módulo HAR. Todos los experimentos se han realizado utilizando un conjunto de datos disponible públicamente denominado UCI Human Activity Recognition Using Smartphones. Los resultados finales utilizando HMMs obtienen resultados comparables a otros métodos de reconocimiento. Se han obtenido algunas mejoras al considerar un procedimiento de entrenamiento de HMMs discriminativo. El mejor resultado obtiene una tasa de error de reconocimiento de actividad (ARER) del 2,5%. Este trabajo se centra en el reconocimiento de actividad independiente y amplía otros trabajos de los mismos autores centrados en la segmentación de actividad y la extracción de características.

Título traducido de la contribución	Monitoreo de actividad humana basado en modelos ocultos de Markov utilizando un teléfono inteligente
Idioma original	Inglés estadounidense
Páginas (desde-hasta)	1-5
Publicación	Instrumentation & Measurement
Volumen	19
N.º	19
DOI	https://doi.org/10.1109/MIM.2016.7777649
Estado	Publicada - 12 dic. 2016

Palabras clave

Hidden markov models
Training
Feature extraction
Error analysis
Computational modeling
Human factors

Areas de Conocimiento del CACES

116A Computación

Acceder al documento

10.1109/MIM.2016.7777649

Otros archivos y enlaces

Other Link

Citar esto

@article{a877cb3d1f2546c9b019989b3e7b2c80,

title = "Human Activity Monitoring Based On Hidden Markov Models Using a Smartphone",

abstract = "This paper presents an human sensing (HS) system based on Hidden Markov Models (HMMs) for classifying physical activities: walking, walking-upstairs, walking-downstairs, sitting, standing and lying down. The system includes a feature extractor (developed by the authors and presented in a previous work), an HMMs training module and an HAR module. All experiments have been done using a publicly available dataset named UCI Human Activity Recognition Using Smartphones. The final results using HMMs obtain comparable results to other recognition methods. Some improvements have been obtained when considering a discriminative HMM training procedure. The best result obtains an activity recognition error rate (ARER) of 2.5%. This work is focused on independent activity recognition and extends other works from the same authors focused on activity segmentation and feature extraction.",

keywords = "Hidden markov models, Training, Feature extraction, Error analysis, Computational modeling, Human factors, Hidden markov models, Training, Feature extraction, Error analysis, Computational modeling, Human factors",

author = "{San Segundo}, Rub{\'e}n and {Echeverry Correa}, {Julian David} and {Salamea Palacios}, {Christian Raul} and Pardoa, {Jos{\'e} M.}",

year = "2016",

month = dec,

day = "12",

doi = "10.1109/MIM.2016.7777649",

language = "English (US)",

volume = "19",

pages = "1--5",

journal = "Instrumentation & Measurement",

issn = "1094-6969",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "19",

}

TY - JOUR

T1 - Human Activity Monitoring Based On Hidden Markov Models Using a Smartphone

AU - San Segundo, Rubén

AU - Echeverry Correa, Julian David

AU - Salamea Palacios, Christian Raul

AU - Pardoa, José M.

PY - 2016/12/12

Y1 - 2016/12/12

N2 - This paper presents an human sensing (HS) system based on Hidden Markov Models (HMMs) for classifying physical activities: walking, walking-upstairs, walking-downstairs, sitting, standing and lying down. The system includes a feature extractor (developed by the authors and presented in a previous work), an HMMs training module and an HAR module. All experiments have been done using a publicly available dataset named UCI Human Activity Recognition Using Smartphones. The final results using HMMs obtain comparable results to other recognition methods. Some improvements have been obtained when considering a discriminative HMM training procedure. The best result obtains an activity recognition error rate (ARER) of 2.5%. This work is focused on independent activity recognition and extends other works from the same authors focused on activity segmentation and feature extraction.

AB - This paper presents an human sensing (HS) system based on Hidden Markov Models (HMMs) for classifying physical activities: walking, walking-upstairs, walking-downstairs, sitting, standing and lying down. The system includes a feature extractor (developed by the authors and presented in a previous work), an HMMs training module and an HAR module. All experiments have been done using a publicly available dataset named UCI Human Activity Recognition Using Smartphones. The final results using HMMs obtain comparable results to other recognition methods. Some improvements have been obtained when considering a discriminative HMM training procedure. The best result obtains an activity recognition error rate (ARER) of 2.5%. This work is focused on independent activity recognition and extends other works from the same authors focused on activity segmentation and feature extraction.

KW - Hidden markov models

KW - Training

KW - Feature extraction

KW - Error analysis

KW - Computational modeling

KW - Human factors

KW - Hidden markov models

KW - Training

KW - Feature extraction

KW - Error analysis

KW - Computational modeling

KW - Human factors

UR - https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7777649

U2 - 10.1109/MIM.2016.7777649

DO - 10.1109/MIM.2016.7777649

M3 - Article

SN - 1094-6969

VL - 19

SP - 1

EP - 5

JO - Instrumentation & Measurement

JF - Instrumentation & Measurement

IS - 19

ER -

Human Activity Monitoring Based On Hidden Markov Models Using a Smartphone

Resumen

Palabras clave

Areas de Conocimiento del CACES

Acceder al documento

Otros archivos y enlaces

Huella

Citar esto