Human Activity Recognition for Indoor Localization Using Smartphone Inertial Sensors

Moreira, Dinis; Barandas, Marília; Rocha, Tiago; Alves, Pedro; Santos, Ricardo; Leonardo, Ricardo; Vieira, Pedro; Gamboa, Hugo

doi:10.3390/s21186316

Cited by 13 publications

(6 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, it is not necessary to perform feature engineering before training the model, and as a result, the data preparation becomes more straightforward. Deep learning models can be classified into three types ( Moreira, 2021 ): deep generative models, deep discriminative models, and deep hybrid models. Deep generative models aim to learn useful representations of data via unsupervised learning or to learn the joint probability distribution of data and their associated classes.…”

Section: Introductionmentioning

confidence: 99%

Effects of sliding window variation in the performance of acceleration-based human activity recognition using deep learning models

Jaén-Vargas

Leiva

Fernandes

et al. 2022

PeerJ Computer Science

View full text Add to dashboard Cite

Deep learning (DL) models are very useful for human activity recognition (HAR); these methods present better accuracy for HAR when compared to traditional, among other advantages. DL learns from unlabeled data and extracts features from raw data, as for the case of time-series acceleration. Sliding windows is a feature extraction technique. When used for preprocessing time-series data, it provides an improvement in accuracy, latency, and cost of processing. The time and cost of preprocessing can be beneficial especially if the window size is small, but how small can this window be to keep good accuracy? The objective of this research was to analyze the performance of four DL models: a simple deep neural network (DNN); a convolutional neural network (CNN); a long short-term memory network (LSTM); and a hybrid model (CNN-LSTM), when variating the sliding window size using fixed overlapped windows to identify an optimal window size for HAR. We compare the effects in two acceleration sources’: wearable inertial measurement unit sensors (IMU) and motion caption systems (MOCAP). Moreover, short sliding windows of sizes 5, 10, 15, 20, and 25 frames to long ones of sizes 50, 75, 100, and 200 frames were compared. The models were fed using raw acceleration data acquired in experimental conditions for three activities: walking, sit-to-stand, and squatting. Results show that the most optimal window is from 20–25 frames (0.20–0.25s) for both sources, providing an accuracy of 99,07% and F1-score of 87,08% in the (CNN-LSTM) using the wearable sensors data, and accuracy of 98,8% and F1-score of 82,80% using MOCAP data; similar accurate results were obtained with the LSTM model. There is almost no difference in accuracy in larger frames (100, 200). However, smaller windows present a decrease in the F1-score. In regard to inference time, data with a sliding window of 20 frames can be preprocessed around 4x (LSTM) and 2x (CNN-LSTM) times faster than data using 100 frames.

show abstract

Section: Introductionmentioning

confidence: 99%

Effects of sliding window variation in the performance of acceleration-based human activity recognition using deep learning models

Jaén-Vargas

Leiva

Fernandes

et al. 2022

PeerJ Computer Science

View full text Add to dashboard Cite

show abstract

“…Deep learning approaches have been explored in OOD settings by testing the models on data from unseen domains [ 4 , 29 , 30 , 31 , 32 ]. Gholamiangonabadi et al.…”

Section: Related Workmentioning

confidence: 99%

Comparing Handcrafted Features and Deep Neural Representations for Domain Generalization in Human Activity Recognition

Bento

Rebelo

Barandas

et al. 2022

Sensors

Self Cite

View full text Add to dashboard Cite

Human Activity Recognition (HAR) has been studied extensively, yet current approaches are not capable of generalizing across different domains (i.e., subjects, devices, or datasets) with acceptable performance. This lack of generalization hinders the applicability of these models in real-world environments. As deep neural networks are becoming increasingly popular in recent work, there is a need for an explicit comparison between handcrafted and deep representations in Out-of-Distribution (OOD) settings. This paper compares both approaches in multiple domains using homogenized public datasets. First, we compare several metrics to validate three different OOD settings. In our main experiments, we then verify that even though deep learning initially outperforms models with handcrafted features, the situation is reversed as the distance from the training distribution increases. These findings support the hypothesis that handcrafted features may generalize better across specific domains.

show abstract

“…Moreira et al [ 79 ] proposed an interesting application based on sensory data acquired through a smartphone and a convolutional long short-term memory (ConvLSTM) for the classification of human activities within an indoor environment network. They considered nine activities for classification.…”

Section: Systems For Harmentioning

confidence: 99%

An Overview of Indoor Localization System for Human Activity Recognition (HAR) in Healthcare

Bibbò

Carotenuto

Corte

2022

Sensors

View full text Add to dashboard Cite

The number of older people needing healthcare is a growing global phenomenon. The assistance in long-term care comprises a complex of medical, nursing, rehabilitation, and social assistance services. The cost is substantial, but technology can help reduce spending by ensuring efficient health services and improving the quality of life. Advances in artificial intelligence, wireless communication systems, and nanotechnology allow the creation of intelligent home care systems avoiding hospitalization with evident cost containment. They are capable of ensuring functions of recognition of activities, monitoring of vital functions, and tracking. However, it is essential to also have information on location in order to be able to promptly intervene in case of unforeseen events or assist people in carrying out activities in order to avoid incorrect behavior. In addition, the automatic detection of physical activities performed by human subjects is identified as human activity recognition (HAR). This work presents an overview of the positioning system as part of an integrated HAR system. Lastly, this study contains each technology’s concepts, features, accuracy, advantages, and limitations. With this work, we want to highlight the relationship between HAR and the indoor positioning system (IPS), which is poorly documented in the literature.

show abstract

Human Activity Recognition for Indoor Localization Using Smartphone Inertial Sensors

Cited by 13 publications

References 25 publications

Effects of sliding window variation in the performance of acceleration-based human activity recognition using deep learning models

Effects of sliding window variation in the performance of acceleration-based human activity recognition using deep learning models

Comparing Handcrafted Features and Deep Neural Representations for Domain Generalization in Human Activity Recognition

An Overview of Indoor Localization System for Human Activity Recognition (HAR) in Healthcare

Contact Info

Product

Resources

About