Efficient Human Activity Recognition Solving the Confusing Activities Via Deep Ensemble Learning

Zhu, R. Y.; Xiao, Zhenyu; Liu, Ying; Yang, Mingkun; Tan, Ya-Wen; Zhou, Liang; Lin, Shih‐Yen; Wen, Hongkai

doi:10.1109/access.2019.2922104

Cited by 74 publications

(50 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Other studies combined accelerometer and magnetometer simultaneously [39], acceleration and gyroscope with magnetometer [40,41], accelerometer with microphone and GPS [6], and other combinations [42].…”

Section: Data Acquisitionmentioning

confidence: 99%

“…Random Forest (RF) is a classifier consisting of a collection of tree-structured classifiers {h(x, k ), k = 1, ...} where the { k } are independent identically distributed random vectors and each tree casts a unit vote for the most popular class at input x [83]. Random Forest generally achieves high performance with high-dimensional data by increasing the number of trees [29,40,48,56,75,84,85].…”

Section: Traditional Machine Learningmentioning

confidence: 99%

See 1 more Smart Citation

Trends in human activity recognition using smartphones

Ferrari

Micucci

Mobilio

et al. 2021

J Reliable Intell Environ

View full text Add to dashboard Cite

Recognizing human activities and monitoring population behavior are fundamental needs of our society. Population security, crowd surveillance, healthcare support and living assistance, and lifestyle and behavior tracking are some of the main applications that require the recognition of human activities. Over the past few decades, researchers have investigated techniques that can automatically recognize human activities. This line of research is commonly known as Human Activity Recognition (HAR). HAR involves many tasks: from signals acquisition to activity classification. The tasks involved are not simple and often require dedicated hardware, sophisticated engineering, and computational and statistical techniques for data preprocessing and analysis. Over the years, different techniques have been tested and different solutions have been proposed to achieve a classification process that provides reliable results. This survey presents the most recent solutions proposed for each task in the human activity classification process, that is, acquisition, preprocessing, data segmentation, feature extraction, and classification. Solutions are analyzed by emphasizing their strengths and weaknesses. For completeness, the survey also presents the metrics commonly used to evaluate the goodness of a classifier and the datasets of inertial signals from smartphones that are mostly used in the evaluation phase.

show abstract

Section: Data Acquisitionmentioning

confidence: 99%

Section: Traditional Machine Learningmentioning

confidence: 99%

Trends in human activity recognition using smartphones

Ferrari

Micucci

Mobilio

et al. 2021

J Reliable Intell Environ

View full text Add to dashboard Cite

show abstract

“…Specifically, some other studies used convolutional neural network (CNN) to process the data gathered from sensors so as to detect different human activities [50] and recognize some hand gestures by using an IMU [51]. We decided to apply this principle based on [52] and [53] on the rover to detect its state. According to this algorithm, the swarm will have all the necessary information to carry out some choices and adapt itself to the environmental constraints.…”

Section: Related Workmentioning

confidence: 99%

“…These features are shown in Table 3 (where A x,i , A y,i , A z,i are measurements taken in the axis of accelerometer, and G x,i , G y,i , G z,i are measurements of the gyroscope; i is the sample of the window). With these measurements, we created an image of 9×9×1 pixels as input to the CNN based on the method used in [52] and [53]. To do that, we took the value of the feature and duplicated it on all the lines, one line for one feature.…”

Section: Cnn For States Differentiationmentioning

confidence: 99%

Toward a robot swarm protecting a group of migrants

Vaidis

Otis

2020

Intel Serv Robotics

View full text Add to dashboard Cite

Different geopolitical conflicts of recent years have led to mass migration of several civilian populations. These migrations take place in militarized zones, indicating real danger contexts for the populations. Indeed, civilians are increasingly targeted during military assaults. Defense and security needs have increased; therefore, there is a need to prioritize the protection of migrants. Very few or no arrangements are available to manage the scale of displacement and the protection of civilians during migration. In order to increase their security during mass migration in an inhospitable territory, this article proposes an assistive system using a team of mobile robots, labeled a rover swarm that is able to provide safety area around the migrants. We suggest a coordination algorithm including CNN and fuzzy logic that allows the swarm to synchronize their movements and provide better sensor coverage of the environment. Implementation is carried out using on a reduced scale rover to enable evaluation of the functionalities of the suggested software architecture and algorithms. Results bring new perspectives to helping and protecting migrants with a swarm that evolves in a complex and dynamic environment.

show abstract

“…Thusly, expansion in information in various people develops the person's action acknowledgment exactness. [4].…”

Section: Radha Mothukuri Tunuguntla Aishwarya Chalasani Himasree Dmentioning

confidence: 99%

Human Activity Recognition using Smartphone

Rasekh¹,

Chen²,

Lü³

2019

IJRTE

View full text Add to dashboard Cite

This paper consists of all the actions that are done by the people by using mobile phone. We used to record the actions by using the mobile phone. The main aim of this paper is to construct a classification model that is required for identifying actions of the people. This paper is used mainly to get solution of multi-classification problems. By using in a theoretic way, we can understand what the problem is, but we can only solve the problem by performing it in mathematical way. We can get very perfect result by solving in the mathematical way. Here we are deriving the actions of the persons using mobile phone, in a phone there are many sensors present in it. The sensors used in this paper are Accelerometer, Gyroscope. They are required for determining the actions of the person. The output obtained in this paper is used to compare the values in the term’s accuracy and precision. It uses a 3-dimension based accelerometer in order to collect the values obtained; there we determined that 31 values we contained in it. All the actions that are present are derived by using machine learning algorithms, they are, Naïve Bayes Classifiers, support vector machine, and neural networks. The output of the action determination by using the dataset required is used to determine a decrease of marking work to accomplish similar execution with machine learning.

show abstract

Efficient Human Activity Recognition Solving the Confusing Activities Via Deep Ensemble Learning

Cited by 74 publications

References 31 publications

Trends in human activity recognition using smartphones

Trends in human activity recognition using smartphones

Toward a robot swarm protecting a group of migrants

Human Activity Recognition using Smartphone

Contact Info

Product

Resources

About