Comparison of exact static and dynamic Bayesian context inference methods for activity recognition

Frank, Korbinian; Röckl, Matthias; Nadales, María Josefa Vera; Robertson, Patrick; Pfeifer, Tom

doi:10.1109/percomw.2010.5470671

Cited by 11 publications

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Bayslets enable a complex bayesian network to be subdivided into smaller RV groups (see VII). For example, motion and presence activity may be treated as separate Bayesian Networks, implemented using different JBI "java beans" and mathematically combined using the rules described in [22]. This is an inherently more scalable approach than the one we have taken as context inference is more easily expressed as smaller networks that are "fused" based on well defined and rules.…”

Section: Conclusion and Discussion Of Ongoing Workmentioning

confidence: 99%

Using Ambient Intelligence to Infer Availability and Meta-status in a Picocell Environment

Dempsey¹,

Jennings²,

Brazil³

et al. 2011

2011 Seventh International Conference on Intelligent Environments

View full text Add to dashboard Cite

Abstract-Pervasive mobile communication is set to be greatly enhanced by the development and deployment of miniature mobile base stations, suitable for workplace or home. These base stations, known as Femtocells and Picocells, connect to traditional broadband connections, to offer greatly improved 3G communications to mobile handsets. Developing an architecture that allows for targeted applications to be developed for end users of these enhanced mobile networks is of great importance. Examining how contextual data, extracted from these mobile networks, can be used to make inference and provide benefit to end users is both a technical and sophisticated challenge to be discussed in this paper.

show abstract

Section: Conclusion and Discussion Of Ongoing Workmentioning

confidence: 99%

Using Ambient Intelligence to Infer Availability and Meta-status in a Picocell Environment

Dempsey¹,

Jennings²,

Brazil³

et al. 2011

2011 Seventh International Conference on Intelligent Environments

View full text Add to dashboard Cite

show abstract

“…В работах по распознаванию видов активности человека [14,15,18,20,37,39] датчики размещали на следующих частях тела человека: руки, запястье, грудь, пояс, лодыжки, а также в карманах, расположенных в районе бедра человека.…”

Section: размещение и ориентация датчиковunclassified

Survey on Approaches and Practical Areas of Human Activity Recognition Application

2019

CMSE

View full text Add to dashboard Cite

Распознавание видов физической активности человека является одним из актуальных направлений исследования в области машинного обучения, так как результаты распознавания необходимы при решении многих практических задач. В статье приводится обзор подходов и практических областей применения методов распознавания видов физической активности человека. Рассматриваются датчики, используемые для распознавания видов физической активности человека, и представлены критерии их выбора. Представлены возможные пути решения проблемы выбора места размещения и ориентации носимых датчиков. В статье рассматриваются основные этапы распознавания видов физической активности человека. Представлены извлекаемые признаки и методы их отбора для повышения точности классификации видов физической активности человека и снижения вычислительных затрат за счет уменьшения числа признаков. Сформулированы достоинства и недостатки популярных методов классификации. Рассматриваются метрики, используемые для оценки качества обучения модели классификации. Наиболее применяемой метрикой качества является кривая ошибок. Также представлены практические задачи, в которых необходимы результаты распознавания видов физической активности человека. Основными областями применения метода распознавания являются медицина, производство, фитнес и безопасность людей. В заключении представлены возможные направления будущих исследований. Ключевые слова: распознавание образов, машинное обучение, виды физической активности человека.

show abstract

“…Also, it can provide great help for users to identify the type of a given context before using it. As context could be categorized from different perspective, the list in Table 1 summarizes a few well-known forms of context classification given in the current literature [ 16 , 17 , 18 , 19 , 20 , 21 , 22 ].…”

Section: Principles Of Context Aware Environmentsmentioning

confidence: 99%

Context Aware Middleware Architectures: Survey and Challenges

Eckert

Martínez

et al. 2015

Sensors

123

View full text Add to dashboard Cite

Context aware applications, which can adapt their behaviors to changing environments, are attracting more and more attention. To simplify the complexity of developing applications, context aware middleware, which introduces context awareness into the traditional middleware, is highlighted to provide a homogeneous interface involving generic context management solutions. This paper provides a survey of state-of-the-art context aware middleware architectures proposed during the period from 2009 through 2015. First, a preliminary background, such as the principles of context, context awareness, context modelling, and context reasoning, is provided for a comprehensive understanding of context aware middleware. On this basis, an overview of eleven carefully selected middleware architectures is presented and their main features explained. Then, thorough comparisons and analysis of the presented middleware architectures are performed based on technical parameters including architectural style, context abstraction, context reasoning, scalability, fault tolerance, interoperability, service discovery, storage, security & privacy, context awareness level, and cloud-based big data analytics. The analysis shows that there is actually no context aware middleware architecture that complies with all requirements. Finally, challenges are pointed out as open issues for future work.

show abstract

Comparison of exact static and dynamic Bayesian context inference methods for activity recognition

Cited by 11 publications

References 14 publications

Using Ambient Intelligence to Infer Availability and Meta-status in a Picocell Environment

Using Ambient Intelligence to Infer Availability and Meta-status in a Picocell Environment

Survey on Approaches and Practical Areas of Human Activity Recognition Application

Context Aware Middleware Architectures: Survey and Challenges

Contact Info

Product

Resources

About