Human activity monitoring based on hidden Markov models using a smartphone

San-Segundo, Rubén; Echeverry-Correa, Julián D.; Salamea, Christian; Pardo, José Manuel

doi:10.1109/mim.2016.7777649

Cited by 26 publications

(18 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because of its computational power, its wide acceptance and since it has been shown to work well in recognizing physical activities [8] [11], the smart-phone has been chosen as the primary input of data in this study. Besides, this system could have also worked with other IoT configurations such as with a smart-watch (alone or added to the smart-phone).…”

Section: Methods and Implementationmentioning

confidence: 99%

See 1 more Smart Citation

Energy expenditure estimation through daily activity recognition using a smart-phone

Bois

Amroun

Ammi

2018

2018 IEEE 4th World Forum on Internet of Things (WF-IoT)

View full text Add to dashboard Cite

This paper presents a 3-step system that estimates the real-time energy expenditure of an individual in a nonintrusive way. First, using the user's smart-phone's sensors, we build a Decision Tree model to recognize his physical activity (running, standing, ...). Then, we use the detected physical activity, the time and the user's speed to infer his daily activity (watching TV, going to the bathroom, ...) through the use of a reinforcement learning environment, the Partially Observable Markov Decision Process framework. Once the daily activities are recognized, we translate this information into energy expenditure using the compendium of physical activities. By successfully detecting 8 physical activities at 90%, we reached an overall accuracy of 80% in recognizing 17 different daily activities. This result leads us to estimate the energy expenditure of the user with a mean error of 26% of the expected estimation.

show abstract

Section: Methods and Implementationmentioning

confidence: 99%

“…Physical activity recognition has been a dynamic field of study in the past few years. First using IoT [6] [7], and then using smart-phones [8] [9] [10]. Through machine learning techniques applied to the smart-phone's accelerometer, they recognize the user's physical activities such as running, walking or standing.…”

Section: Introductionmentioning

confidence: 99%

Energy expenditure estimation through daily activity recognition using a smart-phone

Bois

Amroun

Ammi

2018

2018 IEEE 4th World Forum on Internet of Things (WF-IoT)

View full text Add to dashboard Cite

show abstract

“…Rather than the tag spatial coordinates, in practical scenarios with a large amount of tags, it could be reasonable to determine just the spatial region the tag belongs to [21]. This represents a classification issue, analogously to that addressed for discriminating among moving and static RFID tags [22], for person activities recognition in ambient assisted living applications [23], for imaging recognition [24] or for smell recognition [25].…”

Section: Introductionmentioning

confidence: 99%

RSSI Measurements for RFID Tag Classification in Smart Storage Systems

Buffi

Michel

Nepa

et al. 2018

IEEE Trans. Instrum. Meas.

View full text Add to dashboard Cite

This paper presents a measurement method for tagged item classification in radio frequency identification (RFID)-based smart drawers. Specifically, the amplitude of the signal backscattered by the UHF-RFID passive tags attached to the items inside the drawer is used to identify at which of an assigned set of drawer subregions each detected tag belongs. The received signal strength indicator (RSSI) may be acquired by almost any UHF-RFID commercial reader chipset compatible with the Electronic Product Code (EPC) Class1 Gen2 protocol. The RSSI model in terms of the position of the tag inside the drawer is introduced, and its selective properties are discussed through both numerical data and measurements. The cases of a single reader antenna measuring during natural opening/closing operations as well as of multiple reader antennas measuring when the drawer is static are investigated. The measurement method performance is shown through an experimental analysis in a realistic scenario, where the RSSI values are collected by using commercial UHF RFID hardware components, and then used as the feature vector of the classification algorithm. As an example, two different classifiers are used to map each detected tag to one of the regions the drawer is subdivided into. The classification performance of the method based on a single antenna exploiting the drawer natural sliding movements is comparable to that achievable with multiple reader antennas, while representing a cost-effective and easy-to-implement practical solution. Index Terms-Antenna near-field region, classification techniques, radio frequency identification (RFID) tag classification, received signal strength indicator (RSSI) measurements, RFID tag indoor localization, RFID-based smart drawers, RFID-based smart storage spaces, UHF-RFID measurements.Andrea Michel (S'08-M'15) received the B.E., M.E. (summa cum laude) and PhD degrees in telecommunications engineering from the University of Pisa, Pisa, Italy, in 2009, 2011 Since 2015 he has been a Post-Doctoral Researcher in applied electromagnetism at the Microwave and Radiation Laboratory, Department of Information Engineering, University of Pisa, focusing on the design of integrated antenna for communication systems and smart antennas for near-field UHF-radio frequency identification (RFID) readers. In 2014, he joined the ElectroScience Laboratory, The Ohio State University, Columbus, OH, USA, as a Visiting Scholar, where he has been involved in research on a theoretical analysis on the accuracy of a novel technique for deep tissue imaging. His current research interests include the design of antennas for automotive applications, multiple-in multiple-out systems, and wearable communication systems, in collaboration with other research institutes and companies.Dr. Michel received the Young Scientist Paolo Nepa (M'07) received the Laurea (Doctor) degree (summa cum laude) in electronics engineering from the University of Pisa, Italy, in 1990. Since 1990, he has been with the Department of Information Engineering, Uni...

show abstract

“…During last years, increased attention to fall detection has led to the development of several fall detection approaches [3,4]. Some methods are generally based on information gathered by sensors like vibration and acceleration sensors [3,4].…”

Section: Introductionmentioning

confidence: 99%

“…Some methods are generally based on information gathered by sensors like vibration and acceleration sensors [3,4]. These methods use vibrations, sound, and human movements in fall detection [4,5,6].…”

Section: Introductionmentioning

confidence: 99%