Thyme: Improving Smartphone Prompt Timing Through Activity Awareness

Aminikhanghahi, Samaneh; Fallahzadeh, Ramin; Sawyer, Matthew; Cook, Diane J.; Holder, Lawrence B.

doi:10.1109/icmla.2017.0-141

Cited by 11 publications

(9 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To accommodate this difference in approach, CPAM moves a sliding window over the data. For this paper, the window size, w, is set to 5 s motivated by experiments reported from our group and others [42,43]. Features are extracted from a window and the supervised learning algorithm maps this feature vector onto an activity label, <f statistical , f relational , f temporal , f navigational , f personal , f positional >→A.…”

Section: Real-time Activity Recognitionmentioning

confidence: 99%

“…Formally, given a time series stream of elements X = {x1,..., Activity recognition is performed on the watch using the CoreML libraries. Earlier experiments indicated that random forest with 100 trees performs well on activity recognition from wearable data [42] and we utilize this algorithm for CPAM. The collected features are generalizable, so we build a model that can be used for any existing or new user.…”

Section: Sep Change Point Detectionmentioning

confidence: 99%

“…Rather than considering the data itself as an indication of when location is needed, the baseline method collects user location information every five minutes then turns it off until the next five-minute increment is reached. Sensing activity information every five minutes represents a strategy that has been previously used to monitor activity without draining the battery [42]. Additionally, the shortest monitored activity is approximately five minutes.…”

Section: Energy Reductionmentioning

confidence: 99%

See 2 more Smart Citations

Easing Power Consumption of Wearable Activity Monitoring with Change Point Detection

Culman

Aminikhanghahi

Cook

2020

Sensors

Self Cite

View full text Add to dashboard Cite

Continuous monitoring of complex activities is valuable for understanding human behavior and providing activity-aware services. At the same time, recognizing these activities requires both movement and location information that can quickly drain batteries on wearable devices. In this paper, we introduce Change Point-based Activity Monitoring (CPAM), an energy-efficient strategy for recognizing and monitoring a range of simple and complex activities in real time. CPAM employs unsupervised change point detection to detect likely activity transition times. By adapting the sampling rate at each change point, CPAM reduces energy consumption by 74.64% while retaining the activity recognition performance of continuous sampling. We validate our approach using smartwatch data collected and labeled by 66 subjects. Results indicate that change point detection techniques can be effective for reducing the energy footprint of sensor-based mobile applications and that automated activity labels can be used to estimate sensor values between sampling periods.

show abstract

Section: Real-time Activity Recognitionmentioning

confidence: 99%

Section: Sep Change Point Detectionmentioning

confidence: 99%

Section: Energy Reductionmentioning

confidence: 99%

See 1 more Smart Citation

Easing Power Consumption of Wearable Activity Monitoring with Change Point Detection

Culman

Aminikhanghahi

Cook

2020

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…Bouchard et al [22] used raw latitude and longitude as the features to improve activity recognition performance and also discussed the potential of distance, position, shape, and gesture as features. Aminikhanghahi et al [23] developed an approach, called Thyme, for adapting prompt timing that is based on the context of the user’s activity. Liao et al [24] attempted to recognize activities from GPS traces through training a conditional random field to iteratively re-estimate significant places and activity labels.…”

Section: Related Workmentioning

confidence: 99%

Improving IoT Predictions through the Identification of Graphical Features

Akter

Holder

2019

Sensors

View full text Add to dashboard Cite

IoT sensor networks have an inherent graph structure that can be used to extract graphical features for improving performance in a variety of prediction tasks. We propose a framework that represents IoT sensor network data as a graph, extracts graphical features, and applies feature selection methods to identify the most useful features that are to be used by a classifier for prediction tasks. We show that a set of generic graph-based features can improve performance of sensor network predictions without the need for application-specific and task-specific feature engineering. We apply this approach to three different prediction tasks: activity recognition from motion sensors in a smart home, demographic prediction from GPS sensor data in a smart phone, and activity recognition from GPS sensor data in a smart phone. Our approach produced comparable results with most of the state-of-the-art methods, while maintaining the additional advantage of general applicability to IoT sensor networks without using sophisticated and application-specific feature generation techniques or background knowledge. We further investigate the impact of using edge-transition times, categorical features, different sensor window sizes, and normalization in the smart home domain. We also consider deep learning approaches, including the Graph Convolutional Network (GCN), for the elimination of feature engineering in the smart home domain, but our approach provided better performance in most cases. We conclude that the graphical feature-based framework that is based on IoT sensor categorization, nodes and edges as features, and feature selection techniques provides superior results when compared to the non-graph-based features.

show abstract

“…As mentioned in Section 5.1.2, a starting point here could be to avoid prompting for deliberative tasks at times when one is obviously preengaged or already in the situation that the prompt is about. That said, herein clearly lies a major research challenge, one that has been receiving more and more attention recently (e.g., see Aminikhanghahi, Fallahzadeh, Sawyer, Cook, and Holder, 2017;Bidargaddi, Almirall, Murphy, Nahum-Shani, Kovalcik, Pituch, Maaieh, and Strecher, 2018), but still requires further investigation.…”

Section: Supporting Users' Vigilancementioning

confidence: 99%

Autonomy-Respectful E-Coaching Systems

Kamphorst¹

View full text Add to dashboard Cite

The argument in favor of a more narrowly construed denition of e-coaching systems formed the basis for Kamphorst (2017), "E-Coaching Systems: What They Are, and What They Aren't", published in in a special issue of Personal and Ubiquitous Computing on Supporting a Healthier Lifestyle with e-Coaching Systems, volume 21, issue 4, pp. 625-632. A rst exploration of the relation between personal autonomy and the use of e-coaching systems can be found in Kamphorst & Kalis (2015), "Why option generation matters for the design of autonomous e-coaching systems", in AI & SOCIETY, volume 30, issue 1, pp. 77-88. Finally, elements from the subsection on privacy concerns in Chapter

show abstract

Thyme: Improving Smartphone Prompt Timing Through Activity Awareness

Cited by 11 publications

References 34 publications

Easing Power Consumption of Wearable Activity Monitoring with Change Point Detection

Easing Power Consumption of Wearable Activity Monitoring with Change Point Detection

Improving IoT Predictions through the Identification of Graphical Features

Autonomy-Respectful E-Coaching Systems

Contact Info

Product

Resources

About