Abstract
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.
Highlights
Observing, recognizing, and analyzing human activities form a foundation for scientific fields such as anthropology, archeology, sociology, and psychology
For our change point approach to mobile energy reduction we propose SEP, a SEParation distance strategy, because we showed it to be more sensitive to subtle changes in sensor time series data than other unsupervised methods and it is non-parametric [45,66]
How does Change Point-based Activity Monitoring (CPAM) compare with baseline methods for activity recognition performance?
Summary
Observing, recognizing, and analyzing human activities form a foundation for scientific fields such as anthropology, archeology, sociology, and psychology. Given the 127 million smartwatches that were sold last year alone, the volume of already-collected activity is unprecedented Researchers can analyze this data to validate theories of human behavior and practitioners can gain insights that allow them to provide personalized recommendations and treatment plans. Activity Monitoring (CPAM), an algorithm that performs continual monitoring and recognition of activities of daily living while using change point detection and change point-adaptive sampling to reduce energy consumption. Our long-term goal is to recognize such complex activities, in real time as people perform them To achieve this goal, we must create new approaches to sensing that reduce energy consumption and save battery life. CPAM collects sensor and location data, detects activity changes, adjusts the sampling rate correspondingly, and recognizes activities of daily living in real time. We investigate an enhancement to CPAM that uses activity labels to estimate sensor values between sampling periods as a strategy to further reduce sampling rates while maintaining the ability to accurately detect and recognize critical activities
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have