Abstract

This paper presents an unobtrusive, energy-efficient approach to human activity sensing through the intelligent scheduling of built-in sensors on mobile phones and light-weight compressed sensing. We refer to this framework as pattern-based compressed phone sensing (P-CPS) where two challenging issues are studied, the energy drainage issue due to continuous sensing which may impede the normal functionality of the mobile phones and the requirement of active user inputs for data collection that may place a high burden on the user. The proposed P-CPS framework consists of two stages - training stage and sensing stage. In the training stage, a Pattern Matrix (PM) is constructed and an adaptive sensing scheme is used to update the PM automatically in order to keep records of a user's activity occurrences. In the sensing stage, P-CPS incorporates a Gaussian mixture model-based activity modeling and the adaptive sensing scheme for sensing scheduling. Compressed sensing (CS) is applied during the activity signal acquisition process. P-CPS uses a sparse binary measurement matrix which results in only simple matrix additions at the mobile side for energy efficiency purpose. Experimental results on driving activity sensing show that P-CPS can have, on average, the sensing scheduling accuracy about 70% but with 62.86% less energy consumption as compared to the continuous sensing.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call