NESL Technical Report #: 2011-3-1
Authors:
Abstract: In recent days, smartphones offer more advanced computing ability and connectivity than in the past. A smartphone has a lot of on-the-phone sensors, such as GPS, accelerometer, microphone, camera etc. In addition, a smartphone application can employ remote wireless body sensors, such as electrocardiogram (ECG) sensors or temperature skin sensors to track a user’s behavior and provide useful information for his health. There have been studies where a smartphone was tracking the sensor’s raw data and was processing them on the phone-side. However, now we can use “smart” sensors, which can do some computation on sensor-side and feed the smartphone with processed data. In such a way, the overall performance will be better as some computation would be done much closer to the hardware (on sensor-side). In this paper, the design of a dynamic, data-flow architecture for extracting context-aware information is presented. In the proposed design, the whole process of tracking and handling the sensor’s data is provided to the smartphone’s applications as a Data-Flow Service and is totally separated from the smartphone’s application itself. The user can remotely turn-on/off the sampling of the sensor’s signal and dynamically choose whether the computation will be done on the sensor-side or on the phone-side. In order to present the data-flow architecture, an application on Android HTC Wildfire has been implemented. It uses the DFS, tracks ECG data from a remote ECG sensor and determines if the user has arrhythmia or not.
Page (Count): 10
Date: 2011-03-01
Publisher: Zafeiria Anagnostopoulou
Public Document?: Yes
NESL Document?: Yes
Document category: Other
Back