Selected Articles from IEEE Xplore - June 2016
Introduction by Yen-Kuang Chen, An-Yeu (Andy) Wu, Magdy A. Bayoumi, and Farinaz Koushanfar
In the IoT era, data analysis will play a key role, as we need layers of intelligence to transform the data into wisdom. However, many analytic algorithms assume that the system has all the data on the server. It takes power and bandwidth to communicate the data to the server. Moreover, not all the data is important or useful. Architecture-wise, it is natural for sensor data to be processed in a hierarchical and distributed fashion. Data may be analyzed and fused in sensors or gateways before arriving at the data center to save energy and bandwidth. If the system can recognize what context is important in the very beginning, it only needs to transmit the relevant information to the backend server or the cloud. As devices become more computationally capable, intelligent computation may easily be distributed among the sensors and the backend servers. This month we would like to introduce three papers on low-power data analysis.
First, among different sensors used in IoT, video sensors can provide the richest amount of information. However, the power consumption and bandwidth requirement from video sensors may hinder the wide deployment of them wirelessly. In order to provide a foundation for future distributed video sensor designs, “Power Consumption Analysis for Distributed Video Sensors in Machine-to-Machine Networks” by Chien et al. analyzes and compares the power consumption distributed video coding versus the state-of-the-art H.264 coding.
Second, in order to enable good context analysis on the energy-constrained devices, it is important to consider algorithm-level optimization. “Cascading Signal-Model Complexity for Energy-Aware Detection” by Jun and Jones presents an analysis and design of a low-power detection algorithm.
Third, for accurate signal classification, there is often a need for high computational complexity. However, in reality, the classification must run at an extremely low power in order to survey for a long period of time. “Design of a Low-Power On-Body ECG Classifier for Remote Cardiovascular Monitoring Systems” by Chen et al. provides a comprehensive analysis on the trade-off between power consumption and classification accuracy. Furthermore, it presents a practical implementation of the classifier.
IEEE Xplore References
- S. Y. Chien et al., "Power Consumption Analysis for Distributed Video Sensors in Machine-to-Machine Networks," in IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 3, no. 1, pp. 55-64, March 2013.
- D. Jun and D. L. Jones, "Cascading Signal-Model Complexity for Energy-Aware Detection," in IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 3, no. 1, pp. 65-74, March 2013.
- T. Chen, E. B. Mazomenos, K. Maharatna, S. Dasmahapatra and M. Niranjan, "Design of a Low-Power On-Body ECG Classifier for Remote Cardiovascular Monitoring Systems," in IEEE Journal on Emerging and Selected Topics in Circuits and Systems, vol. 3, no. 1, pp. 75-85, March 2013.
Subscribe to the IEEE Internet of Things Technical Community to receive complimentary access to our selected articles each month via email.