Energy Enhancement of Multi-application Monitoring Systems for Smart Buildings

Abstract

Abstract. High energy consumption is a major problem in smart building systems. Existing studies focus on energy consumption of building not the deployed wireless sensors. These approaches are often fitted to a single monitoring application, and lead to static configurations for sensor devices. Moreover, immense raw data generated by the smart building should be used in terms of service. In this paper, we focus on the energy of the monitoring architecture itself and services that use this dataset. We study impacts of services on the energy consumption of monitoring architecture. We consider a monitoring system as a set of applications that exploit sensor measures in real-time, where these applications are declaratively expressed as (service-oriented) continuous queries over sensor data streams. We focalise on multi-application, tackle the optimization of application requirements to manage energy consumption. We introduce a Smart-Service Stream-oriented Sensor Management (3SoSM) that optimizes sensor configurations and manages sensor data streams.

Lightweight Metric Computation for Distributed Massive Data Streams

“…We have then shown that it behaves pretty well whatever the considered input streams. We are convinced of the indisputable interest of such a metric in various domains including Internet of Things statistical usages as network monitoring and information retrieval [7], and we think that it should pertinent in machine learning, and data mining applications as discussed in [9].…”

“…Finally, the sustainable development of smart cities is expected to handle large amounts of data generated from large number of sensors with the consequent necessity for quick aggregation of the data, which could be exploited to detect correlated events. Among possible applications, smart building management systems rely on service-oriented continuous queries over sensor data streams in case of energy consumption monitoring [7], air pollution monitoring applications heavily rely on sensors to detect threshold crossings [8]. More generally, Stankovic [9] argues that the real time analysis of large and distributed data streams is of utmost importance to tackle issues related to creative knowledge, robustness, privacy, and security.…”