M2M-based Anticipatory Reasoning for Contexts and Services

PI: Prof. Li-Chen Fu

Co-PI: Prof. 吳兆麟

Champion: Dr. Charlie Tai


    Most of traditional energy saving systems often need user interventions to manually preset some static energy-saving rules via not readily accessible interfaces. More importantly, these monolithic systems often fail to take users’ contexts and preferences into account. Therefore, we employ the M2M Infrastructure, which allows cost-effective integration and fast deployment of vast amount of remote devices, to achieve context-aware energy-saving and to minimize user's interventions in the determination of energy saving policies. The overall system architecture is illustrated as below. In this system, there are three major components, which are Distributed Hybrid Context (DHC) Inference Engine, Collaborative Energy-Saving (CES) Decision Support Engine, and Ambient Universal Control Interface (AUCI). Unlike the existing approaches often ignoring the context information such as users’ situations and preferences, the DHC Inference Engine infers users’ activities, preferences, and PMV (Predicted Mean Vote) value through techniques of activity recognition (AR), preference recognition (PR), and PMV evaluation; the system then aggregates them into useful context data. With the context data inferred by the DHC Inference Engine and environment factors extracted from environment, the CES Decision Support Engine deals with conflicts between users’ preferences and energy saving policies. That is, the CES Decision Support Engine will provide recommendations to user who selects his/her preferable policy concerning power saving or comfort level based on that policy while taking into accounts all the extractable factors. Finally, we propose a decentralized and friendly control interface called AUCI to present environment or system information in a coherent way to help users to quickly control/configure multiple devices and provide comprehensive device information via readily accessible interfaces so that efforts for human intervention can be minimized. Besides, AUCI also acts as a bridge between internal networks and cloud services, which will further facilitate service innovation for energy saving.

Fig. 1. Our smart home environment as one testbed for energy-saving evaluations