HyperspaceFlow: A System-Level Design Methodology for Smart Space

Abstract

Smart space technology gains intensive attentions of the worldwide due to the advancement of pervasive computing. However, the design of smart space is inherently complex as it resides in highly varying environment with multiple devices, also the power consumption, cost, and user preference are usually ignored by designers, resulting in huge design cost, high power dissipation, and unsatisfactory user experience. In this paper, we present a system-level design methodology for smart space, which is able to refine the function specification of smart space into underlying implementations leveraging an intermediate representation model HyperspaceFlow. The tasks in smart space are deposited into distributed execution platforms, and a network is automatically synthesized to satisfy their needs of communication at given QoS constraints. Meanwhile, the proposed design methodology is able to offer relatively satisfactory cost, power consumption, and user preference of the design solution at given constraints. At last, a home health care case is used to elucidate our design methodology, and the simulation results demonstrate the effectiveness and feasibility of our methodology.