RTOS-based power management in embedded systems

• Main Author: Muhamed Fauzi Abbas
• Other Authors: Thambipillai Srikanthan
• Format: Theses and Dissertations
• Language: English
• Published: 2011
• Subjects: DRNTU::Engineering::Computer science and engineering::Software::Operating systems
• Online Access: https://hdl.handle.net/10356/46307
• Institution: Nanyang Technological University

Power Management has increasingly become one of the most important features in modern embedded systems. This is in addition to other important constraints such as Non Recurring Engineering (NRE) costs and shorter Time to Market (TTM) due to shrinking product life cycles. In this research work, a systematic approach to incorporating power management feature into RTOS has been proposed. Comprehensive literature reviews of existing power management techniques that rely on operating systems were performed in an attempt to establish suitable techniques for constraint-aware embedded systems. This led to the detailed examination of the popular method of OS based power management technique known as Advanced Configuration and Power Interface (ACPI). ACPI was adapted to cater for RTOS based embedded systems and the proposed power management framework was implemented on the Renesas SH7727 using T-Kernel as the target RTOS. It was shown that a centralised power management could be realized using this framework so as to facilitate the rapid incorporation of new devices and integration of new power management policies without necessitating substantial engineering effort. In addition, the added overhead to RTOS is minimal and it can be easily adapted to suit a variety of RTOS. A systematic technique to deploy power policy algorithms along with RTOS-ACPI has been proposed as part of this work. The proposed technique ensures that it is capable of allowing power policy algorithms to quickly access and to manage the platform specific power features. A technique to improve application-awareness was also evaluated to show that application specific information can be relied upon to identify power-switch points that RTOS-ACPI can depend on to manage power. The proposed methods pave way for incorporating a comprehensive and scalable RTOS centric power management support into resource constraint-aware embedded systems.