Abstract
Despite advances in low power system design, short battery life remains a significant user concern. Effective management of energy resources available on a mobile device requires understanding of the principles of battery behaviour. We propose a time-delay model of a battery, which depends upon three non-linear processes: rate-capacity effect, recovery effect and software scheduling effect. We provide an analysis of the power consumption results using 3DMark’06 chipset benchmarks and demonstrate that a moderate-to-strong correlation between power consumption vs. CPU load, memory allocation and memory release is observed. Finally, we apply our model for chipset energy efficiency profiling and propose a power benchmark metric.DOI: http://dx.doi.org/10.5755/j01.eee.19.6.4577
Highlights
Short battery life was, and still remains, a significant concern for mobile device users
There is a big gap between the energy resources needed by the mobile device and the energy available from the battery; the average battery life of actively used mobile devices is usually less than two days [1]
The most important factors influencing its lifetime are the battery’s capacity and the battery’s discharge rate. This rate is influenced by three non-linear processes, two of which are determined by the electrochemical properties of the battery
Summary
Still remains, a significant concern for mobile device users. Another matter of concern is the ability to predict battery life using available information on the application usage habits, the device’s modes of operation and power management schemes so that the user could decide how to use the remaining battery time most effectively Such prediction is only possible when the behaviour of the battery can be modelled accurately taking all internal (electro-chemical) and external (CPU load, memory usage, display rendering, etc.) factors that influence the State of Charge (SoC) of the battery into account. A third nonlinear process that has influence on the state of the battery is software scheduling schemes introduced at application [3] or operating system level, which control, e.g., CPU rate and energy consumption level of peripheral devices that are considered non-essential for some applications such as display brightness The result of these effects is the dependency of the battery lifetime upon battery discharge distribution over time [4], which in turn depends upon user behaviour and usage patterns.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
