KLEM: A Method for Predicting User Interaction Time and System Energy Consumption during Application Design
2007 11th IEEE International Symposium on Wearable Computers (2007)
Boston, MA, USA
Oct. 11, 2007 to Oct. 13, 2007
Lu Luo , School of Computer Science, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213 USA, email@example.com
Daniel P. Siewiorek , School of Computer Science and Department of Electrical and Computer Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213 USA, firstname.lastname@example.org
The impact of user interactions on the electric energy consumption of a portable computer system and on user efficiency is often not obtainable until after the software application is implemented and deployed on a specific hardware platform. In this paper, we present the Keystroke-Level Energy Model (KLEM), a method that can predict the user time and system energy consumption it will take to perform an interactive task at run time during the phase of application design. KLEM is based on the Keystroke-Level Model (KLM), a psychological theory of human cognitive and motor capabilities that can predict execution time for a skilled user. We first create a design storyboard and define a set of tasks whose KLMs are to be constructed. We then construct KLEM of each task by correlating system activities to the user actions modeled in the corresponding KLM. We obtain the energy profiles of system activities from running a set of user interaction benchmarks on the target hardware platform. To verify KLEM, we conducted a user study of 10 participants on executing an information query task using eight different methods. The user time and system energy of the participants were measured on two popular handheld platforms: a Windows Mobile iPaq and a Palm OS Tungsten. Our experimental results show that KLEM has an average prediction error of 5.6% and 8.8% on user time, and 4.4% and 8.4% on energy consumption on the two platforms, respectively.
L. Luo and D. P. Siewiorek, "KLEM: A Method for Predicting User Interaction Time and System Energy Consumption during Application Design," 2007 11th IEEE International Symposium on Wearable Computers(ISWC), Boston, MA, USA, 2007, pp. 1-8.