隨著微電子技術的廣大的發展，許多的感測器現在可以整合到智慧型手機上。它們利用豐富的感測器來偵測環境背景，系統狀態和用戶活動，來開發出許多情境感知的應用程序(Context-Aware applications)。對於智慧型手機而言，不同的應用程序可能會要求相同的情境是很常見的事情，情境可以通過感測器的不同組合來推斷。目前的手機系統並沒有去協調情境感知的應用程序中所要使用的感測器，導致開啟冗餘的感測器和造成不必要的能源消耗。在這篇論文中，我們提出一個介於感測器和應用程式之間的精準度/耗能優化之情境感測框架，它能挑選最佳的感測器組合，來滿足所有應用程序的情境感知的需求。我們提供不同兩個優化條件：耗能優化，在滿足所有需求的精準度要求下，盡可能降低能源的消耗。精準度優化，在給定能量預算的情況下，盡可能最大化整體的精準度。我們的實驗結果表明，跟原本Android不協調的感測器管理機制下比較，使用我們提出的情境感測框架，在耗能優化下，耗電平均降低30.45%；在精準度優化下，耗電平均降低19.91%。

With the vast developments of microelectronics, a multitude of sensors can now be packed into a smart phone. This has enabled many context-aware applications that utilize the rich set of sensors to sense the environmental contexts, system status, and user activities. For a smart phone, it is common that different applications may request the same context, while a context may be inferred by different combinations of sensors. Current phone systems do not coordinate the use of sensors by the context-aware applications, leading to redundant sensors activated and unnecessary energy consumption. In this thesis, we propose an Energy/Accuracy-Optimized Framework (EAOF), which sits between the applications and the low-level sensors to provide a coordinated and optimized use of sensors to satisfy the context-sensing requirements of the applications. The use of sensors may be optimized based on two criteria: energy-optimized, which minimizes the total energy consumption while maintaining a target accuracy, and accuracy-optimized, which maximizes the overall accuracy under a given energy budget. Our experimental results show that the power consumption can be reduced by 30.45% in the energy-optimized mode and 19.91% in the accuracy-optimized mode, compared with the original, uncoordinated sensor management on Android.

Abstract
Contents
Acknowledgments 1
1 Introduction 1
2 Framework Design 4
2.1 EAOF Overview 4
2.1.1 APIs 4
2.1.2 Request Manager 4
2.1.3 Resource Manager 6
2.1.4 Context Analyzer 6
2.1.5 System Model 7
2.2 Methods 8
2.2.1 Energy-Optimized Mode 8
2.2.2 Accuracy-Optimized Mode 10
3 Implementation 12
4 Evaluation 14
4.1 Evaluation Setup 14
4.2 Evaluation Results 15
5 Related Work 19
6 Conclusion 20

[1] Suman Nath. Ace: exploiting correlation for energy-efficient and continuous context sensing. In Proceedings of the 10th international conference on Mobile systems, applications, and services, MobiSys ’12, pages 29–42, New York, NY, USA, 2012. ACM.
[2] Hong Lu, Jun Yang, Zhigang Liu, Nicholas D. Lane, Tanzeem Choudhury, and Andrew T. Campbell. The jigsaw continuous sensing engine for mobile phone applications. In Proceedings of the 8th ACM Conference on Embedded Networked Sensor Systems, SenSys ’10, pages 71–84, New York, NY, USA, 2010. ACM.
[3] Seungwoo Kang, Jinwon Lee, Hyukjae Jang, Hyonik Lee, Youngki Lee, Souneil Park, Taiwoo Park, and Junehwa Song. Seemon: scalable and energy-efficient context monitoring framework for sensor-rich mobile environments. In Proceedings of the 6th international conference on Mobile systems, applications, and services, MobiSys ’08, pages 267–280, New York, NY, USA, 2008. ACM.
[4] Dana Popovici, Mikael Desertot, Sylvain Lecomte, and Thierry Delot. A framework for mobile and context-aware applications applied to vehicular social networks. Social Netw. Analys. Mining, 3(3):329–340, 2013.
[5] I. Anderson and H. Muller. Practical context awareness for gsm cell phones. In Wearable Computers, 2006 10th IEEE International Symposium on, pages 127–128, 2006.
[6] Y. Kawahara, H. Kurasawa, and Hiroyuki Morikawa. Recognizing user context using mobile handsets with acceleration sensors. In Portable Information Devices, 2007. PORTABLE07. IEEE International Conference on, pages 1–5, 2007.
[7] Norbert Györbíró, Ákos Fábián, and Gergely Hományi. An activity recognition system for mobile phones. Mob. Netw. Appl., 14(1):82–91, February 2009.
[8] Jennifer R. Kwapisz, Gary M. Weiss, and Samuel A. Moore. Activity recognition using cell phone accelerometers. SIGKDD Explor. Newsl., 12(2):74–82, March 2011.
[9] Yi Wang, Jialiu Lin, Murali Annavaram, Quinn A. Jacobson, Jason Hong, Bhaskar Krishnamachari, and Norman Sadeh. A framework of energy efficient mobile sensing for automatic user state recognition. In Proceedings of the 7th international conference on Mobile systems, applications, and services, MobiSys ’09, pages 179–192, New York, NY, USA, 2009. ACM.
[10] The future of context aware mobile computingbal. http://www.coconutdaily.com/ contextawarecomputing.
[11] Endomondo sports tracker. http://www.endomondo.com/.
[12] Yelp mobile. http://www.yelp.com/yelpmobile.
[13] In-meeting. https://groups.google.com/forum/#!forum/just-4-android.
[14] Seyed Amir Hoseini-Tabatabaei, Alexander Gluhak, and Rahim Tafazolli. A survey on smartphone-based systems for opportunistic user context recognition. ACM Comput. Surv., 45(3):27:1–27:51, July 2013.
[15] Aaron Carroll and Gernot Heiser. An analysis of power consumption in a smartphone. In Proceedings of the 2010 USENIX conference on USENIX annual technical conference, USENIXATC’ 10, pages 21–21, Berkeley, CA, USA, 2010. USENIX Association.
[16] Zhixian Yan, Vigneshwaran Subbaraju, Dipanjan Chakraborty, Archan Misra, and Karl Aberer. Energy-efficient continuous activity recognition on mobile phones: An activity-adaptive approach. In Proceedings of the 2012 16th Annual International Symposium on Wearable Computers (ISWC), ISWC ’12, pages 17–24, Washington, DC, USA, 2012. IEEE Computer Society.