- 學位論文
類視訊壓縮之無線感測網路平行資料存取方法
Video-Like Compression for Parallel Data Access on Wireless Sensor Networks
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
無線感測網路(wireless sensor networks, WSN)是由許多資源有限的感測器所組成，感測器可以蒐集並且監測環境上的變化，無線感測網路能夠滿足環境監測上多樣的需求，為了蒐集並儲存這些資訊，無線感測網路必須採取適當的方法來組織並且壓縮感測資料，否則，感測資料必定會佔據大量的存儲空間並且降低資料伺服器(data server)的效能。基於這些原因，我們之前提出類視訊無失真壓縮方法，稱為VLLC (Video-Like Lossless Compression)，目標在於利用感測資料的特性。空間相關性以及時間相關性來提高資料壓縮率和減少資料壓縮的時間；在類視訊無失真壓縮系統中，會參考相關性將原始的感測資料無失真地轉換並排列成固定格式的資料幀(data frame)，而這些資料幀形成一個3D的立體像素(voxel)，並可用H.264進行視訊壓縮，由於立體像素結構的關係，資料能被直接存取而不需要解壓縮所有的壓縮檔案。本論文將針對類視訊無失真壓縮提供有效率的資料查詢流程，並提供語法讓使用者查詢感測資料，除此之外，我們還設計平行處理方法以提高壓縮和解壓縮速率，分析並比較不同資料擺放法(data placement)的效能，提高整體效率。在我們的實驗中，對於4.53GB的感測資料進行類視訊壓縮能較未壓縮省下超過92%的資料空間，並且壓縮時間能低於43秒。另外，感測資料在16台平行處理平台下，採用合適的資料擺放法，將可比隨機擺放節省下62%的處理時間。
並列摘要
Wireless Sensor Networks (WSNs) consist of groups of resource-restricted sensor nodes that collect sensory data and monitor environmental changes. WSN environment services gather sensory data for various purposes. To store the collected information, systems should organize and compress sensory data using proper methods. Otherwise, sensory data will occupy a large amount of storage and decrease the server’s performance. In this thesis, we proposed a video-like lossless compression (VLLC), which aims to adopt the spatial correlation of sensory data in WSNs to enhance the degree of space saving and reduce the data compression time. In VLLC, systems will transform and arrange raw data as formatted video frames without loss according to the spatial correlation. The video frames form 3D voxels that can be highly compressed by H.264. Based on the voxel structure, data can be directly accessed without extracting all the compressed data. VLLC provides an efficient processing flow for querying sensory data and a query command that allows clients to access the proposed database. In our experiment, a space saving of more than 92% was achieved, and the data compression time for 4.53 GB of sensory data was less than 43 seconds. Furthermore, VLLC also offers a parallel processing method to enhance compression and decompression speed. To enhance the efficiency of the system, we also analysis and compare the different data placement methods. In our experiments, if we take proper personal data placement method, we will save 62% processing time with 16 personal computers more than random placement.
參考文獻
[2] D. Culler, D. Estrin, and M. Srivastava, "Guest Editors' Introduction: Overview of Sensor Networks," Computer, vol. 37, pp. 41-49, 2004.
[3] R. Szewczyk, A. Mainwaring, J. Polastre, J. Anderson and D. Culler, "An Analysis of a Large Scale Habitat Monitoring Application," Embedded Networked Sensor Systems, November 2004.
[6] Y. Liang and W. Peng, "Minimizing energy consumptions in wireless sensor networks via two-modal transmission," SIGCOMM Comput. Commun. Rev., vol. 40, pp. 12-18, 2010.
[7] V. Singla, R. Singla and S. Gupta, "Data compression modeling : Huffman and Arithmetic," International Journal of The Computer, vol. 16, pp. 64-48, December 2008.
[9] Z. Zhang and O. Berger, "Cluster based data query analysis and optimization for Wireless Sensor Networks," Advanced Communication Technology, February 2008.