有限誤差壓縮在閘道器上之應用

Abstract

隨著物聯網(Internet-of-Things, IoT)的應用日漸興起，與其相關的無線感測網路(Wireless Sensor Networks, WSN)使用研究也日益顯得重要。由於物聯網之節點大多配置在無線環境，且不易供電，如何降低節點耗電量來延長生命週期便成為一個很重要的研究議題，而物聯網節點主要之行為可以分為資料感測、資料運算及資料傳輸，其中又以資料傳輸最為耗電，也因此成為主要的研究目標。本研究以有限誤差壓縮的方式，透過容許一定量的誤差，將近似的數值視為相同，進而使得傳輸量減少，最終達到壓縮的目的。此外我們也以紡織機台數據為資料集，在實驗中比較了Bounded-Error Run-Length Encoding (BERLE)與Bounded-Error Huffman Coding (BEHC)兩種演算法在邊界運算(Edge Computing)閘道器上的實際壓縮效果。

With the application of internet-of-things (IoT) has become more popular, research about using wireless sensor networks (WSN) has also become more important. Due to the nodes in IoT usually set up in the wireless environment, difficult to supply power, how to reduce consumption to extend the lifetime becomes a significant issue. The main behavior of the nodes in IoT are data sensing, data operation, data transmission. Among above, data transmission always takes the largest part of sensors’ power consumption. Thus, data transmission becomes on main study. In this thesis, we use a way of bounded-error compression by allowing certain amount of errors, regarding similar value as the same. Thus reduce the amount of data transmission and finally achieve the compression. In addition, we also use the textile machine data set to compare the actual compression on edge computing gateway with Bounded-Error Run-Length Encoding (BERLE) and Bounded-Error Huffman Coding (BEHC) in the experiment.