發展無線感測網路防碰撞機制進行影像傳輸

Abstract

隨著無線傳輸技術的進步與發展,無線感測網路的應用範圍越來越廣，也因此吸引了許多研究團隊的高度興趣。無線感測網路在應用上，主要都集中在低資料量的資料傳輸。然而，高資料量資料像是影像可以提供更豐富的資訊。因此，如何設計一個有效的無線感測影像平台成為近年來相當熱門的研究議題。回顧過去的研究，大多都致力於開發新的影像處理技術，像是不同的壓縮演算法或辨識演算法來降低影像資料對於無線網路傳輸上的負載，而鮮少有針對改善網路傳輸的品質這方面的研究,而且大多都集中在討論點對點傳輸，而非多點間的傳輸架構，因為在多點傳輸中會容易造成嚴重的封包碰撞問題。本研究對上述問題提出解決方案，實做了一款以ZigBee為主並含有時間排程機制的多點無線感測影像傳輸平台。 本研究使用機率與期望值來模擬與分析ZigBee網路中的有效資料傳輸量,與傳輸時間，進而討論出最適合時間排成機制 (TTS) 的一個時槽的時間，接著分析自由競爭傳輸機制 (CSMA/CA) 與時間排成機制 (TTS) 之間的效能與傳輸成功率的分析，並且探討在時間排成機制下不同影像大小的傳輸效能。最後為了要使的無線影像傳輸平台有最佳化的傳輸效果，因此跟著探討適合影像傳輸的資料封包長度為何還有最適合影像傳輸的網路拓撲為何。 在實測結果與分析的部份，除了與模擬的結果互相對照比較之外，還分析二者差異並且在結果分析的部份做說明討論，本研究保留了自由競爭機制的重傳的優點更合併了時間排成機制，並透過實際發展一套系統平台來做大資料量的傳輸驗證，例如影像資料，除此之外為了使系統有更好的傳輸效能，也對系統網路資料傳輸與網路拓譜型態做了最佳化的動作，而無論是模擬或是實測的結果都證實了這套系統的可行性與優異性。未來除了影像資料傳輸外也可以應用在音訊或是其他大資料的資料傳輸上。

In recent years, driven by microelectronic technology improvement, network protocols development and wireless communications, the applications of the wireless sensor network (WSN) are more widespread and have attracted the attention and interest of research communities. The applications of the WSN are focused on simple scalar data. Large quantities of data such as image data provide very rich information about the environment. Hence, how to design an effective image sensor network system is one of the quite popular discussion topics. In related works, most of these efforts only focused on improvement of the image processing such as image encoding method and object identification to decrease transmission efforts but little discusses about their development of sensor network. Most of these works are in node-to-node topology, nor multi-node topology. Because of the packet collision problem is happened seriously in multi-node topology. In this study, a ZigBee image sensor network (ISN) system with transmission time scheduling (TTS) is proposed and utilizing the proposed platform to develop a remote surveillance system. In result, this study utilizes simulations and experiments to analysis the effective data capability, TTS time slot, the suitable data payload length, the suitable network topology for image transmission, and comparing performance of the ISN system with TTS and without TTS. In conclusion, the proposed system offers lower packet collision and lower network traffic than the ISN system without transmission time scheduling protocol. In the future, in spite of image data, the proposed platform may work in other applications such as audio data or other large quantities of data.