在無線射頻辨識(RFID)系統中，讀取器(reader)係經由共享的無線通訊頻道藉由詢問電子標籤(tag)上的識別碼的方式來辨識電子標籤的存在。當有多個電子標籤同時回應識別碼時，將會產生訊號碰撞而降低了電子標籤辨識的效能。如何減低碰撞以提升辨識的效能就變成是一件重要的工作。有一些反碰撞(anti-collision)的協定被提出來以解決電子標籤碰撞的問題。他們可以被分類成兩大類：以阿囉哈(ALOHA)為基礎的協定以及樹狀(tree)為基礎的協定。其中樹狀基礎協定可以再分為確定性樹狀基礎以及機率性計數基礎兩個子類別。 在本論文中，我們提出了兩個屬於機率性計數基礎的協定(「適應性分群及預先信號」及「平行分裂」協定)以及一個確定性查詢樹狀基礎的反碰撞協定(「平行反向回應」協定)。其中，「適應性分群及預先信號」協定利用「適應性分群」方案來預估將要處理的計數器中電子標籤的數量，並予以適當分裂，期望可以達到每個電子標籤均有一個獨一的技術器值；併採用「預先信號」方案預先收集下一個計數器中電子標籤的數量為零、一或是多個，進而採取不同的處理程序，以簡省不需的迭代(iteration)數量。而「平行分裂」協定的概念則是讓所有尚未辨識的電子標籤一起分裂，而非只有計數器值為0的電子標籤進行分裂，以提升辨識效能；同時輔以「辨識樹高度調整」方案，使辨識樹的葉節點數與實際的電子標籤數接近，以縮短所需的迭代數量。至於「平行反向回應」協定則是使用「平行字首符合」方案，利用字首互補的方式增加符合字首辨識的電子標籤數量，並將符合條件的電子標籤透過「平行雙副載波回應」的設計方案提高回應成功率。我們亦對所提的三種協定進行分析以及模擬，並與相關的協定進行比較以了解所提協定的優點。In the RFID system, the reader identifies tags by interrogating their IDs through a shared wireless communication channel. Collisions occur when multiple tags transmit their IDs to the reader simultaneously, degrading the performance of tag identification. How to reduce tag collisions to speed up the identification is thus important. There are several anti-collision protocols proposed for dealing with tag collisions. They can be categorized into two classes: ALOHA-based protocols and tree-based protocols that include deterministic tree-based and probabilistic counter-based subclasses of protocols. In this dissertation, we propose two probabilistic counter-based (“Adaptive Splitting and Pre-Signaling” and “Parallel Splitting”) protocols and one deterministic query tree-based (“Parallel Reverse Response”) anti-collision protocol. 1) “Adaptive Splitting and Pre-Signaling” protocol uses the idea of “Adaptive Splitting” scheme to estimate the number K of oncoming identification tags and split them into K proper subgroups, besides “Pre-Signaling” scheme is proposed to reduce the number of messages sent between the reader and tags by detecting the number of tags (null, one or multiple tags) with counter=1 previously. 2) “Parallel Splitting” protocol uses the idea of “parallel splitting” scheme to split the unidentified tags concurrently instead of only the tags with counter = 0 as in the ISO 18000-6B standard. It also utilizes “adaptive identification-tree height adjustment” scheme to adjust the number of leaf nodes of identification tree to approach the number of tags. 3) “Parallel Reverse Response” protocol integrates the ideaof “parallel prefix matching” and “Parallel Two Sub-carriers Response” schemes to let the tags with prefix ID matching the request bit string S or complementary of S be arranged to respond in two sub-carriers simultaneously in order to speed up RFID tagidentification. We analyze and simulate these three proposed protocols and compare them with related protocols to demonstrate their advantages.
