本論文提出基於無線射頻（RFID）技術，利用其對於標籤（Tag）訊號自動追蹤之特性，並針對使用於此設備應用於人機介面（Human Computer Interface）的設計上所遇到的環境障礙的干擾、訊號辨識複雜度、事件觸發點的前置訓練與動態偵測並且即時調整感知訊號的範圍來進行系統設計上的考量。因此，若要以此設備作為人機互動的主要工具，所應該先解決的問題是：標籤的訊號強度（Received Signal Strength Indication）會因為環境不同或使用者使用的方式不一樣，導致系統在即時產生互動事件的正確率會降低。然而，透過我們設計出的架構與機制，配合一個音樂互動式系統，藉由使用前的訊號訓練等前置處理程序，達到與電腦系統進行互動等觸發音樂播放或是特效的產生。最後，分析系統執行後而產生的資料後，的確發現本論文提出的方法會明顯提升系統成功執行事件觸發的次數和辨識正確率。 This paper proposes a human computer interface which is based on Radio frequency identification (RFID) technique to let human communicate with computer by analyzing the signal from tags. For retrieving those signals from tags, how to decreasing the noises created by surrounding environment and detecting useful information from variant signals are the most important. In our proposed method, we adopt a train procedure as pre-processing to categorize all of signals into two categories: noise and real data. After the real data is retrieved, we use “music director” as an application to let user play with computer. The experimental results show the proposed method can analyze the signals from tags successfully with high detection rate.
