使用固定星座圖與沒有固定結構星座圖的碼搜尋已經被提出了，但沒有固定結構星座圖的碼搜尋，過去只有碼字長度等於2的兩個例子。因此，本論文針對碼字長度、碼字個數增加的非同調區塊碼，從給定的非同調距離去搜尋碼字，並且考慮沒有固定結構的碼，提出了幾種演算法來搜尋碼字組成碼，為的是讓碼具有大的最小非同調距離，模擬結果亦顯示此種沒有結構的區塊碼的錯誤效能比使用固定星座圖的區塊碼來得好。近來有篇論文提出使用查表法的差分編碼。根據此篇論文，另一種建立相差編碼表的方式被提出，稱為補碼字(codeword-added)演算法。補碼字演算法可簡單的建出星座點個數比群組個數多的相差編碼表，像是使用三十二點的正交振幅調變(32QAM)建立出16個群組的相差編碼表。本篇論文對補碼字演算法加了一個步驟，修改後的演算法所產生16個群組的三十二點的正交振幅調變(32QAM)相差編碼表有更好的錯誤率。除此之外，也把修改後的補碼字演算法用來建立區塊長度更長的相差編碼表。Searching codewords with certain and uncertain constellation has been proposed. In the case of using uncertain constellation, there were only two cases while codeword length is equal than 2. Therefore, we search noncoherent block codes for greater codeword length and numbers according to a given noncoherent distance with uncertain constellation. We propose several algorithms to find codewords for building up unstructured codes with longer noncoherent distance. The results show that the error performance of unstructured block code is better than the block code using certain constellation.Recently, a new paper of differential encoding by a look-up table was proposed. According that paper, there was another algorithm proposed as well, called codeword-added algorithm to build up a differential encoding table. Codeword-added algorithm can easily build up a differential encoding table when the numbers of constellation points are larger than the numbers of groups like the differential encoding table for 32QAM with sixteen groups. In this thesis, we add one step to codeword-added algorithm and construct a differential encoding table for 32QAM with sixteen groups as well. And its error performance is better than before. Besides, we also use the changed codeword-added algorithm to construct differential encoding tables which has longer block length.
