On Using the Low-Duty Operation Mode for Interference Mitigation in 802.16 Femtocells

在超微型基地台（Femtocell）網路中，由於超微型基地台服務範圍小、數量多以及非規劃佈建之特性，與上層基地台的干擾處理是一個重要的議題。因此，在次世代IEEE 802.16m標準中，特別針對超微型基地台提出低負載模式（Low-Duty Operation Mode）處理干擾問題。當超微型基地台無任何服務中的使用者或是所有服務中的使用者都進入睡眠模式時，超微型基地台可以週期性地控制訊號傳輸模式以減少對周圍基地台的干擾。此「被動式」的低負載模式由於有其使用的侷限性，所以無法廣泛而有效率地處理干擾問題。在本論文中，我們提出「主動式」的低負載模式，結合資源分配機制來處理超微型基地台網路的干擾問題。我們建構一個超微型基地台網路之最佳化數學模型，能在最大化網路系統容量及使用者滿意度的目標下減少對上層基地台的干擾。為了解決此最佳化問題，我們將問題轉化成線性規劃，提出一最佳化演算法，並限制最大近似誤差以求得最佳解。利用此最佳化模型，我們首先分析在不同網路環境下最佳之低負載模式設計參數，包括其負載周期以及傳輸區間比例。其次，考慮到最佳化演算法需要搜集所有使用者資訊，複雜度過高，因此我們提出一個兩階段的演算法以減少計算複雜度。使用電腦模擬的結果顯示，此兩階段演算法在犧牲約10%的通道容量下，可減少90%以上的計算複雜度與執行時間，而能有效地解決超微型基地台網路的干擾問題。In two-tier networks, due to characteristics of short service range, great of number and ad-hoc deployment in femtocell, the processing of interference to the upper tier is a critical issue. In next-generation mobile system, IEEE 802.16m, the low-duty operation mode is proposed to reduce interference from neighboring femtocells. When the femtocell without any service user or all users operating in sleep mode, the femtocell can enter low-duty operation mode and periodically control transmission for reducing interference to neighbors. Due to the limitation of operation, this reactive low-duty operation mode can not mitigate interference widely and effectively. In this thesis, we propose a proactive low-duty operation mode, combined with resource allocation to manage interference in femtocell networks. Then we construct a optimization model, and it can maximize system capacity and reduce the interference to upper tier under satisfying user''s requirement. To resolve this optimization problem, we then reformulate it into linear programming problem with controlling approximation error. Using this optimization model, we first investigate in different network environments the optimal parameters of low-duty operation mode, including its duty cycle length and available interval ratio. Secondly, due to global information requirement and high complexity of optimization framework, we propose a two-stage algorithm to reduce computational complexity. The simulation results show that this two-stage algorithm around 10% at the expense of system capacity, it can reduce more than 90% computation complexity, and can effectively manage interference in femtocell networks