在LTE/LTE-A上行網路下的一個跨層睡眠排程方法

A PHY-MAC Cross-Layer Approach for Sleep Scheduling in LTE/LTE-A Uplink

無線行動裝置支援行動性，使用電池供給裝置電力但只能維持裝置有限時間的運作；以3G手機為例，根實際測試，手機運作時以無線通訊介面損耗的電力最為可觀，佔全部耗電比例的40%（處理器為20%，顯示螢幕則為13%），而最新的4G無線存取技術採用高效率的力OFDMA(Orthogonal Frequency-Division Multiple Access)搭配MIMO(Multi-input Multi-output)增進傳輸速率，無線網路介面的耗電因此更多。在本論文中，我們討論了在3GPP LTE/LTE-A(3rd Generation Partnership Project Long Term Evolution/Long Term Evolution-Advanced)的無線網路下，睡眠調度的跨層設計。設計同時考慮了媒體存取制(Medium Access Control; MAC)層的多用戶睡眠調度（藉著DRX/DTX(Discontinuous Reception/Discontinuous Transmission)參數的設置）和實體層功率和無線電資源分配。本論文提出的方法還同時考慮了延遲時間、封包遺失率和最大發射功率的限制下，使行動裝置能夠獲得更長的電池續航時間和提升無線資源的使用率。本研究的設計透過模擬進行比較與驗證，可以發現對省電效能有很大的改善。

To support mobility, wireless mobile devices are powered by batteries; however, a battery can only store a limited amount of energy. Take a 3G handheld device for example, according to the real test, the wireless interface consumes the largest proportion of the total amount of power (up to 40%). In particular, the 4G wireless communications adopt OFDMA and MIMO as the wireless access technology, where the former improves the bandwidth efficiency and the latter enhances throughput. In this case, the wireless interface in a mobile device will consume much more power than that for 3G devices. To save power, we have designed a PHY-MAC cross-layer sleep scheduling method which considers power saving, energy efficiency, and QoS at the same time over the 4G LTE/LTE-A wireless networks. Several papers have been published or submitted and three master theses have been accomplished. In this paper, we discuss a cross-layer design of sleep scheduling in 3GPP LTE/LTE-A wireless networks. The design jointly considers the Medium Access Control (MAC) layer multi-user sleep scheduling (via the settings of DRX/DTX parameters) and the physical layer power and radio resource allocations. The new approach enables UEs to gain longer battery life time and better utility of radio resource while guarantee the delay, packet drop rate, and maximum transmit power constraints. The design is validated through comparisons with simulation results and we find significant improvement on power saving.