針對能量收集、低成本與緊急任務性、以及鄰近性物聯網情境之蜂巢式無線網路資源管理研究

Cellular Resource Management for Three Internet-of-Things Scenarios: Energy-Harvesting, Low-Cost & Mission-Critical, and Proximity-Based

施美如(Mei-Ju Shih)

指導教授 ： 魏宏宇

國立臺灣大學/電機資訊學院/電信工程學研究所/博士(2016年)

https://doi.org/10.6342/NTU201603714

摘要

在次世代無線通訊網路中，物聯網(又稱為機器對機器通訊或機器類型通訊) 已經吸引大量關注。更具體地說，有鑑於第五代蜂巢式網路的設計目標：超行動寬頻、超可靠機器類通訊和大量機器類通訊，物聯網在第五代蜂巢式網路扮演關鍵角色。另一方面，伴隨著資料流量的需求增加，以及無線資源缺乏，無線電資源管理深具挑戰卻又十分重要。由於物聯網根據不同服務會有不同需求，單一資源管理解決方案很難滿足所有物聯網情境。在本篇博士論文，我們著重於在蜂巢式網路下的物聯網資源管理之研究，並以三大物聯網特色為目標：具能量收集性的物聯網、低成本和緊急任務性共存的物聯網、以及具鄰近性的物聯網, 根據不同物聯網特色提出妥善資源分配方法以解決網路擁塞和減少資源碰撞的現象。

為了處理物聯網在蜂巢式網路下的資源管理，我們研究兩個基本的物聯網問題：網路擁塞和減少資源碰撞。這些問題起因於大量連網裝置。例如：在傳統以上行和下行鏈結為主的蜂巢式網路中，物聯網的資料流量以上行為主。因此，我們針對具能量收集性的物聯網、以及低成本和緊急任務性共存的物聯網，進一步探討網路接取程序和上行資料傳輸。近來，隨著邊緣鏈結通訊(又稱為直接通訊、裝置對裝置通訊) 興起，具鄰近性的物聯網將得以實現，裝置們可以彼此交換區域資訊而不需要透過基地台。因此，我們也探討在具鄰近性物聯網中的裝置對裝置通訊資源分配議題，以減少資源碰撞。值得一提的是，裝置對裝置通訊可以重複使用上行鏈結的資源。

針對物聯網之上行資源管理，我們首先考慮物聯網裝置具有能量收集性，代表這些裝置可以自行擷取環境中的能量以供自行運作而不需要人為介入。針對被安置在難以到達的物聯網裝置、以及為了節省成本，具能量收集性的物聯網情境變得愈來愈廣泛被應用。我們根據先進長期演進技術系統，針對上行傳輸程序建立實際的模擬平台，並進而鑽研網路接取程序。根據隨機接取通道資源如何被選擇及運作，我們提出具能量感知的推式、拉式及混合式設計方案，並和現行的長期演進技術系統比較。我們進而鑽研這三項設計方案的優缺點。以具能量收集性的物聯網來說，我們建議具能量感知性的設計是必備。在推式設計中，當裝置規模超過上限時，隨機接取通道資源碰撞、延遲和能量效率等表現開始變差。相對地，拉式設計總能達到最大吞吐量與能量效率，而代價是幾近線性成長的延遲和排程信令成本。然而，具有自適性網路存取和裝置數量估測的混合式設計具有更大的彈性，在不同資料流量和能量收集率的情境下，皆能達到最大吞吐量、能量效率、以及合宜的延遲。

接著，我們針對低成本物聯網、緊急任務性物聯網和人與人間通訊共存的情境下進行上行專屬資源分配的研究。我們透過裝置在意能量程度為裝置差異性建立模型，其中，低成本物聯網裝置具有較低能量感知，緊急任務型物聯網裝置具有較高能量感知。我們設計一套間接機制，其中包含以排隊時間為主的資源拍賣和直接存取資源的付費。透過貝氏史塔克伯格賽局，此具有優先性的資源分配框架得以被分析，並能達到唯一貝氏奈許均衡以及區域間並以排隊時間為基礎的誠實告知偏好之性質。在機制中，貝氏奈許誘因相容性、期中效益、期中個體理性以及弱收支平衡的性質得以成立。與長期演進技術標準相比，此架構除了引入基地台宣布資訊，並未增加額外的信令交換。更值得一提的是，針對緊急任務性物聯網而設計的付費資源池將可以作為營運商的財務補償。營運商可以根據機器對機器間通訊與人對人間通訊的流量和資源池分割，設計動態的價格。

最後一部分的研究是物聯網邊緣鏈結資源管理。我們根據第三代合作夥伴計畫的標準制定流程和限制進而設計裝置對裝置通訊的資源分配機制以實現具鄰近性的物聯網。根據沒有回饋訊息和半雙工傳輸模式的假設，我們設計排程指派和資料資源的運作機制，以期能在基地台失去功能時，能最小化資源碰撞。針對排程指派資源池與資料資源池間的對應關係，我們提出兩種強化性機制：顯式和隱式。這兩個機制採用資源感測結果，就資料碰撞率與吞吐量而言，表現優於第十二版裝置對裝置通訊之模式二資源分配。強化性顯式機制較近似於第十二版模式二，但能透過資源跳躍進而避免排程指派的碰撞。強化性隱式機制透過排程指派許可控制，進而能在短時間內達到零資料碰撞；然而，成功的鄰近性物聯網裝置數量受限於通道數目。這三種機制的比較結果得以提供車對車間通訊和加強型裝置對裝置間通訊作為研究基礎。

總結，本博士論文期能為客製化物聯網情境之上行鏈結與邊緣鏈結資源管理注入豐富研究能量，包含：針對能量收集物聯網的網路存取設計、針對低成本和緊急任務性物聯網的上行鏈結資源管理、以及針對鄰近性物聯網的邊緣鏈結自主性資源選擇。

關鍵字

物聯網 ； 機器對機器通訊 ； 裝置對裝置通訊 ； 資源分配 ； 能量收集 ； 隨機存取通道 ； 賽局理論

並列摘要

Internet-of-Things (IoT), also known as Machine-to-Machine (M2M) communications and Machine-Type Communication (MTC), has attracted much attention in the next-generation wireless communications network. Specifically, in view of 5G targets, i.e., extreme mobile broadband (eMBB), ultrareliable MTC (uMTC) and massive MTC (mMTC), “IoT” has played a key role in the evolution of the 5G cellular network. On the other hand, with increasing data traffic demand and deficient wireless resources, radio resource management is challenging and of great importance. Since IoT has varied requirements
according to needed services, a unitary solution of resource management may not be appropriate for all IoT scenarios. In this dissertation, we investigated IoT resource management in the cellular network, focusing on
three IoT features: energy-harvesting IoT, low-cost and mission-critical IoT, and proximity-based IoT.

To review IoT cellular resource management, we investigated two fundamental IoT problems: network congestion and collision minimization. Such problems are caused from massive numbers of devices. For example, in a traditional uplink/downlink cellular network, IoT data traffic is mostly uplink-oriented. Therefore, we first studied the network entry procedure and uplink data transmission for energy-harvesting IoT and low-cost and mission-critical IoT. Recently, with the advent of sidelink (direct link) communications, also known as Device-to-Device (D2D) communications, proximity-based IoT allows devices to exchange local information without bypassing the evolved NodeB (eNB). Thus, we also explore the D2D resource allocation for collision minimization problem in proximity-based IoT. It is worthy to note that D2D communications can reuse uplink resources.

For IoT uplink resource management, we first considered that IoT devices are energy-harvesting, which means they can harvest ambient energy on their own and operate without human intervention. This scenario becomes more popular for IoT devices deployed in places hard to access and for the purpose of cost saving. Based on Long-Term Evolution-Advanced (LTE-A) systems, we built a practical simulation platform for an uplink procedure and studied the network entry procedure. Compared to the naïve LTE scheme, energy-aware push-based, pull-based, and hybrid schemes are further proposed based upon how the Random Access CHannel (RACH) resources are selected and operated. Energy-aware mechanisms are recommended for energy-harvesting IoT. In a push-based scheme, performance such as RACH collision, delay and energy efficiency begins to degrade when the device scalability exceeds an upper bound. In contrast, a pull-based scheme achieves its maximal throughput and energy efficiency at the cost of almost linearly increasing delay and schedule signaling cost. Instead, the hybrid scheme, with adaptive network access and estimation of device number, is more flexible. Thus, the hybrid scheme can achieve energy efficiency, maximal throughput and adequate delay across different traffic loads and energy-harvesting rates. Afterwards, we studied the uplink dedicated resource allocation for the low-cost and mission-critical IoT co-existing with Human-to-Human (H2H) communications. We modeled the device heterogeneity by energy awareness levels, i.e., low-cost IoT devices with low energy awareness, and mission-critical IoT devices with high energy awareness. An indirect mechanism was designed with the waiting-time auction and direct access price. This prioritized resource allocation framework was analyzed by the Bayesian Stackelberg game, and achieved unique Bayesian Nash equilibrium, and interregional and waiting-time-based truth-telling. Bayesian Nash incentive compatibility, interim efficiency, interim individual rationality, and weakly budget balance are maintained with this mechanism. Except for the eNB’s announcement, the framework does not introduce extra signaling exchange compared to the LTE standard. What’s more, the design of a price-based resource pool for mission-critical IoT compensates the operator financially. An operator can design dynamic prices according to M2M/H2H traffic loads and resource pool partitions.

Finally, for IoT sidelink resource management, we designed D2D resource allocation schemes following the standardization progress and constraints in 3rd Generation Partnership Project (3GPP) to realize the proximity-based IoT scenario. Based on an assumption of no feedback and half-duplex mode, we designed the operation of Scheduling Assignment (SA) and data resources, in order to achieve collision minimization when the eNB is out of function. According to whether the SA pool is mapped to the data resource pool, we propose two enhanced schemes: explicit and implicit. These approaches, utilizing the sensing results, outperform Rel-12 Mode 2 in terms of data collision probability and throughput. The enhanced explicit approach, similar to Rel-12 Mode 2, avoids SA collisions by resource hopping. The implicit approach can reach data collision free in short iterations via SA admission control, while the amount of successful proximity-based IoT devices is restricted by the channel number. A comparison of the three schemes are provided, which serves as the foundation for the future standard development such as Vehicle-to-Vehicle (V2V) communications and D2D enhancement.

To sum up, this dissertation provides abundant research on uplink and sidelink resource allocation for customized IoT scenarios: a network entry design for energy-harvesting IoT, an uplink resource management for low-cost and mission-critical IoT, and a sidelink autonomous resource selection for proximity-based IoT.

並列關鍵字

IoT ； M2M ； D2D ； resource allocation ； energy-harvesting ； RACH ； game theory

參考文獻

[2] P.-H. Lee, M.-J. Shih, G.-Y. Lin, and H.-Y. Wei. Optimal Resource Reservations to Provide Quality-of-Service Guarantee in M2M Communications. In 8th Int. Wireless Commun. and Mobile Comput. Conf. (IWCMC), pages 769–773, Aug. 2012.

the Grid Integration of Renewable Energy Sources: A Survey. IEEE Trans. Ind. Electron., 53(4):1002–1016, Jun. 2006.

[6] A. Osseiran et al. Scenarios for 5G Mobile and Wireless Communications: the Vision of the METIS Project. IEEE Commun. Mag., 52(5):26–35, May 2014.

[7] M.-Y. Cheng, G.-Y. Lin, H.-Y. Wei, and AC.-C. Hsu. Overload Control for Machine-Type-Communications in LTE-Advanced System. IEEE Commun. Mag., 50(6):38–45, Jun. 2012.

[12] 3GPP TR 36.843 v1.2.0. Study of LTE Device to Device Proximity Services; Radio Aspects. Mar. 2014.