Next-generation wireless wide-area network (WWAN) standards, such as IEEE 802.16 and 3GPP LTE-Advanced, are established for achieving the 4G standard requirements proposed by ITU-R [1]. One of the most challenging requirements in 4G standard is 1Gbps peak data rate for low-mobility user equipment (UEs) and 100 Mbps peak data rate for high-mobility UEs. In theory, the data rate requirement could be achieved by expanding the carrier bandwidth to 100 MHz. Nevertheless, a contiguous spectrum with such a wide bandwidth is rarely available in developed and developing countries since most feasible spectrum has been licensed to existing wireless techniques such as TV, GSM, or 3G. The unlicensed spectrum is mostly in a non-contiguous narrowband form. Given the carriers formed by the narrowband spectrum, next-generation WWAN standards are unlikely to satisfy the peak data rate requirements if each UE is served by a single carrier.
4.1.1 Carrier Aggregation
Carrier aggregation is introduced in LTE-Advanced for aggregating non-contiguous spectrum into a virtual carrier [53]. Multiple narrowband carriers can be aggregated into
a wideband virtual carrier for better spectrum utilization. UEs with carrier aggregation capability can increase their peak data rates by transmitting through the aggregated virtual carrier that virtually provides a larger transmission bandwidth. The carrier aggregation configuration can be UE-specific, which means that it provides a new dimension of UE configurations. UEs that require high throughput or have delay constraints can be assigned with a virtual aggregated carrier for high peak data rate, while others may be assigned with the traditional carrier for energy conservation. In summary, carrier aggregation can increase UE peak data rate, enhance spectrum utilization efficiency, and provide flexible configurations on a per-UE basis.
Beside the benefits, there are challenges in the design and implementation of carrier aggregation in LTE-Advanced system. New hardware and protocol designs, for instance, are required for all LTE-Advanced infrastructure and devices to support this new function.
Nevertheless, it should have minimal impacts on the existing and running LTE protocol and maintain the compatibility with the existing LTE-legacy devices. For configuration part, carrier activation on UEs is one of the key issues in carrier aggregation. Which car-rier(s) should be activated for each UE is a complex problem with concerns on the diverse carrier quality, coverage, spectrum efficiency, and UE-specific data requirements [54].
Using the system we illustrated in Fig. 4.1 as an example, LTE devices with different CA capability are located at different locations in the cell. The cell is offering two carriers C1 and C2, each with different coverage in the cell. In this example, only the CA-supported device within the coverage of C1 can burst its throughput using both C1 and C2. For others, the choice of activated carrier for those two devices outside of C1 is limited to C2 due to carrier reception. Additionally, the legacy device within the coverage of C1 supports only one carrier due to device capability. In addition, it may not be necessary to enable carrier aggregation for devices within C1 if they are in idle state or with low data rate requirement. The carrier activation problem is complex that it deserves further studies.
Additionally, in order to make the virtual carrier function properly, one of the activated carriers should be the primary carrier, which is responsible for not only the data
transmis-eNodeB C1 C2
legacy CA-supported Figure 4.1: An illustration of CA-enabled LTE Cell
sion but also the delivery of control signals, which is called cross-carrier scheduling [55].
By aggregating the control signal transmission into one or several carriers, the system may prevent unnecessary interference from neighbor cells. The choice of the primary carrier further increases the complexity of the carrier activation problem. It can be seen that this problem is a NP-hard problem. Thus, an optimal solution may not be derived in a reason-able time. A practical solution with acceptreason-able calculation time is desirreason-able and will be one of the main objectives of this work.
4.1.2 Truth-telling
Nevertheless, most of the existing works studying cellular systems make an assump-tion that all infrastructure and devices will faithfully follow the designed protocol. This assumption may not be valid in real world scenario [56]. For instance, some works assume that the evolved NodeB (eNB) knows the data requirements of UEs when she is calculating the final carrier and resource block allocations. However, the data requirements of UEs should be reported by UEs themselves since only they known the information. The re-porting process introduces a possibility of cheating for the UEs: A UE may report the data
requirement untruthfully if the resulting resource allocation, which is altered by her un-truthful report, increases her utility. This impairs the allocation efficiency since the eNB does not receive the correct information and the allocation is therefore manipulated by UEs under their selfish concerns. Therefore, a solution which functions properly when all devices faithfully follow the design may perform poorly when the devices (UEs) behave rationally. A detailed analysis on the rational behaviors of users in the system is neces-sary in order to understand and prevent this kind of malfunction. Game theory, which aims to analyze the rational behaviors and interactions of players under certain game rules, is suitable for this purpose.
In this work, we study the downlink carrier activation and resource block allocation in carrier aggregation capable LTE system. UEs are assumed to be rational and have different QoS requirements on the receiving data amount and delays. Their utilities are determined by the received data amount and the delay. The eNB is in charge of the carrier and re-source block allocation according to the carrier quality, coverage, and UE-specific data requirements, while its objective is two-fold: It aims to maximize the system efficiency, which is the total utility of UEs in the system, and prevent the rational UEs from reporting their information untruthfully. Our main contributions are as follows:
1. We address the heterogeneous characteristics of carrier quality, coverage, and UE QoS requirements in our proposed model and solutions. We make use of the carrier aggregation to enhance the system performance by satisfying the QoS requirements of UEs more efficiently. Specifically, we consider two type of UEs, throughput-sensitive and delay-throughput-sensitive UEs, in this work. The proposed solution effectively reduces the delay for delay-sensitive UEs while satisfying the throughput require-ments of throughput-sensitive UEs by properly triggering the carrier aggregation function of each UE accordingly.
2. To the best of our knowledge, this is the first attempt to propose and address the truth telling issue in carrier aggregation. We propose a truthful auction design specifically for the heterogeneous carrier quality and QoS requirements of UEs. We theoreti-cally prove that the proposed design indeed provides proper incentive for the UEs
to truthfully report their QoS requirements even if they selfishly behave.
3. The proposed solution is evaluated through extensive simulations in our LTE-Advances simulator, which uses the models and parameters suggested in 4G evaluation docu-ment [1] and 3GPP LTE standard [2]. The simulation results show that the proposed solution outperforms traditional implementations significantly in terms of social welfare, which represents the ability of the system satisfying the QoS requirements of UEs.