Quantized feedback scheduling for MIMO-OFDM broadcast networks with subcarrier clustering

Abstract

Feedback of user channel quality information (CQI) to base station (BS) is essential for efficient scheduling of multi-user (MU) multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) broadcast networks. This feedback overload increases linearly with the cardinality of users, transmit antennas and subcarriers present in MIMO-OFDM systems. Hence, clustering of adjacent subcarriers has been considered in literature for reduction in feedback load. Here users communicate the CQI of the center subcarrier of the clusters assuming that subcarriers of a cluster experience similar channel conditions. However, this method leads to inefficient resource scheduling for higher cluster sizes as it does not take advantage of CQI of other subcarriers that experience non-uniform channel conditions. This reason motivated us to develop a new limited feedback scheduling scheme where quantization of CQI with multiple bits is implemented along with clustering of adjacent subcarriers. In this scheme, the multi-bit quantized CQI of all the subcarriers of a cluster are examined for the efficient scheduling along with further reduction in the feedback overhead. Moreover, we proposed a less time consuming method to find the optimum quantization thresholds that play significant role in multi-bit quantization process by using genetic algorithm (GA). This GA methodology is computationally more efficient than exhaustive search process.

Introduction

Multiple-input multiple-output (MIMO) systems are known for providing delay tolerant high data rate communications with good quality of service (QoS) [1], [2], [3]. This is achieved by performing special kind of multiplexing, and diversity by exploiting the spatial dimension of multi antenna systems. Moreover, a new kind of selection diversity gain known as multiuser diversity is achieved in multiuser environment [4]. This multiuser diversity arises due to the independent fading paths exist between different users and the base station (BS).

Orthogonal frequency division multiplexing (OFDM) is a favourable multiplexing technique for frequency selective fading channels as it decomposes the frequency selective channel into a set of frequency flat fading channels. Hence, each frequency flat fading channel can be easily modulated independently from the others [5]. This makes OFDM a robust technique. Present and next generation wireless standards like 3GPP-LTE (Third Generation Partnership Project Long Term Evolution), WiMAX (Worldwide interoperability of Microwave Access) (IEEE 802.16e) etc. have integrated both MIMO and OFDM to take the advantages of both these techniques [6]. Hence, MIMO-OFDM has become an essential technique for multiuser communications with multiple carriers and multiple antennas.

Dirty paper coding (DPC) [7] is the optimal scheme for MIMO and MIMO-OFDM broadcast networks to achieve the data rate according to Shanon’s capacity limit [2]. However, as the complexity of implementation of DPC is very high, less complex scheduling schemes are discussed in literature for both MIMO and MIMO-OFDM broadcast networks. Authors in [8] discussed that DPC achieves the maximum capacity region of Gaussian MIMO broadcast network by serving multiple users (as many transmit antennas installed at BS) simultaneously. Users send their channel quality information (CQI) to BS for supporting BS in scheduling multiple users simultaneously. However, this feedback from users occupies a reasonable chunk of the uplink spectrum. Hence, limited feedback schemes are discussed in literature [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19] to reduce the feedback overhead in MIMO broadcast networks. Authors in [20], [21] have proposed ideas of applying soft computing techniques for scheduling resources in wireless communications.

Similarly, limited feedback scheduling schemes are also discussed in literature for MIMO-OFDM broadcast network. Svedman et al. discussed some opportunistic limited feedback scheduling schemes and beamforming schemes for OFDM/OFDMA systems in [22], [23]. Some related works are also discussed in [24], [25], [26]. In [27] authors proposed a method of dividing the available frequency and time resources of MU MIMO-OFDM systems into tiles or chunks for reduction of complexity in signal processing and feedback overhead. A specific number of adjacent subcarriers are grouped into chunks in the frequency domain. In time domain, specific number of adjacent symbols are grouped into tiles. Same kind of signal processing is done for each subcarrier present in the chunk. This effort reduces the signal processing and overhead in feedback. Similar work is done by authors in [28] where clusters of adjacent subcarriers are made by assuming adjacent subcarriers experience similar channel conditions for achieving a feedback load reduction. In [28], authors discussed that each user sends only the maximum CSI and the corresponding transmit antenna index to BS of the center subcarriers of the clusters for user scheduling. This scheme is not efficient in scheduling the best user and the receive antennas because this scheme is not able to take advantage of CSI of other subcarriers present in the clusters. Hence, in this paper we proposed an efficient scheduling scheme where the CSI of all the subcarriers present in a cluster are processed by the BS for scheduling the best users and the receive antennas. Multi-bit quantization process is proposed in [29] for MIMO broadcast networks for achieving reduction in feedback overhead. This work motivated us to evaluate the effect of multi-bit quantization process with optimal quantization thresholds for MIMO-OFDM broadcast networks. According to the proposed scheme each user feeds the multi-bit quantized CSI to the BS, which further reduces the feedback overhead of the MIMO-OFDM system.

Important results obtained in this paper are mentioned as:

• A new scheduling algorithm with random beamforming and employing spatial multiplexing at the BS is proposed for efficient scheduling of MU MIMO-OFDM broadcast network by quantized feedback of CQI with multiple bits.

• The proposed scheduling algorithm ensures that all antennas at BS gets utilized in communications by assigning a user randomly to all transmit antennas which remain idle. This enhances the achievable system throughput by the proposed scheduling algorithm.

• This scheme achieves better system throughput than [28] and [19] with incurring reduced feedback overhead.

• A framework using GA is given for finding out the optimum quantization thresholds for quantization process with any number of bits.

• We also showed that for large number of users ( ≈ 1000) 4-bit quantization with fixed optimum quantization threshold is sufficient to have optimum system throughput unlike 1-bit quantized feedback scheduling [14] where adaptive quantization threshold is required (different quantization thresholds for different number of users).

• Further, we discussed that the optimal quantization thresholds for 4-bit quantization process depend on the average system signal-to-noise ratio (SNR), the number of transmit antennas, and the cluster size.

The organization of rest of the paper is as follows. In Section 2 the system model of the multi-user MIMO-OFDM broadcast network is discussed. A new scheduling algorithm using quantization of CQI with multiple bits is proposed for efficient scheduling of MU MIMO-OFDM broadcast network in Section 3. The quantization process and finding the optimum quantization thresholds using GA are discussed in Section 4. In Section 5, various simulation results are presented. The reduction achieved in computational complexity by the proposed GA technique is discussed in this section. Moreover, the feedback overhead comparison of various scheduling schemes is also presented in this section. The concluding remarks of the paper is discussed in Section 6.