MIMO in 3G Cellular Systems: Challenges and Future Directions

For 3G cellular systems to compete in the mobile data market with emerging technologies like WiMax/802.16 in the medium to long-term, multi-antenna transmission and reception (known as MIMO) will be required to achieve the requisite high data rates. One of the greatest challenges facing MIMO in the context of 3G is that present MIMO systems do not cope gracefully with high levels of interference. Since any well-designed cellular system is by nature interference-limited, this poses a fundamental conflict: increasing the spectral efficiency with MIMO appears to require reducing the interference level, which traditionally requires increased frequency reuse or other spectral efficiency reducing measures. In this paper, we overview and compare recent approaches for multicell MIMO and explain their shortcomings for 3G systems, or indeed any heavily-loaded cellular network. We then introduce two simple and complementary interference-reducing techniques that may prove effective and practical for adding MIMO to 3G cellular, and other future wireless broadband systems. The authors are with the Wireless Networking and Communications Group, Department of Electrical and Computer Engineering, The University of Texas at Austin, 1 University Station C0803, Austin, TX 78712, USA. Email:{jgandrews, wchoi, rheath}@mail.utexas.edu. Phone:1-512-471-0536. This work was supported in part by SOLiD technologies, Korea, and Freescale Semiconductor. April 29, 2005 DRAFT SUBMISSION FOR IEEE COMMUNICATIONS MAGAZINE 2 I. THE PROMISE OF MIMO The idea of using multiple receive and multiple transmit antennas has emerged as one of the most significant technical breakthroughs in modern wireless communications. Theoretical studies and initial prototyping of these multiple-input multiple-output (MIMO) systems have shown order of magnitude spectral efficiency improvements in point-to-point communication. As a result, MIMO is considered a key technology for improving the throughput of future wireless broadband data systems, which presently are mired at data rates far below their wired counterparts. The multidimensional MIMO channel can be exploited to increase the diversity of the system or to provide parallel spatial channels, which is known as spatial multiplexing. Diversity is generally considered lower risk, and a well-known example is space-time codes [1], [2], which have found adoption in 3G CDMA cellular systems [3]. Diversity increases the robustness of the system by eliminating fades, and also raises the average received signal to noise ratio (SNR). Since the SNR increases linearly with the diversity order, the capacity growth is logarithmic, easily verified with Shannon’s formula C = B log2(1 + SNR). On the other hand, spatial multiplexing divides the incoming data into multiple parallel substreams and transmits each on a different antenna. If successfully decoded, it is logical that this increases the capacity linearly with the number of transmit antennas (or communication dimensions), which indeed has been proven using information theory [4]. Spatial multiplexing is thus more exciting than spatial diversity from a high-data rate point of view, but due to the fading in the wireless channel, some diversity is generally needed in a practical system to get an acceptable SNR and hence error probability. Therefore, a future high throughput system is likely to use some of the available dimensions for spatial multiplexing, and some for diversity. In this paper, we focus primarily on the spatial multiplexing aspects of MIMO as they are more exciting as far as increasing capacity and more challenging. Since high data rates are particularly April 29, 2005 DRAFT SUBMISSION FOR IEEE COMMUNICATIONS MAGAZINE 3 interesting for the downlink, it can be assumed that the number of transmit antennas Mt will be larger than the number of receive antennas Mr, due to extreme space and cost restrictions on the mobile unit. It is reasonable to expect that Mr of the transmit antennas will be used for spatial multiplexing, while the remaining Mt − Mr can be used for transmit diversity, using antenna subset selection for example [5]. II. MIMO IN 3G CELLULAR SYSTEMS Despite all the excitement surrounding MIMO’s promised capacity increases, it is quite a different matter to apply it successfully to a commercial cellular system. The vast majority of academic and even industrial research has focused on the point-to-point model (ignoring nearby competing interference sources), which is well-summarized in [6]. Nevertheless, MIMO’s enormous data rate incentives have motivated much interest from 3G cellular providers and equipment manufacturers, and MIMO is being widely considered for cdma2000 (3GPP2) and WCDMA (3GPP), particularly for the high-data rate modes such as EV-DV, EV-DO, and HSDPA [7]. For example, for 3GPP a combination of V-BLAST and spreading code reuse has been considered and the hypothetical peak data rates with this combination are given in Table I [8]. Table I is a typically naive interpretation of how MIMO can be applied to an interferencelimited cellular system. All well-designed cellular systems are by nature interference-limited: if they were not, it would be possible to increase the spectral efficiency by lowering the frequency reuse or increasing the average loading per cell. In the downlink of a cellular system, where MIMO is expected to be the most profitable and viable, there will be an effective number of NMt interfering signals, if the number of non-negligible neighboring base stations is N . Figure 1 illustrates the impact of other-cell interference in cellular MIMO systems. III. THE OCI PROBLEM IN CELLULAR MIMO COMMUNICATIONS There have been a few notable preliminary studies on interference-limited MIMO networks [9], [10] with the general conclusion that other-cell interference severely degrades the overall April 29, 2005 DRAFT SUBMISSION FOR IEEE COMMUNICATIONS MAGAZINE 4 capacity of spatial multiplexing MIMO systems. MIMO receivers are able to decode the parallel data streams by suppressing the spatial interference between the signals sent from the Mt adjacent transmit antennas. The root cause of the degradation is that the MIMO receiver has to focus on suppressing the spatial interference introduced by the multiantenna transmitter; consequently, there are not sufficient degrees of freedom available to also suppress the co-channel interference, be it from other users in the same cell or from other cells (other-cell interference). Recently, a closed-form expression for interference-limited MIMO cellular system performance was developed, specifically for the outage probability and outage capacity of a MIMO-CDMA system with linear receivers [11]. Although one of the primary advantages of CDMA is its resilience to co-channel interference, this paper demonstrated that OCI can severely degrade the overall capacity of spatial multiplexing even in CDMA systems. Figure 2 shows this disappointing result, which assumes that the spatial interference is completely suppressed by a zero-forcing linear receiver in order to separate out of the different spatial streams. In particular, it can be seen that a conventional 1 × 1 cellular CDMA system in fact performs similarly to an 8 × 8 system with the same total rate (and hence 8 times the spreading factor). Although a suboptimal Minimum Mean Square Error (MMSE) receiver can reduce the noise enhancement by allowing some residual spatial interference, the overall improvement is incremental. While the problem of other-cell interference has existed in cellular systems for many years, its effect on MIMO systems is far more severe, and also more intriguing. Traditional OCI-reduction techniques have been quite limited, such as sectoring cells using simple directional antennas, or increasing the frequency reuse distance. In the rest of the paper, we discuss more sophisticated recent strategies that have been proposed for interference-limited multicell MIMO systems, and consider their promise for 3G wireless networks. April 29, 2005 DRAFT SUBMISSION FOR IEEE COMMUNICATIONS MAGAZINE 5 IV. OCI MITIGATION TECHNIQUES FOR 3G CELLULAR MIMO SYSTEMS In this section, we outline possible OCI mitigation techniques in 3G cellular MIMO systems and discuss their feasibility.