Noncoherent eigenbeamforming for a wideband cellular uplink

Abstract

We investigate wideband space-time communication on the uplink of an outdoor cellular system, in which the base station is equipped with N antennas and the mobile has a single antenna. We assume noncoherent reception at the base station, which incurs significantly less overhead than pilot-based estimation of the space-time channel from each mobile to the base station. Noncoherent communication techniques are particularly well suited to outdoor cellular systems for which channel time variations are significant due to mobility at vehicular speeds. As is common in outdoor cellular systems, we assume little or no scattering around the base station, so that, from the viewpoint of the base station antenna array, the incoming signal from a given mobile has a narrow power angle profile. Thus, the spatial channel covariance matrix is typically highly colored, having one or two dominant eigenmodes. We demonstrate that large beamforming gains are achievable without explicit estimation of the space-time channel, while only processing a relatively few number of dominant eigenmodes.Consider the received signal Y f at frequency bin f in an OFDM system (1). The N × 1 space-time channel H f for each subcarrier is well-modeled as identically distributed complex Gaussian random vectors that decorrelate across frequency [1].Thus, the spatial covariance matrix C = E[H f H † f ] can be obtained by averaging over subcarriers, without requiring any pilot overhead. A spectral decomposition of the channel co-variance matrix yields (2), where the eigenvector matrix U = [U1 . . . UN ] is unitary, and Λ is diagonal with eigenvalues {λ l } arranged in decreasing order. The eigenvalue λ l represents the strength of the channel on l th eigenmode U l . We propose an eigenbeamforming receiver that projects the received signal in each subcarrier along the L (typically much smaller than the number of receive elements N ) dominant eigenmodes of the estimated channel covariance matrix, {U l } L l=1 . Beamforming along the dominant eigenmodes of the channel creates parallel, independently fading channels for the same transmitted data. For each of the L eigenmodes, we employ noncoherent coded modulation strategies with turbo-like joint data and channel estimation, as in prior work on single antenna channels [2,3,4]. Thus, by appropriately leveraging the covariance estimate available in a wideband system, the noncoherent eigenbeamforming receiver provides many of the benefits of explicit space-time channel estimation without incurring its overhead. For example, beamforming gains in received SNR (relative to a single antenna system) are realized, 1 This work was supported by Motorola and the University of California Cooperative Research Program under a Communications Research (CoRe) grant.while incurring reasonable complexity by using a small number of dominant eigenmodes for demodulation and decoding.In addition to an SNR advantage from scaling up the number of receive elements, diversity gains are provided by the L parallel, independent fading channels...

Noncoherent eigenbeamforming and interference suppression for outdoor OFDM systems

“…NONCOHERENT EIGENBEAMFORMING Accurate estimates of the spatial covariance matrix, available from averaging in wideband systems, allow eigenbeam-forming at the receiver with the dominant spatial modes of the channel [13]. Much of the Multi-Input Multi-Output (MIMO) communication literature assumes channel coherence (see [16] and the references therein) and classical receive beamforming is based on explicit channel estimations [7].…”

“…Appendix A details the calculation for the parallel block fading channel model. Results for idealized channels, where the total power is distributed evenly amongst the top L spatial modes, are provided in [13]. The capacity notations in Figure 4 correspond to the assumptions in [13].…”

“…Results for idealized channels, where the total power is distributed evenly amongst the top L spatial modes, are provided in [13]. The capacity notations in Figure 4 correspond to the assumptions in [13]. Figure 3(a) provides capacity results for realistic fading channels, where the top L modes are used from spatial covariance matrices generated by Laplacian PAPs and a 10-element linear array with halfwavelength spacing.…”

“…Nevertheless, the experimental results are encouraging, since the focus is on outdoor channels with relatively few dominant modes. Further, there is a larger potential diversity gain in capacity for the two mode channel than there is by increasing the number of modes beyond two [13]. Finally, as the number of clusters grows in an outdoor environment, they are more likely to be attenuated Performance of noncoherent eigenbeamforming receiver at 1.35 bits/sec/Hz with idealized (where the received energy is divided equally among the given number of dominant spatial modes) parallel block fading channels, T = 10.…”

“…Here we consider spatially correlated channels, employing block fading to model time-frequency coherence. In [13], the capacity with QPSK signals of L parallel block fading channels is quantified. Such computations reveal a modest gain of up to 2 dB at midto-high Signal-to-Noise Ratio (SNR) for channels with equal strength dominant spatial modes.…”