Study of Certain Subband-Based Adaptive Modulation Schemes in an Ofdm System
L. J. Shi,L. Rong,R. F. Lin,Y. Du
DOI: https://doi.org/10.2529/piers061113041022
2007-01-01
Abstract:In applying adaptive schemes to an OFDM system to adaptively adjust the trans- mission format based on channel state information, a practical scheme would group the subcarriers into similar-quality subbands to reduce the associated modem mode-signaling requirements. One way to determine the modulation mode for a speciflc subband is based on the average signal- to-noise ratio (SNR) of the subband. To flnd the optimal switching levels, knowledge of the probability density function of this average SNR is required. Another popular way to determine the modulation mode is based on the SNR of the worst subcarrier of the suband, which may rep- resent an overconservative choice of order statistic. This approach can be extended to deploying other statistic as well, for example, the second or the third worst statistic. Analysis of the above approaches is the focus of this study, and performances of these schemes are compared through numerical simulations. DOI: 10.2529/PIERS061113041022 For an OFDM system in wireless communication, due to multipath fading, subcarriers experi- ence signal-to-noise ratio (SNR) degradation to varying degree. Adaptive modulation schemes are efiective to mitigate such detrimental efiect, where in principle the transmitted power and bit rate of each subcarrier are allowed to change dynamically according to channel variations. Yet such a full treatment implies high complexity in modem hardware and high signalling overhead over the feedback channel from the receiver to the transmitter, a scheme that may not be plausible in high-speed wireless data applications. Such concern leads to some suboptimal schemes. Grunheid etal. (1) and Keller and Hanzo (2) proposed to group the subcarriers into similar-quality subbands with single constellation being used for the subcarriers within a subband. Dardari (4) proposed an ordered subcarrier selection algorithm where among the available N subcarriers only the K sub- carriers (KN) with the highest SNRs are used, to which uniform bit and power level allocation are applied. For a mode-switching-assisted adaptive modulation scheme, the bit-error-rate (BER) and through- put performances are determined by the set of switching levels. Chung and Goldsmith (4) and Choi and Hanzo (5) analyzed for single carrier system the constant power variate rate adaptive mod- ulation scheme. The optimal switching levels that maximize the average bits-per-symbol (BPS) throuput for a target average BER are found to depend on the channel quality statistics. In ex- tending this technique to multicarrier system where subband based adaptive modulation scheme is adopted, Keller and Hanzo (2) proposed to use the subcarrier experiencing the worst SNR. This approach may appear somehow conservative, which motivates us to analyze the problem using other order statistic of the subcarrier SNRs, such as the second or the third order worst statistic. Another plausible approach is to use the average SNR of all the subcarriers within a subband, where the optimal switching levels can be determined by using the probability density function (pdf) of the average SNR in conjunction with Lagrangian optimization technique. These approaches form the body of the current study. In the order statistic approach, the subcarriers are usually assumed to be independently and identically distributed (i.i.d) Rayleigh fading channels for mathematical convenience (3), although they are in general correlated. Consider a subband containing P subcarriers, where the pdf of the SNR ∞ of each subcarrier is f∞ (∞) = 1 ∞ e