Wireless Indoor Flexible High Bitrate Modem Architecture

The simultaneous perturbation of an OFDM receiver by phase noise plus a residual frequency offset (due to synchronization errors) was modelled as combined phase impairment, and its effect has been evaluated analytically for the case of a frequency-selective fading channel. Several novel non-pilot-aided (decision-directed) schemes have been proposed, which compensate for the common (over all the sub-carriers) phase impairment effect. By representing the resulting Inter-Carrier Interference (ICI) as an uncorrelated, unequal-variance process in the frequency domain, Maximum Likelihood (ML) and approximate ML estimators of the complex-vector and phase-only type are derived and analytically evaluated. The present schemes are also compared with other current methods based on individual phase trackers, one per sub-carrier. The results will be disseminated in international journal papers throughout the current and the next year. IASA has also presented and evaluated two Post-FFT, decision-directed (i.e., non-pilot based) algorithms for joint channel equalization and phase noise and residual frequency offset compensation, in the case of slowly varying channels. The first algorithm is a Normalized Least-Mean-Squares tracker, which compensates for the total equivalent dynamic channel, comprised of the propagation channel plus the PHN and RFO effect. The second is a novel 2-Dimensional (2-D) algorithm, which compensates separately for the effect of the phase impairments (in the frequency axis) and the dynamic channel (in the time axis). Both of the algorithms have been evaluated through simulations. The 2-D algorithm is shown to be more efficient and more robust, thus, it has been selected for implementation within the WIND-FLEX project. The corresponding algorithm will be rotationally extended for use with other schemes as with Multi-Carrier Code Division Multiple Access.

Several post-FFT fine frame synchronization algorithms have been proposed foe use with OFDM systems. Two different approaches have been followed: the first is a channel-blind approach, which does not take into consideration the effect of the channel. The second is a channel-parametric approach, which is a novel in that it employs simple models in order to approximate the effect of the unknown channel on a set of post-FFT frame samples. The proposed algorithms have been evaluated and compared through simulations for different types of channels. It is shown that they can produce reliable estimates even for one dedicated OFDM symbol only. These algorithms will be useful for other related projects and for the research work of PhD and MSc candidates within IASA who focus on the same research topics.

IASA has proposed two practical algorithms for adaptively changing the transmission parameters on a turbo-Coded, Orthogonal Frequency Division Multiplexing (COFDM) communication link, in order to minimize the transmission power while satisfying constraints for a given Quality of Service (QoS) level. These algorithms use an approximation of the coded system performance in order to minimize the transmission power based on the proper selection of a set of the adaptable parameters (on a frame base). This set of adaptable parameters was the constellation size, the number of sub-carriers used for transmission, and the code rate. The selection of this set was done base on the WF system constraints. The idea of not using the weakest sub-carriers for transmission was called Weak Sub-Carrier Excision (WSCE). It is actually a "hard" (or on-off) bit loading scheme. The first proposed algorithm uses WSCE with a fixed percentage of excised carriers and the second one with variable. The significance of variable WSCE is the ability to choose for the same target rate the best code rate-WSCE percentage option. The gain in using these algorithms is twofold: First, the transmission power of the system is decreased, and second, the QoS is practically guaranteed per sample channel realization. These algorithms can be used with minor modifications to apply to a variety of systems that experience channels which are static long enough to necessitate the adaptation of the transmission parameters for a specific channel sample path, but are still fairly dynamic, such that classic bit-loading algorithms (which require high-throughput, error-free feedback channels to work well) would not be applicable. Future work focuses on finding closer approximations to coded performance, for a larger SNR range, when transmitting in the said environment Throughout this effort IASA has gained knowledge concerning modern adaptive modulation techniques and the designing process, at the algorithmic level, of reconfigurable transceivers. This knowledge will help our group for other related projects and for the research work of PhD and MSc candidates within IASA who focus on the same research topics.

IASA presented a VLSI architecture for minimizing the transmission power for coded or un-coded Orthogonal Frequency Division Multiplexed Systems. The hardware architecture estimated the power required for a given set of modem parameters, including the Bit Error Rate and the Signal to Noise Ratio, adapting them according the Quality of service requirements. The proposed architecture is based around a control unit exploiting the CORDIC algorithm, used for calculating exponentials in fixed-point representation. Further, the iterative bisection technique was applied to solve the power estimation problem. The architecture performs at wire-speed, uses minimal area and has shown the performance gain in an indoor wireless application. An implementation using Field Programmable Gated Array technology has validated the results.

Livrables

Partager cette page

Télécharger