Ali Ozgur Yilmaz

A General Framework for Optimum Iterative Blockwise Equalization of Single Carrier MIMO Systems and Asymptotic Performance Analysis

The paper proposes a general framework for both time-domain (TD) and frequency-domain (FD) iterative blockwise equalization in single carrier (SC) wideband multiple-input multiple-output (MIMO) channels. First, a novel turbo blockwise operating equalizer structure is proposed by jointly optimizing the feed-forward and feedback filters at each iteration based on the minimum mean squared error (MMSE) criterion. Optimization of the filter coefficients, utilized for feed-forward equalization and decision feedback, is performed in both time and frequency domains, and the equivalence between them is established by unifying the iterative block equalization with the feed-forward and feedback filters in various domains. The corresponding filters are obtained analytically in a closed form as a solution of the constrained Wiener-Hopf equation with the use of average reliability information. Second, asymptotic performance and diversity analysis of the proposed single carrier frequency domain equalizer (SC-FDE) is carried out that leads to insights on the structures of the filters in the high SNR regime. Furthermore, we elaborate on the rate and diversity tradeoff of the space-time coded SC-FDE based on the proposed equalizer structure and the space-frequency coded OFDM systems. Performance comparison between the proposed SC-FDE scheme and OFDM is made. The proposed SC-FDE scheme performs very close to the genie-aided performance bounds and better than OFDM based transmission.

On the error floor analysis of turbo codes: weight spectrum estimation (WSE) scheme

In this paper, we propose an efficient method to find the smallest-weight codewords of a turbo code with specific interleavers. This enables us to obtain accurate approximations to the error floor performance of turbo codes.

Cooperative diversity in carrier frequency offset

Cooperative diversity can enjoy the same benefits of MIMO systems when there is no possibility of carrying multiple antennas on a communication terminal. The main distinction between MIMO and cooperative diversity systems is the inherent asynchronism in the latter. We propose a model to study the effects of carrier frequency offset along with symbol asynchronism. Based on the proposed model, a Viterbi equalizer is utilized as a solution for demodulation. Through simulations, it is observed that such a solution suffers a graceful loss with increasing carrier frequency offset

Collision free row column S-random interleaver

Parallel decodable turbo codes (PDTCs) are suitable for concurrent decoding and hence have low latency. Memory collision issue is an important problem met during parallel processing. In this article, we propose a collision free interleaver for parallel processing operations. The performance of PDTCs is analyzed with the proposed random interleaver preventing the memory collision problem. Distance spectra of PDTCs with the proposed interleaver are computed and compared to those with S-random interleaver.

Cooperative Multiple-Access in Fading Relay Channels

Virtual antenna arrays can be constructed via relaying even in the case that there is insufficient physical space or other resources for multiple antennae on wireless nodes. When there is a multiple access scenario, relaying offers a variety of ways to establish communication between source and destination nodes. We will compare a scheme based on space division multiple access to previously studied time division based ones. We observe that space division improves especially the ergodic capacity.

ML performance analysis of digital relaying in bi-directional relay channels

The research challenges to enhance the throughput of communication networks revealed the concept of network coding which is based on the idea of coding at packet level. The bi-directional relay channel in which two user terminals exchange independent information with the help of a relay terminal is among the prominent applications of network coding in wireless cooperative communications. In this paper, maximum likelihood (ML) performance analysis of digital relaying based bi-directional communication is presented. The three stage communication scenario is considered where the first and second stages are allotted to the transmissions of the users and after demodulation the relay broadcasts the network encoded packet obtained by a bit-wise XOR operation to the users in the third stage. The average bit error rate of this scheme with ML detection is derived for Rayleigh fading channels under noncoherent BFSK and coherent BPSK signaling. Simulation results are also presented to validate the theoretical analysis. Analytical and simulation results show that digital relaying based three stage bi-directional communication scheme with ML detection not only outperforms the noncooperative system that involves direct transmissions between user terminals but also prevents the diversity level reduction due to error propagation. Copyright

Blind Channel Estimation Based on the Lloyd-Max Algorithm in Narrowband Fading Channels and Partial-Band Jamming

In wireless communications, knowledge of the channel coefficients is required for coherent demodulation. In this work, a blind channel estimation method based on the Lloyd-Max algorithm is proposed for single-tap fading channels. The algorithm estimates the constellation points for the received signal using the Lloyd-Max algorithm. The algorithm is investigated for frequency hopping systems with small hop durations and operating under partial-band jamming for both detecting the jammer and estimating the channel. The performance of the Lloyd-Max channel estimation algorithm is compared to the performance of pilot-based channel estimation algorithms and non-coherent demodulation and decoding. A special concatenated channel code is constructed to work with the proposed blind channel estimation due to the phase ambiguity in estimates.

Fast Decodable Turbo Codes

Decoding delay is an important consideration for the use of turbo codes in practical applications. We propose a new structure for turbo codes which is very suitable for parallel decoding. It is shown by union bound analysis and simulation results that the proposed system performance is comparable to that of the classical turbo codes

Error Rate Analysis of GF(q) Network Coded Detect-and-Forward Wireless Relay Networks Using Equivalent Relay Channel Models

This paper investigates simple means of analyzing the error rate performance of a general q-ary Galois Field network coded detect-and-forward cooperative relay network with known relay error statistics at the destination. Equivalent relay channels are used in obtaining an approximate error rate of the relay network, from which the diversity order is found. Error rate analyses using equivalent relay channel models are shown to be closely matched with simulation results. Using the equivalent relay channels, low complexity receivers are developed whose performances are close to that of the optimal maximum likelihood receiver.

Subcarrier allocation in OFDMA with time varying channel and packet arrivals

This study considers the design of efficient low-complexity algorithms for dynamic allocation of subcarriers to users in a multiuser transmitter using Orthogonal Frequency Division. The effects upon throughput and delay performances of several canonical algorithms of varying the number of users, the number of subcarriers, and the statistical characteristics of incoming packets are investigated. Consequently, a subcarrier allocation algorithm with low computational complexity and satisfactory performance in almost all cases of interest is developed. The significance of channel v.s. queue state information with respect to various statistical properties of packet arrival processes is explored through extensive simulations. The results carry implications for the amount of channel estimation and feedback necessary. Finally, a new experimental method is proposed and applied for obtaining the stability regions of the algorithms in consideration.