Amir Laufer

Fourth order cumulants in distinguishing single carrier from OFDM signals

In this paper, we propose to use the fourth order cumulants to distinguish OFDM from single carrier signals. We show analytically the value of C40(N) and C42(N) as a function of the number of subcarriers of the OFDM signal and its SNR. By taking these values as an estimate of these cumulants, we compare it to the estimate of different single carrier modulation signals including SC-BC as it was given in the literature or by experiments we performed as for others. From these values we create thresholds detectors use either one cumulant or both depending on the assumed environments which include timing offset, phase rotation, frequency offset of the received signals, and pulse shaping. Last, we use Monte Carlo simulations under different scenarios to examine the probability of detecting OFDM from different single carrier signals.

A new improved-performance decoding technique for Asymmetrically-Clipped Optical-OFDM

In this work, an improved receiving technique for Optical Wireless Intensity Modulation and Direct Detection (OW IM-DD) communication system, with Asymmetrically-Clipped Optical OFDM (ACO-OFDM) modulations, is proposed. The (socalled in the literature) “clipping noise”, which carries 50% of the transmitted signal power is customarily discarded. Instead we show that the clipping noise carries information related to the original unclipped signal. Hence it is used in the proposed receiver scheme to generate an improved estimator for the transmitted signal. It is shown both theoretically and numerically, that the new approach improves the receiver performance of the system by up to 3dB without a change to the modulated signal.

Mean square error analysis of improved receivers for unipolar communications - Asymmetrically Clipped Optical OFDM

Recently, an improved receiving technique for Optical Wireless (OW) with Intensity Modulation and Direct Detection (IM-DD) communication systems, implementing Asymmetrically-Clipped Optical OFDM (ACO-OFDM) modulation, has been proposed. It uses the previously discarded (so called in the literature) “clipping noise” to generate an improved estimator for the transmitted signal, up to 3dB in the receiver performance, without a change in the modulated signal, and with negligible complexity increase at the receiver. In this work we analyze the performance of the improved receiver, as a function of the input SNR; find the optimal parameters of the proposed receiver; show a range of parameters that guarantee performance improvement over the regular receiver; and find a set of fixed parameters that give near optimal performance for all input SNR, thus establishing a simple receiver without the need to estimate the input SNR.

Improved transmission scheme for orthogonal space time codes

The major drawback of the orthogonal space time codes (OSTC) is the limited rate achieved by this family of codes. The known answer to this problem is the quasi-orthogonal space time codes (QSTC) which enables full rate with the penalty of high complexity decoding. Both degrades the performance of a system with high data rate demand due to the use of large constellation size to compensate the low symbol rate for the OSTC and the use of non-optimal but tractable decoding scheme for the QSTC. In this paper we present a way to increase the OSTC code rate while maintaining its orthogonal property which results in low complex decoding and improved performance for high data rate systems.

Adaptive decoding for space time codes with imperfect channel estimation, using the bootstrap algorithm

Space time coding is a powerful technique to exploit transmitter diversity. Nevertheless, to enjoy its full potential performance, a perfect knowledge of the channel state information should be available at the receiver. Since even small channel estimation error results in severe performance lose, a long, rate consuming training sequences should be used to minimize these estimation errors. Hence, presenting an undesired tradeoff between rate and performance. Applying the ”bootstrap” algorithm; an adaptive decoder, can mitigate the effect of estimation errors, hence enables a way out of this deadlock, by achieving high performance transmission while reducing the required length of the training sequences.

Full rate space time codes for large number of transmitting antennas with linear complexity decoding and high performance

Space time codes (STC) have been shown to be used well with the Multiple Input Multiple Output (MIMO) channel. The Orthogonal STC (OSTC) family of codes is known to achieve full diversity as well as very simple implementation of the Maximum Likelihood (ML) decoder. However, it was proven that with a complex symbol constellation one cannot achieve a full rate code when the number of transmitting antennas is larger than two. Quasi-OSTC can have full rate even for more than two transmitting antennas but with the penalty of decoding complexity which becomes severe if the constellation size is high. In order to tackle these inherent drawbacks of the OSTC/QSTC we propose a new STC code that, when used with a new transmission and decoding methods, achieves full rate while maintaining linear complexity decoding for any number of transmit antennas. Also if the transmitter knows the strongest channel (through minimal feedback) the code also achieves full diversity along with better error rate than the OSTC and the QSTC.

On converting OSTC scheme from non-full rate to full-rate with better error performance

In the paper, ldquoImproved Transmission Scheme for Orthogonal Space Time Codesrdquo, a new scheme is proposed that uses an iterative Expectation Maximization (EM) algorithm for decoding and provides full rate and full diversity. For full rate, some of the codeword symbols are not transmitted but rather estimated at the receiver using the expectation maximization (EM) algorithm. In this paper, we prove analytically that the EM algorithm converges exponentially and unconditionally to the least squares (LS) estimate and the rate of convergence depends on the channel parameters and not on the initial vector. We also propose a new very low complexity decoding for the aforementioned transmission scheme with identical error performance.

Linear computational complexity decoding for semi orthogonal full rate space time codes

Communication with low complexity decoding is an important attribute for many applications. From low cost, distributed sensors arrays with desirably long operative life to ultra smart handheld battery operated devices, low complexity decoding results in better power consumption and simpler hardware implementation. In this paper a linear decoding complexity is applied and analyzed to the semi-orthogonal space time codes which are rate 1 codes originated from a regular orthogonal space time codes. The proposed transmission and decoding schemes are very appealing in the sense that they improve the code rate of the orthogonal codes family, yet, maintain its simple, symbol by symbol decoding with no filter calculation complexity overhead. In addition, a performance boosting modifications are suggested and investigated for the basic transmission and decoding schemes under different system settings such as the availability of limited feedback and multiple receiving antennas.

Bootstrap decoding for the Alamouti space-time scheme with imperfect channel estimation

Imperfect channel estimation results in the introduction of symbols coupling for system with space time codes encoding when decoded by the mismatched filter. The bootstrap algorithm was suggested to be used as a signal separator at the output of the mismatched filter. In this work, this solution is studied in details for the Alamoutis code. An inherent problem of the bootstraps weights control process causes the algorithm not to converge to the optimal weights, making it inapplicable to this code. Alternatively, a new method of weights calculation is presented and analyzed which enables the use of the bootstrap algorithm.

Full Rate Space Time Codes for Large Number of Transmitting Antennas with Linear Complexity Decoding

Among the specification of the 5G networks two crucial aspects are the support of fast mobility and high data rates. With fast mobility, the fading channels phenomenon become crucial, resulting in the need for multiple input/output channel to create spatial diversity. Space time codes (STC) have been shown to be well used with the Multiple Input Multiple Output channel. The Orthogonal STC (OSTC) family of codes is known to achieve full diversity as well as very simple implementation of the Maximum Likelihood (ML) decoder. However, it was also proven that with a complex symbol constellation one cannot achieve a full rate code when the number of transmitting antennas is larger than two. Quasi-OSTC (QSTC) can have full rate even for more than two transmitting antennas but with the penalty of decoding complexity which becomes severe if the constellation size is large. In order to tackle this inherent drawback of the OSTC/QSTC and to be able to support the 5G high data rate demand, we have come up with a different STC code that, when used with a new transmission and decoding methods, achieves full rate while maintaining linear complexity decoding for any number of transmit antennas. It can also be shown that when the transmitter knows the strongest channel (through minimal feedback) the code also achieves full diversity along with better error rate than the OSTC and the QSTC.