ECM and SAGE based joint estimation of timing and frequency offset for DMIMO–OFDM system

Distributed multiple-input multiple-output (DMIMO) technology is a key enabler of coverage extension and enhancement of link reliability in wireless networks through distributed spatial diversity. DMIMO employs classic relay channels in between the source and the destination to opportunistically form a virtual antenna array (VAA) for emulating cooperative diversity. Use of multiple antennas at the relays further increases capacity and reliability of the relay–destination channel through multiplexing and diversity of MIMO antennas respectively. In such network, the signal received at the destination is characterized by multiple timing offsets (MTO) due to different propagation delay and multiple carrier frequency offsets (MCFO) due to independent oscillators of the relays. Hence, synchronization becomes a crucial issue in DMIMO in order to realize the distributed coherence. In this paper, we address joint estimation of MCFO and MTO in DMIMO orthogonal frequency division multiplexing (OFDM) with MIMO configuration at the relays for estimate-and-forward (EF) relaying protocol. Two iterative algorithms, based on expectation conditional maximization (ECM) and space alternating generalized expectation–maximization (SAGE) are proposed for joint estimation in presence of inter carrier interference (ICI). The robustness of both the estimators to ICI is evaluated by mathematical analysis and supported by extensive simulations. The performance of the proposed estimators is assessed in terms of mean square error (MSE) and bit error rate (BER). The theoretical Cramer–Rao lower bound (CRLB) of estimator error variance is also derived.     

Model Order Estimation for Joint Demodulation in DS-CDMA Systems

An iterative algorithm for estimating the number of relevant transmission paths characterized by the model order vector L in the uplink of a direct sequence code division multiple access (DS-CDMA) system is presented. Estimates of simplelower bounds on the probability of correct one-shot symbol estimates are formulated as a function of bit-error rate estimates of all users and maximized w.r.t. L. The joint demodulation for a given value Lis accomplished using a combination of a space-alternating generalized expectation-maximization (SAGE) algorithm for channel parameter estimation and a°multistage detector. Simulation results indicate that the modelorder is estimated correctly even in situations with considerably differentsignal power levels.     

Monotonically Overrelaxed EM Algorithms

We explore the idea of overrelaxation for accelerating the expectation-maximization (EM) algorithm, focusing on preserving its simplicity and monotonic convergence properties. It is shown that in many cases, a trivial modification in the M-step results in an algorithm that maintains monotonic increase in the log-likelihood, but can have an appreciably faster convergence rate, especially when EM is very slow. The method is applicable to more general fixed point algorithms. Its simplicity and effectiveness are illustrated with several statistical problems, including probit regression, least absolute deviations regression, Poisson inverse problems, and finite mixtures. This article has supplemental materials available online.     

In vitro estimation of fast and slow wave parameters of thin trabecular bone using space-alternating generalized expectation-maximization algorithm

In testing cancellous bone using ultrasound, two types of longitudinal Biot’s waves are observed in the received signal. These are known as fast and slow waves and their appearance depend on the alignment of bone trabeculae in the propagation path and the thickness of the specimen under test (SUT). They can be used as an effective tool for the diagnosis of osteoporosis because wave propagation behavior depends on the bone structure. However, the identification of these waves in the received signal can be difficult to achieve.In this study, ultrasonic wave propagation in a 4 mm thick bovine cancellous bone in the direction parallel to the trabecular alignment is considered. The observed Biot’s fast and slow longitudinal waves are superimposed; which makes it difficult to extract any information from the received signal. These two waves can be separated using the space alternating generalized expectation maximization (SAGE) algorithm. The latter has been used mainly in speech processing.In this new approach, parameters such as, arrival time, center frequency, bandwidth, amplitude, phase and velocity of each wave are estimated. The B-Scan images and its associated A-scans obtained through simulations using Biot’s finite-difference time-domain (FDTD) method are validated experimentally using a thin bone sample obtained from the femoral-head of a 30 months old bovine.     

SAGE based iterative receiver for joint synchronization and channel estimation in uplink MIMO-OFDMA systems

A new Turbo iterative receiver structure is proposed for the uplink multiple-input multiple-output orthogonal frequency division multiple access （MIMO-OFDMA） systems. The space-alternating generalized expectation-maximization （SAGE） algorithm is naturally embedded in the framework of iterative receiver to perform synchronization and detection usin~ the Turbo detector outputs. In each iteration, the expectation step intends to remove the multiple access interference （MAI） caused by other asynchronous users, and the maximization step is utilized to estimate the required parameters （i.e., timing offset, carrier frequency offset, channel state information, etc.） sequentially for each user. Simulation results show that the proposed algorithm can approach the performance of ideal receiver closely, while the processing complexity is rather lower than the conventional detectors.     

基于SAGE的分布式虚拟现实框架

文中借鉴了国内外的典型应用,提出了基于SAGE(Simple Adaptive Grid Engine)的分布式虚拟现实框架.该框架简单高效,且易于应用开发.通过性能和可扩展性这两项指标的测试,表明该框架在中小规模的分布式虚拟现实环境中具有良好的性能和可扩展性.     

一种基于SAGE算法的OFDM系统频偏跟踪技术

正交频分复用系统（OFDM）的性能受限于频偏误差的影响，基于迭代方式的频偏跟踪技术可以较好地消除残余频偏，降低计算复杂度，并能够适应时变的信道参数。提出一种基于SAGE（Space-Ahemating Generalized Expectation-maximization）算法的迭代频偏跟踪方法。仿真结果表明，这种方法可以在很少的迭代次数里实现很高的估计精度，有效提高在时变信道中的接收性能。     

Low-complexity SAGE-based MMSE channel estimation for CoMP multi-user systems

Coordinated multiple point(CoMP) transmission/reception has been investigated recently as a promising technology to increase the cell-edge user performance of long term evolution-advanced(LTE-A),and channel estimation is a crucial technology for CoMP systems.In this paper,we consider a reduced-complexity minimum mean square error(MMSE) channel estimator for CoMP systems.The estimator uses space alternating generalized expectation maximization(EM)(SAGE) algorithm to avoid the inverse operation of the joint MMSE estimator.In the proposed scheme,the base stations(BSs) in the CoMP system estimate the channels of all the coordinated users serially and iteratively.We derive the SAGE-based estimators and analyze complexity.Simulation results verify that the performance of the proposed algorithm is close to the joint MMSE estimation algorithm while reducing the complexity greatly.     

OFDM联合信道估计与均衡技术研究

OFDM系统在时变信道中会受到子载波间干扰,单独进行信道估计和信号检测的策略对于提高接收机的抗干扰能力有限,将信道估计和信号检测统一考虑则可更有效地抵抗子载波间干扰。针对此问题,基于迭代SAGE算法提出了一种新的联合信道估计与符号检测算法,为了减低算法的复杂度,引入BEM信道建模方法。仿真评估了BEM算法的归一化均方误差性能,验证了BEM建模的有效性,显示该算法的误码率优于基于BEM算法的线性均衡检测算法和基于MMSE的干扰对消算法,而且该算法只通过少数几次迭代便可达到收敛域,较好地克服了子载波间干扰的问题。