Widely Linear System Estimation Using Superimposed Training

Abstract

In this correspondence, the use of superimposed training (ST) as a mean to estimate the finite impulse response (FIR) components of a widely linear (WL) system is proposed. The estimator here presented is based on the first-order statistics of the signal observed at the output of the system and its variance is independent of the channel components if suitable designed training sequences are employed. The construction of such sequences having constant magnitude both in time and frequency domains is also addressed.Index Terms-Joint channel I/Q imbalance estimation, optimum channel independent sequences, superimposed training, widely linear system estimation.

Orthogonal superimposed training design for block transmission over frequency‐selective fading channels

“…We are now going to present the simulation results of our proposed polyphase sequence‐based PPST scheme and its comparison with conventional ST , DDST , and TM training schemes. The training power is set to be 20% of the total transmitted power.…”

“…Usually, estimating h and recovering b from x in is a nonlinear problem and optimal in maximum likelihood sense but often computationally prohibitive. If the HPb in is treated as part of input noise, then h can be uniquely determined if the following channel identification condition is satisfied : …”

“…This is the blind estimation schemes, which are very attractive from the point of view of the transmitter, but the receiver requires longer data records and must absorb all the implementation complexity . In order to avoid the loss of training or pilot bandwidth incurred by TM/FM approaches and still be able to identify the channel parameters with manageable computational complexity, superimposed training (ST)‐based schemes have received enormous attention in this domain , where a training sequence known to the receiver is algebraically added to the data symbols at a low power for the purpose of aiding the channel estimation but without extra bandwidth to convey training information. Unfortunately, this advantage also conveys a penalty.…”

Constellation rotation and symbol detection for data‐dependent superimposed training

“…Introduction: Channel estimation using superimposed training (ST) has been proposed to improve the bandwidth efficiency as compared to time-multiplexed (TM) or frequency-multiplexed training [1][2][3]. Unfortunately, the ST-based channel estimate is quite poor due to the interference from the unknown data symbols since the training sequence and data symbols are 'co-mingled' at the receiver.…”

Intel Labs Mexico - Leading Industrial Research in Latin America

“…For instance, in[20] the spatial characteristics of a MIMO system are included to the radio model to provide a better understanding of the system's performance under different scenarios and usage. Also, an estimator for the fmite impulse response of widely linear systems was proposed in[21]. This estimator is particularly relevant to communications channel modeling.In [22] the problems derived from increased parametric variations due to manufacturing process scaling, integration of multiple RF front-ends and analog blocks is addressed through the development of a flexible and portable digital framework…”