Single-user channel estimation and equalization

Performance Evaluation of a Multi-band UWB Location and Communication System Based on Modified Gegenbauer Functions

“…In a transportation application, the channel is highly variable: the situations are very different in tunnel and outdoor parts. The theoretical results presented in this article represent then only an optimum, which may be only approached by resorting to (blind or supervised) channel equalization techniques [17,18], to be developed. In parallel typical channel models (theoretical or emanating from experiments) may be established.…”

“…From the presence of the cosinus term in the cross-correlation expression of equation (13) and the randomness of the term φ ij in its argument, one may deduce some properties that lessen the computations involved in the BER estimation procedure from equation (17). One may first notice that, to each value of the cosinus term (corresponding to some value of φ ij ), one may associate its opposite (corresponding to φ ij ± π).…”

Performance Limits of Alphabet Diversities for FIR SISO Channel Identification

“…MSK] modulation for which h becomes globally nonidentifiable but remains locally identifiable. This is in contrast with C AMV(Rỹ) (1) (10) and (11) respectively, for whichH(h) becomes rank deficient. This behavior of the LS/SS algorithm is interpreted by the "noise eigenvector" that is mistaken for a "signal eigenvector" when the channel is close to the non identifiability conditions.…”

“…ISO blind identification has been long considered to need High-Order Statistics (HOS) [1], [2]. Actually, it is now well known that the use of an additional diversity at the receiver permits to build a SIMO channel that can be identified with the sole help of second order statistics, e.g., via subspace techniques [3]; if spatial diversity is not available at the receiver, oversampling allows to increase diversity only in the presence of sufficient excess bandwidth, which is however rarely encountered.…”

On channel estimation using superimposed training and first-order statistics

“…This wastes resources. An alternative is to estimate the channel based solely on noisy exploiting statistical and other properties of [3], [5]. This is the blind channel estimation approach.…”