Model based adaptive channel identification algorithms for digital mobile radio

A class of adaptive algorithms for channel identification within the highly time variant radio communications environment is proposed. The algorithms, referred to as "model based" (MB) algorithms, differ from conventional least squares (LS) or gradient search (GS) algorithms in that a priori knowledge of the time variant characteristics of the channel, in the form of a predetermined model, is incorporated directly into the structure of the algorithm. Time variation compensation is then performed intrinsically within the algorithm recursion. As a result MB algorithms offer the possibility of performance advantage over more conventional techniques. The paper shows that under very highly time variant conditions, for example those found in the digital mobile radio environment, MB algorithms may out-perform other types of adaptive algorithm.<
>