Skip to Main Content
In this paper, we investigate the possibility of using Markov chains to model the error process in the data blocks delivered by the physical layer of wideband code division multiple access a (WCDMA) cellular system. Suitable Markov models (MM) are designed to fulfil the two following objectives: First, an upper layer protocol supplied by the output obtained from the MM should behave as if it were running on the actual physical layer; second, MM parameters should be linked via simple relationships to the main physical layer parameters. Starting from the results on the error statistics obtained from a suitable simulation tool which jointly performs system and link level analysis, we first classify the users on the basis of performance level and burstiness, and then, we provide some guidelines for the design of Markov models in the different system and channel conditions. The performance of an automatic repeat request (ARQ) (Go-Back N) protocol at the link layer is taken as an example to test the accuracy of the proposed models. It is shown that the perspective of using simple error models in the analysis of upper layer protocols is feasible in many cases.