Signals and Systems at Uppsala University, Sweden
Signals and Systems Group Staff Descriptions of all our research areas Publications, in html, pdf and postscript Undergrad  courses and PhD programs News and Events


Master Thesis Seminar at the Signals & Systems Group

Title: A Simulation Model of Coding/Retransmission (Adaptive Modulation with Hybrid Type-II ARQ) in Wireless IP Traffic

Speaker: Mats Vännström


Time and Place:
Thursday, June 13th, at 14:15
Library, floor 2 at Magistern, Dag Hammarskjölds väg 31, Uppsala

Abstract:

In the Wireless IP project, improvements of TCP/IP traffic over wireless fading links are studied. A simulation tool is under development in which computationally demanding algorithms for ARQ with adaptive channel coding and modulation are used. This thesis describes a model that can replace these non-real-time computations. The model in question consists of a fast accessible statistical table for each (re)transmission attempt. Three attempts are allowed. The tables holds the packet error rates corresponding to all investigated combinations of channel coding rate, modulation and channel signal to noise ratio (SNR). All the combinations will form a very large combinatorial tree.

The work was performed in three steps. First, the system model and the simulation tool were studied in order to implement an effective program package for the creation and utilization of statistical tables. Then large simulations were run for all chosen transmission combinations. Finally, the statistical data and the retrieval of data from resulting tables were evaluated.

The main conclusions are: 1) It is possible to do the simulations in a reasonable amount of time (< 35 days) using a smaller cluster of modern computers, if the statistical resolution for the third transmission attempt is allowed to be low, and if the investigated transmission combinations are carefully selected so that the number of combinations is limited. The latter also implies that the retrieval of data from tables will be fast enough to run the simulation tool in real-time. 2) The computed statistical data is reasonable and reliable. 3) Using a limited number of investigated transmission combinations gives only an approximation of the true system behaviour. The accuracy of this approximation should be studied further.

 

Page editor: Mathias Johansson