Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Queueing theoretic optimal scheduling for multiple input multiple output multiple access channel

: Boche, Holger; Wiczanowski, M.


IEEE Signal Processing Society; IEEE Computer Society:
3rd IEEE International Symposium on Signal Processing and Information Technology, ISSPIT 2003. Proceedings : 14 - 17 December 2003, Maritim Rhein/Main Hotel, Darmstadt, Germany
Piscataway, NJ: IEEE, 2003
ISBN: 0-7803-8292-7
International Symposium on Signal Processing and Information Technology (ISSPIT) <3, 2003, Darmstadt>
Conference Paper
Fraunhofer HHI ()

In the framework of new services in wireless networks beyond the third generation, asymmetric and flexible uplink transmission schemes are highly desirable. Hence, especially in "data only" networks, an efficient uplink scheduling must not be neglected in the network design. Although not yet well covered in the literature, this area appears to gain importance especially in the face of high speed uplink packet access (HSUPA) radio link design for future wireless systems. In this work we study an uplink scheduling scheme from the information theoretic and queueing theoretic point of view. The objective of the cross-layer (Goldsmith, A.J. et al., 2003) scheduling approach we present is the stability of bit-queues awaiting their transmission at every user node. With this approach is associated an optimization problem of maximizing the weighted sum of rates, where the weights express the intensity of packet-traffics at every user node. An important property characterizing our approach is considering only spatial channelization of users, i.e. allocating user signals within one successive interference cancellation (SIC) order fixed over a certain time interval T. We show, that despite this neglection of time sharing technique, it is possible to achieve an overall optimal stability achieving scheduling for every set of user rate weights. Furthermore, we show that the choice of the SIC-order allowing for the optimal scheduling is independent of user channels and that the associated optimization problem over transmit covariance matrices is convex.