Nilo Casimiro Ericsson

Licentiate Thesis, Signals and Systems, Uppsala University, June 2001

Abstract:

This thesis addresses problems appearing in the transition from a fixed Internet to a mobile and wireless one. The focus is on scheduling of downlink transmissions of packet data over wireless links with varying quality, due to multipath fading.

Recent work suggests that predictions of channel quality can be achieved for several milliseconds ahead in time, for fast moving (vehicular) users. In this thesis we try to answer the question of how accurate, or, how long predictions we need in order to make efficient resource allocations through time-slot scheduling, and adaptive modulation.

An alternative approach to a deterministic scheduling of link resources, is the usage of incremental redundancy, or so-called Hybrid type-II ARQ. This thesis suggests an approach where predictive scheduling is combined with incremental redundancy, to obtain a fast, and robust, predictive resource allocation scheme.

The conclusions drawn are mainly: 1) A prediction of the signal-to-interference ratio with error standard deviation of up to 3.5 dB significantly improves the quality of the link. It makes it possible to maintain a nearly error free transmission at a low additional cost in delay, when using a simple time-slot scheduler with Hybrid type-II ARQ. 2) For error-sensitive applications, utilizing TCP/IP, a robust link layer protocol with ARQ should be used to avoid expensive transport layer re-transmissions.

These conclusions are drawn from network- and link layer simulations, and are based on expected performance at the transport layer.

Related publications:

Ph.D. Thesis by Nilo Casimiro Ericsson.

PCC Wireless IP project complete publication list.

Adaptive Modulation and scheduling, IEEE VTC99-fall.

Adaptive Modulation and scheduling, IEEE Globecom 1999.

Prediction of mobile radio channels, Licentiate thesis by Torbjörn Ekman, 2000.

Hybrid type-II ARQ/AMS supported by channel predictive scheduling in a multi-user scenario, IEEE VTC00-fall.

Later version in NRS/PCC Workshop 2001, Nynäshamn.

Scheduling as an optimization problem in NRS/PCC Workshop 2001, Nynäshamn.

Source:

Pdf, 1.7K