Advanced Uplink Packet Scheduling for Dedicated Traffic Channels

Project Details


In this project we have focussed on new uplink packet scheduling concepts for WCDMA evolution as a logical counterpart to the high-speed downlink packet access (HSDPA) standard. Apart from the logical evolution, the motivation is also with the possible capacity limitations when future and more demanding uplink applications develop. Specifically, we have considered packet scheduling for non real-time traffic on dedicated channels in WCDMA and evaluated different concepts in which much of the Packet Scheduler (PS) functionality is moved from the radio network controller (RNC) to the base station part of the network. Primarily, this is to allow faster reaction to changes in the non real-time traffic for better utilisation of the air interface capacity, and secondly to allow for more physical layer aware scheduling concepts. For one part, we investigated a blind data rate detection scheduler that relies on the transport format indication already specified for the uplink of WCDMA. From dynamic network simulations under bursty traffic conditions, and low mobility vehicular channels, we have estimated the cell capacity gain from this scheduler to be around 18% compared to a representative RNC based PS. Furthermore, this gain can be achieved irrespective of the gain from adding a fast physical layer retransmission protocol (ARQ) so that the combined gain will be in the order of 25-30% compared to the RNC PS with radio link control ARQ. Fast physical layer retransmission is likely to be implemented together with fast scheduling. We also considered fast packet schedulers based on time and rate scheduling - a concept which is radically different from the normal rate scheduling concept of WCDMA. This allows us to take advantage of the channel state and thereby benefit from multi-user diversity. Taking into account the fairness in serving users we have demonstrated that the channel dependent PSs can achieve a cell capacity gain of almost 50% compared to the RNC scheduler. For this we have assumed a 2 ms transmission time interval and fast physical layer ARQ. For larger transmission time intervals it becomes more difficult to track the channel and the gain will become smaller. Also, limitations from implementation, particularly signalling delays, will effect to the gain numbers. (Claudio Rosa, Konstantinos Dimou, Troels B. Sørensen, Preben E. Mogensen; Jeroen Wigard, Nokia)
Effective start/end date31/12/200331/12/2003


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.