"Advancing the future of wireless communications"

ACCESS CONTROL FOR VOICE/DATA CDMA
By Christina Comaniciu and Narayan Mandayam

Information access nowadays, invariably involves multimedia data in some form or another. As a consequence, there is a need for efficient mapping of the QoS requirements of multimedia systems onto wireless environments. Taking into account the complementary QoS requirements that voice and data have, an access control protocol maintains the total system interference below a certain level.  It does so by scheduling the data users to transmit more packets in periods of low voice activity and reducing the transmission when the voice load is high. This talk describes access control protocols for integrated voice/data CDMA systems that are based on estimating the residual capacity available for data users. For the perfect power control case, the outage condition is defined as the feasibility condition for power control, and the available residual capacity is obtained by using an adaptive prediction technique based on Delta Modulation. The analysis is extended for the case of imperfect power control where the outage condition is defined as the limitation of the total interference power to a fraction of the background noise power, 99% of the time. The design goals for the proposed algorithms are: outage probability less than or equal to 1%, guaranteed minimum data rate throughput for data users, specified average data rate throughput for data users and efficient use of system resources (maximize system capacity). Two access control protocols are proposed: Modified Delta Modulation with Scheduled access (MDM-S) and Modified Delta Modulation with Random access (MDM-R).

Comparisons with earlier proposed protocols reveal a better performance in terms of average throughput for data, for a given system capacity (max number of voice and data users in the system for a 1% outage probability constraint). In addition MDM-S guarantees zero outage for the case of perfect power control and MDM-R is very efficient in terms of downlink utilization (for 10 voice users in the system and 0.4 voice activity, it needs a "2 bit" feedback from the Base Station, less than 15% of the time).

This presentation is available in PDF or Postscript form.

If you have questions, comments or other suggestions please contact us here or email your comments to webmaster@winwww.rutgers.edu