Showing papers by "Rapeepat Ratasuk published in 2002"

Turbo code based incremental redundancy

Abstract: An improved turbo code based incremental redundancy includes a first step of puncturing a data stream for a first transmission to provide a set of first unpunctured trellis sections. A next step includes puncturing a data stream for a second transmission to provide a set of second unpunctured trellis sections. A next step includes incremental redundancy combining the first and second transmissions of the trellises to provide non-adjacent first and second unpunctured trellis sections. The above arrangement results in a uniform distribution of punctured and unpunctured bits to provide lower errors.

Quasi-static method for predicting link-level performance

Abstract: In this paper, we investigate the performance of a quasi-static method used to predict link-level error performance in dynamic system simulations. Two flavors of the quasi-static approach are used to model the link characteristics: (a) static curves with fudge factors; and (b) short-term frame error curves. For the fudge factor based method, we examine the various factors (or penalties) required for the method to work properly, their limitations in capturing the various features of the high-speed packet data channel, and their effects on the overall accuracy of the technique. Here, we concentrate on three important penalties: puncturing penalty, Doppler penalty, and demapping penalty. The quasistatic method based on short-term frame error curves is used to generate frame erasures for voice and control channels associated with high-speed packet data channels. From simulation results, it is shown that both variations of the quasi-static method can provide accurate prediction of link-level performance at certain operating ranges, provided that the penalties are correctly obtained. At low frame error rates and moderate to high speeds, however, the quasi-static method often fails to accurately reflect frame error performance at the link level.

Pilot information gain control method and apparatus

Abstract: Various gain factors for pilot information are used to determine a corresponding plurality of likely channel performance values for a plurality of data channels (and/or services). The resultant information is used to select a specific gain factor to apply when transmitting the pilot information. In one embodiment, the gain factor that correlates to the highest signal to interference plus noise ratio is used.

Performance analysis of time-multiplexed services for UMTS W-CDMA reverse link

Abstract: In UMTS W-CDMA systems, it is possible to offer multiplexed services such as voice, data, and video simultaneously on one physical connection. Several issues such as the QoS requirement for each service, algorithmic complexity, hardware limitations, and performance of the overall system must be considered. In this paper, we investigate some of these issues and present link-level performance for multiplexed services on the reverse-link of an UMTS W-CDMA system. We show that because of an increase in the peak-to-average power ratio when different services are code-multiplexed onto the same physical channel, multiplexed services for UMTS may have to be offered via the time-multiplexing approach. Next, we illustrate how rate matching can be used to satisfy the quality of service requirements for all services. In our results, we illustrate how the rate matching attributes can be selected to satisfy the QoS requirements of 1% speech frame error rate, 10% data frame error rate, and 0.1% video frame error rate. The selection of good rate matching attributes, however, is dependent on many factors such as service rates, channel conditions, and desired error rates. As a result, under some circumstances, excessive transmission power may be required in order to meet the minimum QoS requirements. Next, the performance of the system under compressed mode, where certain frames are compressed in time to allow inter-frequency measurements or handover to other systems, is considered. It is shown that the degradation due to compressed mode is minimal (less than 0.5 dB), and hence compressed mode can safely be used in conjunction with multiplexed services.