Sung-Moon Shin

Burst error smoothing by using the outer loop power control in DS-CDMA systems

In this paper we propose an algorithm for reverse link outer loop power control in COMA systems to smooth burst errors. DS-CDMA systems require tight control of transmit power to obtain the required performance. In DS-CDMA systems, closed loop power control on the reverse link is executed the inner loop and the outer loop. The inner loop allows the mobile station to reach its current signal to interference ratio (SIR) target through using a sequence of power up/down commands. The outer loop dynamically adjusts the target SIR so that adequate performance in terms of frame error rate (FER), the truly quality measure, is achieved. The propagation conditions of different mobile stations are different. For this reason, the effectiveness of the inner loop is not the same for all users. In some cases, it always comes close to the target, while in others the variability could be quite significant. The outer loop compensates for such variability by adjusting the SIR targets so that all users obtain the same performance in terms of error rate. In this work, we propose an outer-loop power control algorithm using the leaky bucket (LB) scheme. The characteristics of the algorithm are based on the frame error count, which is increased or decreased by using the leaky bucket principle. The leaky bucket scheme is based on two parameters: T (bucket limit) and D (drip rate). We adaptively adjust step 5izes, T and D. The algorithm will be simulated to prove its advantages for outer-loop power control.

Adaptive learning rate and limited error signal for multilayer perceptrons with n-th order cross-entropy error

Although an n-th order cross-entropy (nCE) error function resolves the incorrect saturation problem of conventional error backpropagation algorithm, the performance of multilayer perceptrons (MLPs) using the nCE function depends heavily on the order of nCE. In this paper, we propose an adaptive learning rate to make the MLP performance insensitive to the order of nCE. Additionally, we propose a method to limit error signal values at the output nodes for stable learning with an adaptive learning rate. The effectiveness of the proposed method is demonstrated in a handwritten digit recognition task.

Protocol test of the WCDMA smart antenna system

The smart antenna WCDMA IMT-2000 system (SWS) is developed based on three function groups - service resource, service control, and service management groups. The SWS functions, in this paper, are defined as the functions required for the SWS operations. In order to test all these functions efficiently, ETRI developed a testbed for the SWS. In this paper, we discuss the SWS testbed architecture from the point of view of implementing the SWS functions. In the testbed, the commercial user equipment (UE) handsets are used. Therefore, This paper is mainly concerned with implementing the radio access network (RAN). A cell of the SWS consists of 3 sectors to easily cooperate with the existing WCDMA IMT-2000 system. A sector has antenna array system of 8 antenna elements, which makes maximum 12 beams. The propagation characteristics differences among paths between antenna element and transceiver are calibrated in the system. For the down link from Node-B to UE, the beam-form is fixed, while for the up link from UE to Node-B, the beam is formed adaptively to the dynamically changing environment. The SWS follows the Uu interface of the 3GPP specifications. Services supported are such circuit services (CSs) as voice of 12.2kbps and H.263 video of 384kbps. The power control techniques of open loop and closed loop are adopted and the softer and soft handoffs are supported.

Adaptive call admissions for multimedia services in DS-CDMA cellular systems

Link capacity is defined as the number of channels available in a link. In DS-CDMA cellular systems, this is limited by the interference present in the link The interference is affected by many environmental factors and thus the link capacity of the systems changes with the environment. In addition, since various traffic types mutually interfere with different amount, their different power levels affect the link capacity. Actual capacity of the systems must be evaluated accurately and call admission control (CAC) needs to be adapted to the actual capacity which varies with the environment. Since the capacity of CDMA system is limited by the reverse link we are concerned with the reverse link and analyze the actual link capacity for multimedia services under varying environments. The analysis results yield a CAC adaptive to dynamically varying environment for multimedia services and thus improve CDMA capacity significantly.

A power allocation scheme maximizing a discrete capacity function of OFDMA system

In this paper, we propose an adaptive power allocation method for an orthogonal frequency division multiple access (OFDMA) system that employs an adaptive modulation and coding (AMC) scheme. Unlike the water-filling scheme that has been aimed to maximize a continuous capacity function, the proposed scheme is to maximize a discrete capacity function that more precisely characterizes the discrete AMC levels. In terms of complexity, numerical results show that the proposed scheme reduces the processing time by more than 80% comparing with greedy adaptive power allocation (APA), and shows as same performance as greedy APA.

A Mathematical Derivation of a Resource Allocation Matrix for an OFDMA System

In this paper, a resource allocation matrix (RAM) for an orthogonal frequency division multiple access (OFDMA) system is derived to maximize overall data throughput under a strict condition for user fairness. The RAM identifies the most preferable allocation state of OFDMA sub-bands. It is subject to a user fairness condition such that one sub-band is allocated per user during an OFDMA time-slot, allowing each user an equal opportunity to use the sub-bands. The proposed resource allocation scheme using the RAM is evaluated in terms of throughput and user fairness by comparison with the Proportional Fairness (PF) scheme and the Max C/I scheme. Numerical results show that the proposed scheme has much higher throughput under the strict user fairness condition. Even though the PF and Max C/I are given an advantage of ignoring the user fairness condition, the throughput performance of the proposed scheme is quite comparable to that of the PF and Max C/I.

Adaptive channel assignment for different types of traffic in DS-CDMA cellular systems

The link capacity of DS‐CDMA cellular systems is limited by the interference contained in the link. This link interference is affected by many environment factors and thus the link capacity varies with the environment. Since link capacity changes with the varying interference and different traffic types mutually interfere with each other, it is difficult to use link capacity efficiently. Static channel assignment (SCA) based on fixed link capacity is inefficient for DS‐CDMA cellular systems. To improve system capacity, channel assignments need to be adapted to variations in interference. In this paper, we propose an adaptive channel assignment (ACA) for different types of traffic. The proposed ACA is based on the reverse link power received at the base station and is adaptable to dynamically varying environments. It consists of two schemes: nonprioritized and prioritized. In the nonprioritized scheme, there is no difference in channel assignments between calls. In the prioritized scheme, however, the number of nonpriority calls acceptable is limited. In both schemes, a channel is assigned if the link power after assigning the channel is less than the power allowed in the link. The performance is evaluated in terms of link capacity and service grade. Utilizing the proposed algorithm yields more link capacity than using SCA in such environment changes as nonhomogeneous traffic load or varying path loss. Service grade is also improved by properly limiting the number of nonpriority channels.