Yeon Seung Shin

Dynamic Network Selection using Kernels

We present a new algorithm for vertical handover and dynamic network selection, based on a combination of multi- attribute utility theory, kernel learning and stochastic gradient descent. We show that this new method is able to improve network selection in a non-stationary mobile environment. Furthermore, since the kernel employed is based on the utility functions for attributes such as Availability, Quality and Cost, the kernel regression in fact gives interpretable results. We present simulation results that demonstrate our algorithm being able to dynamically learn utilities and efficiently select networks.

Inter-cell Interference Coordination for Uplink Non-real-time Data in CDMA Systems

CDMA systems, interference is one of very important criteria to evaluate system performance. Interference is classified into inter-cell and intra-cell interference. While the intra-cell interference is generated by mobiles sharing the same base station, inter-cell interference is generated by the mobiles connected to different base stations. Many inter-cell interference mitigation schemes have been proposed to improve system performance. Most of them focus only on the downlink transmission. As for the downlink transmission, only one mobile is allowed to receive data from the base station at a moment except the broadcast service. However, as for the uplink transmission, multiple mobiles may have access to the base station simultaneously. The uplink inter-cell interference management is, therefore more difficult and important than the downlink. In this paper, we propose the inter-cell interference coordination scheme for uplink non-real-time transmission. This scheme shuts down a cell that injects lots of interference. The design of this scheme is based on the assumption that each base station applies the proportional fairness (PF) algorithm to the scheduling of uplink packets. The simulation results show that we can achieve improved system throughput by applying the proposed scheme.

Mobile tracking using fuzzy multi-criteria decision making

In the microcell- or picocell-based system the frequent movements of the mobile bring about excessive traffics into the networks. A mobile location estimation mechanism can facilitate both efficient resource allocation and better QoS provisioning through handoff optimization. In this study, we propose a novel mobile tracking method based on Multi-Criteria Decision Making (MCDM), in which uncertain parameters such as PSS (Pilot Signal Strength), the distance between the mobile and the base station, the moving direction, and the previous location are used in the decision process using the aggregation function in fuzzy set theory. Through numerical results, we show that our proposed mobile tracking method provides a better performance than the conventional method using the received signal strength.

A Study on End-to-End QoS Provision for Multimedia Services in Beyond 3G Convergence Networks

In Beyond 3G networks interworking with external IP-based networks, end-to-end QoS provision is a critical issue. To facilitate this, 3GPP proposes the policy-based network management and suggests IntServ and DiffServ as policy enforcement means. Even if DiffServ is preferable to IntServ in scalability and configurability, it has the limitation of not being able to leverage multiple paths, hence undesirable in the network utilization aspect. In this paper, we propose DiffServ-aware Multiple Protocol Label Switching (MPLS) as a policy enforcement means. It enables to provide differentiated levels of QoS as well as to improve network utilization by splitting packets over multiple paths. It is also able to avoid the per-flow packet ordering problem occurring when splitting a flow into multiple paths by using a hashing-based scheme. We verify the effectiveness of our proposed scheme by the NS-2 based simulation.

Wireless packet scheduling for integrated services of real-time and non-real-time applications in mobile broadband wireless access system

This paper considers an opportunistic packet scheduling algorithm to support both RT and NRT service in mobile broadband wireless access system. The design objective of our proposed packet scheduler is to determine the optimal number of RT and NRT users based on the service charges (which differ depending on the type of services) in order to maximize the revenue, while at the same time satisfying their individual QoS requirements, i.e., the maximum allowable packet loss rate for RT traffic and the minimum reserved bit rate for NRT traffic. It takes an urgency of RT users into account only when their head-of-line (HOL) packet delays exceed the given threshold, which allows for maximizing the system capacity of NRT users. Evaluating the proposed approach in orthogonal frequency division multiple access/frequency division duplex (OFDMA/FDD)based mobile access system, it has been shown that, in terms of the number of RT and NRT service users, system capacity can be significantly improved.