• Proceedings of the KIEE Conference
• /
• 1999.07b
• /
• pp.791-793
• /
• 1999

In this paper, neural network-supervisory control method is proposed to minimize the effect of system uncertainty by load change and disturbance in the PID control system. In the proposed method, PID controller performs main control action by performing control within constraint error. And neural network-supervisory controller performs control action when error reaches the boundary of constraint error. Combining neural network-supervisory controller to guarantee the stability into PID control system, the resulting PID control system is expected to show better performance in the system with load change and disturbance. Simulation applying PID controller and neural network-supervisory controller showed excellence of proposed method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.5B
• /
• pp.508-519
• /
• 2004

This paper proposed two multi-path constraint-based routing algorithms for Internet traffic engineering using MPLS. In normal constraint-based shortest path first (CSPF) routing algorithm, there is a high probability that it cannot find the required path through networks for a large bandwidth constraint that is one of the most important constraints for traffic engineering, The proposed algorithms can divide the bandwidth constraint into two or more sub-constraints and find a constrained path for each sub-constraint, if there is no single path satisfying the whole constraint. Extensive simulations show that they enhance the success probability of path setup and the utilization of network resources.

• Journal of Communications and Networks
• /
• v.8 no.1
• /
• pp.85-92
• /
• 2006

In this paper, we study a wireless cellular network design problem. It consists of selecting the location of the base station controllers and mobile service switching centres, selecting their types, designing the network into a tree-topology, and selecting the link types, while considering the location and the demand of base transceiver stations. We propose a constraint programming model and develop a heuristic combining local search and constraint programming techniques to find very good solutions in a reasonable amount of time for this category of problem. Numerical results show that our approach, on average, improves the results from the literature.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.49 no.8
• /
• pp.39-45
• /
• 2012

This paper investigates the capacity for the secondary user with the primary full duplex relay (FDR) network in spectrum sharing systems. For sharing the spectrum of the primary user, sharing scheme under the various primary network should be investigated. However, the conventional works only consider the point-to-point primary communication system. When the FDR node is used for the primary networks, both primary relay and destination suffer the interference from the secondary transmitter simultaneously. Thus, the strict interference constraint should be considered to share the spectrum of the primary user. Therefore, we investigate the capacity of the secondary user with the primary FDR network under the average and peak received-power constraints. In addition, we analyze the performance degradation by the selection of the worst interference channels to consider interference constraint in this system. Through the numerical results, the capacities of the primary and secondary users under the average received-power is superior to the that under the peak received-power constraint.

• The Journal of Society for e-Business Studies
• /
• v.1 no.2
• /
• pp.161-173
• /
• 1996

It Is a typical scenario in concurrent engineering/design that different perspectives of each team members participating In the project exist, and those perspectives lead to conflicting decisions. The model 'resolution network' proposed in this work provides system-mediated resolution for all related team engineers to consider to optimize the manufacture in general. This paper focuses on development of the general architecture of the model and a search engine called Mediator to determine a resolution network from a given constraint network. The Mediator manages the constraint network, determines the most optimistic resolution called optimal point in terms of satisfying overall production goal , and use the optimal point to mediate controversial issues among teams. The biggest merit of our model is that it provides teams with resolution with logical and rational reasoning.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10b
• /
• pp.1805-1810
• /
• 1991

This paper derives the upper and the lower bound of the mean cycle time and the mean service time of the class 6 and the class 4, within which the minimum utilization constrain of the class 4 is guaranteed. Also, derived are conditions under which the token bus network is stable or unstable. These bounds and stable conditions are represented in terms of the high priority token hold time, the token rotation time and the arrival rate and the total station number etc. This paper suggest a parameter tuning algorithm in a partially symmetric token bus network with two classes, which maximizes the token rotation time for a suitable high priority token hold time and at the same time meets the stability condition of the network, the real time constraint and the minimum utilization constraint of the class 4.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.12 no.1
• /
• pp.81-95
• /
• 2009

The CDB(Constraint Database) model is a new paradigm for massive spatial data processing such as GIS(Geographic Information System). This paper will identify the limitation of the schema structure and query processing through prior spatial database research and suggest more efficient processing mechanism of constraint data model. We presented constraint model concept, presentation method, and the examples of query processing. Especially, we represented TIN (Triangulated Irregular Network) as a constraint data model which displays the height on a plane data and compared it with prior spatial data model. Finally, we identified that we were able to formalize spatial data in a simple and refined way through constraint data modeling.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.6
• /
• pp.2044-2059
• /
• 2022

Recent works have shown that the geometric constraint can be harnessed to boost the performance of CNN-based camera localization. However, the existing strategies are limited to imposing image-level constraint between pose pairs, which is weak and coarse-gained. In this paper, we introduce a pixel-level epipolar geometry constraint to vanilla localization framework without the ground-truth 3D information. Dubbed EpiLoc, our method establishes the geometric relationship between pixels in different images by utilizing the epipolar geometry thus forcing the network to regress more accurate poses. We also propose a variant called EpiSingle to cope with non-sequential training images, which can construct the epipolar geometry constraint based on a single image in a self-supervised manner. Extensive experiments on the public indoor 7Scenes and outdoor RobotCar datasets show that the proposed pixel-level constraint is valuable, and helps our EpiLoc achieve state-of-the-art results in the end-to-end camera localization task.

• Proceedings of the Korean Institute of Communication Sciences Conference
• /
• 1995.06a
• /
• pp.381-384
• /
• 1995

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.2
• /
• pp.196-200
• /
• 2008

Recently, the query processing algorithm in spatial network database (SNDB) has attracted many interests. However, there is little research on route-based query processing algorithm in SNDB. Since the moving objects moves only in spatial networks, the route-based algorithm is very useful for LBS and Telematics applications. In this paper, we analyze In-Route Nearest Neighbor (IRNN) query, which is an typical one of route-based queries, and propose a new IRNN query processing algorithm with time constraint. In addition, we show from our performance analysis that our IRNN query processing algorithm with time constraint is better on retrieval performance than the existing IRNN query processing one.