SeokYoon Kang

Region-Based Collision Avoidance Beaconless Geographic Routing Protocol in Wireless Sensor Networks.

Due to the lack of dependency on beacon messages for location exchange, the beaconless geographic routing protocol has attracted considerable attention from the research community. However, existing beaconless geographic routing protocols are likely to generate duplicated data packets when multiple winners in the greedy area are selected. Furthermore, these protocols are designed for a uniform sensor field, so they cannot be directly applied to practical irregular sensor fields with partial voids. To prevent the failure of finding a forwarding node and to remove unnecessary duplication, in this paper, we propose a region-based collision avoidance beaconless geographic routing protocol to increase forwarding opportunities for randomly-deployed sensor networks. By employing different contention priorities into the mutually-communicable nodes and the rest of the nodes in the greedy area, every neighbor node in the greedy area can be used for data forwarding without any packet duplication. Moreover, simulation results are given to demonstrate the increased packet delivery ratio and shorten end-to-end delay, rather than well-referred comparative protocols.

Design of hierarchical routing protocol for heterogeneous airborne ad hoc networks

Due to recent research interests in airborne ad hoc networks, many research works have been conducted in each layer by developing communication protocols which take into unique properties of node (aircraft) as well as operating environments account. Among them, as the research efforts in network layer, several routing protocols have been proposed to improve packet delivery ratio even though a node moves faster than one in existing ad hoc networks. Related to routing protocol, as proved in many previous research works, hierarchical routing protocol can be considered as appropriate approach in airborne ad hoc networks due to its stability against dynamic change on topology. To meet above demand, in this paper, we present how to form two level hierarchies for heterogeneous networks by extending current geographical routing protocol. Furthermore, the proposed scheme introduces how to improve the packet delivery ratio by applying DTN (Delay Tolerant Networks) routing protocol adaptively for network partition. Simulation results are given to prove suitability of proposed protocol in heterogeneous airborne ad hoc networks.

Fuzzy searching and routing in unstructured mobile peer-to-peer networks

Peer-to-Peer (P2P) networks offer a scalable solution for efficient query searching and sharing across the unstructured networks. With an increased overhead due to large amount of object searching and routing in unstructured P2P networks, it is a challenge to continue QoS routing among different mobile peers. Most existing mobile P2P protocols focus on inflexible techniques to route queries and discover objects of interest. Such common techniques incur a relatively high search time due to remarkable network traffic and duplication of query messages. The correlation between routing and mobility is crucial to efficiently search and route the query object in an overlay to avoid unnecessary consumption of network resources. Our previous fuzzy search controller model (Shah and Kim in 12th IEEE international conference on dependable, autonomic and secure computing (DASC), 2014) reduced the search time query processing in P2P networks, but it caused low hit rate and a high overhead due to peer mobility. Thus, this article proposes a scalable fuzzy controller based on probabilistic walk for unstructured mobile P2P networks to reduce the search time with controlled mobility. The search time is reduced by jumping a query walker to a 2-hop away ultrapeer, selected through a fuzzy scheme. Furthermore, each mobile ultrapeer shares its pong cache with its directly connected ultrapeer in order to increase the hit rate and reduce the network overhead. Simulations show that the fuzzy search controller gives better performance than the competing protocols in terms time 10% reduction in response time and 15% increase in hit rate in different mobility scenarios.

A mobility-based temperature-aware routing protocol for Wireless Body Sensor Networks

A large number of temperature-aware routing algorithms have been proposed to prevent the damage of surrounding tissues caused by temperature rise in wireless body sensor network. But, since the existing temperature-aware routing algorithms for WBSN have not taken account of human postural mobility, large of packet are lost due to topological partitioning. To solve this problem, in this paper, we propose a new temperature-aware routing algorithm by employing store and carry scheme. Simulation results are given to provide the suitability of the proposed routing algorithm by demonstrating its performance in terms of reducing the number of hot spots, packet delivery delay and packet loss.

Towards Cluster-Based Real-Time Flow Scheduling in Interference-Aware Wireless Sensor Networks

As wireless sensor networks have been used in various fields, real-time applications are also run in sensor networks, such as real-time data aggregation or periodic image transmissions. In order to guarantee real-time flow deadlines, TDMA scheduling schemes can be applied with consideration of both node interference and flow deadlines. Generally, they require much time complexity as the number of sensor nodes and real-time flows increases in the networks. Thus, in this paper, we provide a cluster-based real-time flow scheduling scheme for the purpose of interference-aware real-time flow scheduling in wireless sensor networks. The proposed scheme includes (i) the intra-cluster scheduling which schedules real-time flows among nodes in each cluster, and (ii) the inter-cluster scheduling which schedules composed real-time flows among clusters. Throughout simulation results, we show the effectiveness of the proposed scheme in many aspects.

Three Dimensional Cloud Modeling Approach Based on L-System

Typical methods for cloud modeling in simulation software are usually classified into static and dynamic modeling. While static modeling is accomplished by modeling tool such as 3ds max, dynamic modeling is implemented by particle system. In the former one, it is very hard to create a variety of realistic cloud models due to long implementation time. On the other hand, in the dynamic modeling case, rendering speed becomes slow because of computing overhead to handle large number of particles. Based on mentioned features of two schemes, it is required to create realistic three dimensional cloud model with small computing overhead. In order to fulfill this requirement, in this paper, we propose a cloud modeling and rendering approach based on L-system which is expected to provide scalability and low complexity. To create cloud, existing cloud dynamic model is transformed to rules of L-system in recursive way. In order to verify applicability of the proposed method, we present experimental result which is implemented by Open Scene Graph (OSG) library.

Development of data collecting system for forecasting with meteorological sensors

As one of sensor applications, meteorological service collects information for weather and uses it for preventing disaster risk. To achieve this, data from sensor networks must be integrated with communication and data processing system. In this paper, we present a specific web-based application with mobile networks to transmit weather data timely by introducing forecasting system. We implement central database for remote weather stations and make use of it in a preventive way, to recognize the occurrence of natural disasters by external geo-dynamics.

Beaconless Geographic Routing Protocol with Adaptive Delay Timer in Wireless Sensor Networks

Even though geographical routing protocol is regarded to energy-efficient and scalable approach in wireless sensor networks, position information through beacon messages should be exchanged prior to forwarding procedure. This is another control overhead when it comes to consider nodes constraints. Thus, recently beacons graphical routing have been proposed to perform data forwarding without information exchange by distance-based delay timer and overhearing technique. However, existing schemes causes additional recurrent and redundant delay in cases of network dynamics or partial network holes in each hop. Specially, if the networks are not uniformly deployed, current schemes cannot be applied. To solve this problem, in this paper, we propose a new fast beaconless protocol, which enables neighbor nodes of a sending node to adjust their delay time depending on whether any node is identified by localized backlearning scheme or not. Also, simulation results are presented and analyzed to prove the decreased end-to-delay delay and increased packet delivery ratio.

Dynamic Culling Scheme Based on Altitude for Real-Time Rendering System

Dynamic culling scheme is usually implemented to handle overhead caused by rendering the massive large-scale terrain data in flight simulator. However, existing culling scheme without considering altitude is not suitable for flight simulator due to additional computational overhead. To solve this problem, in this paper, we propose hybrid approach by applying two dynamic culling schemes depending on altitude. In addition, we remove unnessary computational overhead by creating different z-map resolution when aircraft changes its altitude. The proposed scheme is implemented with open graphic library and tested with real terrain data. Through the experimental results, we can recognize the improved rendering speed about 8 to 73 percents as compared to existing scheme.

A new hybrid culling scheme for flight simulator

In flight simulator, real-time rendering system processes most of scenarios at high altitude. Therefore, overload process of real-time rendering system tends to ignore altitude distinction through applying only view-frustum culling. However, even though the flight of the aircraft operates at high altitude, but dynamic culling according to altitude is demanded because of several situations such as takeoff, landing and flight at low altitude. According to this problem statement, in this paper, we propose how to adjust dynamic culling in case of low altitude. To achieve this goal, first, we classify high altitude and low altitude depending on the height of the aircraft motion simulation. Moreover, we apply fuzzy logic into classifies for ambiguous altitude state. If altitude state is low, we use both view-frustum and back-face culling at once in order to reduce complexity of the operations caused by the visibilitytest to dynamically generate the resolution of the ZMap for back-face culling at low altitude state. We use OSG(Open Scene Graph) graphic library and test real world terrain data. Through test results, our proposed algorithm decreases triangles of 15~80% in real-time rending, and improves frame per second of 8~73% as compared to existing one.