Catálogo de publicaciones - libros

In the Ubiquitous Society, there will be many types of mobile access network surrounding us and we can access the Internet anytime anywhere. At that time, mobile device can select several links from surrounded mobile access networks and access the Internet with multiple interfaces. We have already Mobile IPv6 protocol that supports mobility and try to extend to support multiple Care-of Addresses registration. But, we don’t have any solution for selecting effective path. The effective path has many advantages such as reducing communication overhead. In this paper, we propose that effective path selection method in Multiple Care-of Addresses Mobile IPv6 environment with ‘Parallel Delay Measurement’ technique. With our technique, we can make down average packet delay.

This paper proposes an efficient heuristic for solving the problem in energy- and resource-constrained ad hoc networks. The heuristic organizes the network in clusters aiming to minimize a given cost function. The function used measures the total communication cost between all nodes within the cluster, keeping a minimum amount of resources per cluster.

The clusterhead selection of the proposed heuristic is based on the division of labor encountered in social insects. The idea is that each node has probabilistic tendencies to assume a determined role in the network. For example, nodes with good connectivity and high energy level are good candidates for being clusterheads. The probability of assuming a determined role is based on a node’s fitness for the specific role and the actual (reflected by stimulus) of the role in the current network context. After becoming clusterhead, a node starts recruiting members in order to reach a minimum amount of resources that have to be available in the cluster. The procedure is based on a membership fitness function that evaluates the suitability of a node for the cluster.

The realized simulations demonstrate that the proposed heuristic performance was about in average 25% inferior to the global optimum.

Since the context plays a significant role in ubiquitous computing environment, many researches have studied about context-awareness system to improve the performance. An efficient learning mechanism is in importance of context-aware system, but there are seldom algorithms focused on convenience of systems by elaborating the learning mechanism with user’s context information. As one of the most adaptable algorithm, Back Propagation provides us favorable inference capability. In this paper, we concentrate on improving the predict ability and reducing the system workload by proposing improved self-adaptive back propagation algorithm. The middleware we proposed improves the predicate capability. Thus, the overall performance becomes better than other systems. By adding system cache to middleware, it is possible for the context-aware system to act faster and improve the workload efficiency. Experiments show that there is an obvious improvement in overall performance of the context-awareness systems.

This article presents the concept of an artificial hormone system for a completely decentralized realization of self-organizing task allocation. We show that tight upper bounds for the real-time behavior of self-configuration can be given. We also show two simulation results using the artificial hormone system demonstrating the operation of the artificial hormone system under different workloads.

The benefits of component-based service composition are immense however due to their exponential complexity; their real time implementation is a big challenge. In hybrid networks i.e. ubiquitous zone-based networks (u-Zone Networks), the high intensity of service requests can effect drastically on the performance stability of service gateways. In this paper, we present a self-aware service request optimization algorithm for autonomic self-healing systems. We propose when a service request is received and the system is found in the overload state where it can not entertain more service requests, instead of imposing denial of service, the gateway would evaluate the workload of the worker thread at the gateway and reschedule the service request by deferring it. According to our simulation result, the proposed delay time algorithm, if implied, enable gateways with more stability in order to process high flux of service request loads even beyond the saturation point.

Reconfigurable hardware such as FPGAs combines performance and flexibility, two inherent requirements of many modern electronic devices. Moreover, using reconfigurable devices, time to market can be reduced while simultaneously cutting the costs. However, the design of systems that beneficially explore the reconfiguration capabilities of modern FPGAs is cumbersome and little automated. In this work, a new approach is described that starts from a very high level of abstraction, so-called , and exploits the additional information of this level of abstraction to beneficially execute on reconfigurable devices. Particularly, the approach focuses on dynamic run-time reconfiguration on partially reconfigurable FPGAs. As a first introduction to this approach, we consider stream parallelism paradigms including their composition.

We propose a framework that supports user-friendly configuration of a new location sensor system and its integration with a location manager. The proposed framework abstracts the diversity of heterogeneous sensor technologies using adapters that provide a common interface to the location manager. Configuration of a location sensor system requires information provided by the vendor of the location sensor system, so we propose a configuration protocol with which a newly deployed location sensor system can provide and obtain configuration options and parameters. An integration protocol is proposed as well so that a newly deployed sensing system can be integrated as part of an existing location manager. In order to verify the efficiency of the proposed framework, we measured configuration time with our framework and against manual configuration. Experimental results show that the proposed framework reduces configuration time significantly.

Information retrieval is one of the most challenging areas in which the ontology technology is effectively used. Among them, image retrieval using the image metadata and ontology is the one that can substitute the keyword-based image retrieval. In the paper, the retrieval of visual media such as the art image and photo picture is handled. It is assumed that there are more than one service providers of the visual media, and also there is one central service broker that mediates the user’s query. Given the user’s query the first step that must be done in the service broker is to get the list of candidate service providers that fit the query. This is done by defining various ontologies such as the service ontology and matching the query against the ontology and providers. A novel matching method based on the ASN.1 is proposed in the paper. The experiment shows that the method is more effective than the existing tree-based or interval-based methods.

We propose a new service discovery method based on Shared- Space concept. SharedSpace is a virtual community space for service registration and sharing, which is similar to a chat room in a chat system. Any user can freely create a SharedSpace and register his/her services. Others can join the created SharedSpace as members and register their services. All registered services can be shared by the SharedSpace members as if they are in a single network. Detailed mechanism is designed and implemented with Obje middleware in order to validate our scheme.

SystemC is a popular open source library in C++ for developing embedded system design, from the abstract System Level Design until the accurate Register Transfer Level Design. SystemC simulation, however, runs in console mode (text-based), thus making it difficult for user to interact with the simulation. To extend the capabilities of SystemC simulation, it is necessary to create Graphical User Interface. In this paper we recommend to use RapidPLUS tool for making the interface for embedded system prototype to reduce time to market. We also propose the connectivity between RapidPLUS and SystemC by using socket communication that attached in JavaBean RapidPLUS object