Ubiquitous Computing Systems

The ubiquitous computing produces big volume of messages as the pervasive computability is deployed in large scale. As a middleware between the message producer and message consumer, message routing system enables backend systems to efficiently acquire the interested message. A widely adopted message routing mechanism is the content-based publish-subscribe framework. Based on this paradigm, we propose a rule-based system for supporting message routing in ubiquitous computing. The novel system features in the flexibility of the message computing, which is accomplished through a set of message operators. The message consumer could select the appropriate operator and specify the operating rule to get satisfying messages.

This paper presents a system which extensively exploits user’s eye gaze information for operating services and appliances in the emerging home network system (HNS). We design and implement the system called AXELLA, which captures user’s gaze, then invokes a service operation, and finally announces the response via voice. AXELLA interprets the gaze information together with supplementary information as a

gaze context

, and triggers a service module associated by a service rule. Thus, a simple gazing activity can be used for various service operations. Service developers (or even home users) can easily develop context-aware HNS services with the eye-gaze-based UI. We demonstrate a practical service called “See and Know” implemented using AXELLA, where a user can acquire the current status information of every appliance just by looking at the appliance. It was shown that the proposed system can reduce the artificial dependency significantly with respect to ease-of-learning and system scalability.

As the ubiquitous computing technology (uT) is prevailing, applying uT is more likely to transform the legacy way of providing services to a new service everywhere, every time, with any devices to gain more comparative advantages than the other typical services. This opportunity naturally requires a methodology to assess to what extent the legacy e-services are transformed to the uT-based services. However, research about assessing the level of u-Transformation has been still very few. Hence, this paper aims to propose a methodology for assessing the level of u-Transformation oriented by the teleology of ubiquitous services. A case study, GPS-based navigation services, is performed to show the feasibility of the methodology addressed in this paper.

3D space handwriting is a gesture-like character written in the air, and it is a promising input method for its portability. In this work, we propose a practical development of 3D space handwriting recognition system by combining 2D handwriting and the ligature of 3D handwriting based on that the most different part between these handwritings is the shape of ligature. We designed a ligature model not to depend on each character shape but to depend only on incoming and outgoing vectors. Therefore with a few ligature models, various ligature shapes are modeled. By using 2D space handwriting models and data, we can use existing models for various writing styles, and overcome the problem of the lack of data.

In this paper scenario-based design approach for development of Ambient Intelligence (AmI) is described. Design of any concept is based on communication between different stakeholders. Furthermore design of these concepts should be grounded heavily to the real user needs. Scenarios offer a flexible tool for these purposes and extremely well scenarios seems to be fitting to the early concept definition of AmI solutions. In this paper we describe our experiences of using scenario-based design approach in the MIMOSA project.

This paper suggests a framework for ubiquitous multimedia access with multidimensional information browsing interfaces and service migration technique. In order to achieve this goal, we propose a multi-dimensional liquid browsing concept, which facilitate browsing of comparably large amounts of information on the various types of devices. Furthermore, the service migration concept among the various types of devices is proposed and experimental results for the image information are shown.

This paper describes the

Obje Perception Framework

(

OPF

), a distributed software architecture for context sensing and inference in ubiquitous computing environments.

OPF

provides flexibility, scalability, and robustness even as the sensor configuration changes. For flexibility,

OPF

supports many context inference tasks, ways of achieving those tasks, and heterogeneity in sensor types. With respect to scalability,

OPF

accommodates the needs of a large number of applications simultaneously while conserving power and reducing the amount of data transmitted over the network. And to support robustness to dynamism,

OPF

constructs context inference pipelines to satisfy each applications’ needs in a goal-directed fashion. The value of

OPF

is demonstrated by a case study of an end-user application that helps users establish and manage connections among the various digital resources in their environment.

With the development of a car-mounted navigation system and an in-car network, improved information services for creating drive plans automatically have been realized. We propose a method of developing telematics services through combination of a context-aware agent framework and a planner that provides the predicted feature situation to the framework and creates drive plans composed of POIs (Points Of Interest) proposed by the framework.

Time based TDOA localization systems require time or clock synch-ronization between receivers such as cellular base stations, satellites, and sensor nodes. Imperfection of time synchronization causes degradation in positioning accuracy. However if we know about its characteristics and how to estimate the clock offsets, the localization system can be properly calibrated to provide good quality of services. In this paper, hence, we present how to derive a localization error vector with independent clock offset, and illustrate its effect on the positioning errors, and then, provide a simple method of TDOA clock offset estimation from the observation of error vectors.

Pervasive computing environments need to exhibit highly adaptive behavior to meet the changing personal requirements and operational context of environment. Recently, task computing (TC) paradigm has gained acceptance as the choice computing model for pervasive computing environment. One of the key components of TC is a task model that provides an adequate high-level description of user-oriented tasks. This paper presents a novel context-sensitive task modeling approach capable of supporting complex, user-oriented task definitions, and proposes an algorithm to discover a task and a method to execute it. This work is motivated by the fact that whilst current TC systems allow users to interact with their ambient environments in terms of high level tasks, existing task definitions are still relatively simple, and do not include user-centric and environmental contextual information in the task definition. This paper elaborates the proposed task model through a smart home application example to illustrate steps in context-dependent task modeling, service provisioning and resource organization.

In our previous work, we have described the COMPASS location system that uses multiple information sources to determine the current position of a node. The raw output of this process is a location in geo-coordinates, which is not suitable for many applications. In this paper we present an extension to COMPASS, the so called Translator, that can provide facts about the location like city name, address, room number, etc. to the application. These facts are represented in the Semantic Web RDF/XML language and stored on distributed Geo RDF Servers. The main focus of this paper is a location-based service discovery mechanism which allows a node to find all services that can provide facts about its current location. This discovery service is built upon a structured Peer-to-Peer system implementing a distributed hash table.

In this article, a formal approach for modeling central concepts of context-awareness in ubiquitous and mobile computing is introduced. The focus is on an appropriate handling of issues of vagueness and granularity in ubiquitous computing environments. A formalization of perceptual and sensory uncertainty and a characterization of granularity are applied for modeling three central aspects of context-awareness: context as retrieved from sensors, context for representing relevance, and context as unfocussed background information. The notions are developed and demonstrated with respect to the special case of spatial contexts, but are sufficiently general to also cover other types of context. Use of the characterized concepts is motivated with an example of ongoing work on ontology design for ubiquitous computing environments.

In order to overcome the various constraints of wireless environments and provide content according to device specifications and user preference, research relating to content adaptation is gaining in significance. For content adaptation, existing research either prepares content in advance, a reflection of client types which may have access to server, or describes the adaptation rules for dynamic content conversion. However, these require a lot of effort from the content author or system developer, and prospecting the appearance of a new device is a difficult work in today’s rapidly changing computing environment. This paper proposes an intelligent adaptation system that automatically extends adaptation rules. The system classifies users into basic categories, then dynamically converts content according to the rule mapping category, offering this result to the user. Then, the system monitors the user action, and performs learning based on this feedback. Moreover, the system has characteristics of offering more personalized content as well as reducing the response time due to reuse of the content generated by same group category. A prototype was implemented in order to evaluate the proposed system in terms of system maintainability, by automatic rule extension, correctness of generated rules, and response time. The effectiveness of the system is confirmed through the results.

Context-awareness which provides relevant information and services to situation using context is an important issue to solve problems in ubiquitous environment. In addition, many problems in ubiquitous environment are solved by interaction between computational elements rather than by performing individual actions. Therefore, it needs to approach context-awareness from in a cooperative point of view. In this paper, we discuss about context model for cooperation system in ubiquitous environment and describe context ontology focused on interaction process with general context. Moreover, we propose cooperation architecture based on context ontology supporting ontological processes. This architecture introduces community as an organization of elements solving common problem and provides context-aware mechanism within community.

Physical interfaces have been proposed as a way to realize natural interactions with ubiquitous computing environments. The successful design of such interfaces requires design approaches that integrate aspects of our world which are usually treated separately in traditional system development approaches. This paper describes a design approach based on Gibson concept of affordance. We demonstrate an experimental method for studying object affordance and show how it can be applied to the design of a concrete physical interface artefact.

Object tracking is an important feature of the ubiquitous society and also a killer application of wireless sensor networks. Nowadays, there are many researches on object tracking in wireless sensor networks under practice, however most of them cannot effectively deal with the trade-off between missing-rate and energy efficiency. In this paper, we propose a dynamic clustering mechanism for object tracking in wireless sensor networks. With forming the cluster dynamically according to the route of moving, the proposed method can not only decrease the missing-rate but can also decrease the energy consumption by reducing the number of nodes that participate in tracking and minimizing the communication cost, thus can enhance the lifetime of the whole sensor networks. The simulation result shows that our proposed method achieves lower energy consumption and lower missing-rate.

This paper proposes DPS-MAC (Divided Preamble Sampling-MAC), a carrier sense Medium Access Control (MAC) protocol with ultra low power operations for wireless sensor networks. Due to battery-operated computing and sensing devices in wireless sensor networks, the development of MAC protocols that efficiently reduce power consumption is an important issue. DPS-MAC is a novel energy efficient protocol based on sampling preamble segments that have preamble offsets and next hop addresses. By using these parameters, DPS-MAC can overcome the drawbacks of long fixed preambles and overhearing in B-MAC. The divided preamble sampling technique provides not only very low power consumption when the channel is idle, but also a chance to avoid overhearing without additional overheads compared to B-MAC. This technique makes it possible to extend the lifetime of the wireless sensor networks that contain a large number of nodes. The experiment results show that DPS-MAC protocol reduces much energy consumed by receiving compared to B-MAC and can support various features of wireless sensor networks efficiently.

In this paper, we introduce a service framework using the VPW (Virtual Personal World) and its associated service conflict resolution algorithm based on the VPW. In this new service model, users can manage their own spaces that include all things to support their services. Previous service frameworks are focused on designing adaptive software applications by using heterogeneous devices and softwares. However, these service frameworks have some problems to provide adaptive services to the users. To resolve the limitation that is caused by multi-user environment, we propose a service conflict resolution algorithm. Therefore, our proposed service framework based on the VPW can support more user-centric services, reduce the number of service conflicts, and improve the number of conflict resolutions. The experiment result shows that almost 40% resolution rate can be achieved comparing with the location based system and 30% work time can be improved comparing with the location based system.

In ubiquitous networks, Mobile Nodes (MNs) may often suffer from performance degradation due to the following two reasons: (1) reduction of signal strength by an MN’s movement and intervening objects, and (2) radio interference with other WLANs. Therefore, quick and reliable detection of the deterioration of a wireless link condition is essential for avoiding the degradation of the communication quality during handover. In our previous works, we focused on a handover decision criterion allowing MNs to maintain communication quality and stated the problems of existing decision criteria. Furthermore, we showed the effectiveness of the number of frame retransmissions through simulation experiments. However, a comparison between signal strength and the number of frame retransmissions could not be examined due to the unreliability of signal strength in simulations. Therefore, in this paper, by employing FTP and VoIP applications, we compare signal strength and the number of frame retransmissions as a handover decision criterion with experiments in terms of (1) and (2) in a real environment. Finally, we show the problems of signal strength in contrast to the effectiveness of the number of frame retransmissions as a handover decision criterion.

In this paper we propose an operating system design for Ubicomp applications that are implemented on embedded sensor platforms. The OS provides support for both periodic sensor sampling and sequential appliation logic. Core component is a lightweight real-time runtime system guaranteeing predictable real-time behavior of periodic sampling processes. The design utilizes a novel method, called buffer feedback scheduling (BFS), to adapt the system under unpredictable workload. Processes are automatically coordinated and expensive hardware accesses are reduced when the feedback indicates that the results do not contribute to the application. Real-time behavior is guaranteed throughout the entire runtime. Theoretical analysis and implementation in a Ubicomp application study on the Particle Computer sensor platform demonstrate a significant performance step-up when utilizing BFS.

Sentient artefacts are our everyday life objects augmented with various sensors for value added services. We have been exploiting these artefacts for perceiving user context and proving proactive context aware services while keeping their natural appearances and roles intact. From our observations of these artefacts we have identified two passive advantages besides their explicit value added functionalities. The first is a profile-based participation of the artefacts in the application scenarios thus making the artefacts generalized, independent of the applications and reusable in various scenarios. The second one is the role of the static artefacts in identifying the location of the mobile artefacts thus eliminating the requirement of any dedicated location infrastructure for proactive services. In this paper we have discussed these two issues with illustrations of our findings.

Ubiquitous applications usually involve highly interactive context data management. Traditional general-purpose programming languages are not sufficient for use in this domain, as they do not have the capability to manage such data effectively. We have developed a scenario-based programming language that we call ‘PLUE (Programming Language for Ubiquitous Environment)’, which is a Java-based prototyping language for ubiquitous application development. PLUE supports ECA (event-condition-action) rules and finite state automata-based (FSA-based) interactive responses to dynamic situations. In addition, PLUE programmers are able to manage heterogeneous data with a uniform view of path expressions. We have implemented PLUE on top of CAMUS (Context-Aware Middleware for Ubiquitous Robotic Companion System), a framework for context-aware applications that was originally developed for network-based robots.

Creating applications based on data from individual sensor nodes is typically a two-tiered process: Firstly, a (potentially large) number of sensor nodes is deployed in order to gather comprehensive datasets. After analyzing the collected data, algorithms are then installed on the individual nodes and iteratively fine-tuned using a collect-and-analyze procedure. This approach is not only time consuming, but also prone to errors: the two separate steps of data collection and data analysis complicate algorithm development; the absence of programming abstractions in embedded systems programming languages often introduces hard-to-detect runtime errors; and the lack of modern integrated development environments (IDEs) does not allow for quick trial-and-error prototyping. To mitigate those effects, we have developed

JSense

– a hardware abstraction layer for modern sensor nodes that allows for Java-based access to all sensor and actuator controls. It supports an IDE-based centralized development cycle with real-time debugging of a particular sensor environment, as well as the use of not-yet-available sensor and actuator hardware on each node, such as positioning information. Using

JSense

, designers of sensor-based environments can quickly try out a combination of situations and observe in real-time the data collection processes of their nodes, while programmers are able to prototype applications in their favorite Java-IDE in a hardware independent fashion, even taking into account not-yet-deployed node hardware.

The ubiquitous computing environments requires integration of a variety of current and future wired and wireless networking technologies to support seamless computing and communication environments for user applications. WLAN also need to be a part of a seamless communication infrastructure, as they can be used either as a wireless extension of wired networks or peer-to-peer networks. This paper proposes a new methodology to estimate the number of competing stations in an IEEE 802.11 network. Due to nonlinear nature of measurement model, an iterative nonlinear filtering algorithm, called the “Central Difference Filter” (CDF), is employed. The CDF can provide a better alternative to nonlinear filtering than the conventional extended Kalman filter (EKF) since it avoids errors associated with linearization. This approach shows both high accuracy as well as prompt reactivity to changes in the network occupancy status. Specially, our proposed algorithm is more improved performance in non saturated conditions than the EKF. Numerical results show that it provides a more viable means for estimation of the number of competing stations in IEEE 802.11 network than estimators based on the EKF.

This paper presents an intelligent vehicle safety system, constructed by exchanging emergency-related information such as urgency stop, traffic accident, and obstacles between vehicles. In the majority of vehicle safety communication applications, an emergency message is propagated in the form of broadcasts. However, this causes numerous problems, such as massive radio collision, multi-hop propagation and fast relay. This paper presents a selective message forwarding method by proposing a stem and branch structure based on prospected positions. The logical structure makes it possible to assign a representative vehicle to relay an emergency message among the different level of wireless radio coverage. The proposed scheme improves the efficiency of message transmission with the selective priority for forwarding messages by distance. The structure also limits the number of broadcasting hops. The proposed method is evaluated using a network simulator.

Vertical handoff will be essential for the next generation heterogeneous wireless networks. We propose an

Adaptive Vertical Handoff Decision Scheme

called

UbiComm

to avoid unbeneficial handoffs in the integrated WiBro and WLAN networks. If the mobile node (MN)’s velocity is high and moving pattern is irregular, more unnecessary handoffs can occur. Therefore, MN’s velocity and moving pattern are the important factors of our handoff decision scheme. In order to avoid unbeneficial handoff the

U

biComm adjusts the dwell time adaptively, and it also predicts the residence time in the target network. In addition,

U

biComm’s adaptive dwell timer makes a MN receive service of a better network as long as possible. The simulation results show that the reduction of unnecessary handoffs proposed in

U

biComm improves the MN’s throughput.

Hierarchical Mobile IPv6 (HMIPv6) guarantees to reduce handoff latency, because the MN only registers the new addresses at mobility anchor point (MAP) when the MN moves around access routers (ARs) in the same MAP domain. HMIPv6 still has packet loss problem when the MN moves from one MAP to another. In this paper, a novel location update scheme which further reduces signaling traffic for location update by employing virtual MAP (VMAP) on top of overlapped MAP in HMIPv6, is proposed. This proposed scheme significantly improves performance compared to HMIPv6, in terms of location update rate per user. Also it makes the mobile nodes (MNs) moving around the boundary ARs of adjacent MAP’s become to move within a VMAP. It is certain that this scheme reduce the network resources efficiently by reason of removing the location update. In conclusion, this scheme greatly reduces the packet loss and delay, due to Inter-MAP handoff not occurring.

WPANs (Wireless personal area networks) vary greatly by nature with regard to requirements such as data rate, coverage, subscriber volume, and supported mobile velocity. Mobile devices with multiple network interfaces are very common and to make these multi-mode devices communicate with any WPAN device at anytime, anywhere, a framework for heterogeneous wireless networks is essential. Therefore, we propose an integration framework that handle heterogeneous WPAN protocols simultaneously in a unified ways. In this paper, we explain the requirements and detailed design of our framework, and then present the emulation test results.

Lots of routing protocols have been proposed in the literatures to overcome several challenges in ad hoc networks. The fundamental point we consider in this paper is that most of such protocols are generally based on the assumption that mobile nodes are functionally equivalent to each other in computing power and memory space. Moreover, all of the mobile nodes are required to use a common routing protocol to communicate with each other. However, such assumptions do not reflect the real world, even further the future oriented ubiquitous world. The ubiquitous paradigm requires networking technologies to support the heterogeneity including various capabilities to compute, amounts of storage, radio interfaces, patterns of mobility and others. In real scenario, for instance, some nodes may not want to relay packets for others owing to their power constraints. Also there might be nodes employing different routing protocols in a single communication zone. To cover some of these cases, this paper proposes a simple but efficient approach called HRPC (Heterogeneous Routing Protocol Coordinator) that works well in our previously proposed MANET architecture. HRPC is not a stand-alone routing protocol but a coordinating module for support bridging functionality between heterogeneous routing protocols in MANET. This paper also gives HRPC implementation and its demonstration results, where DYMO and OLSR routing protocols are used as an exemplified scenario to evaluate the operability of HRPC.

DynaMoNET is designed for a multihomed moving network which has various MANETs interconnected through WPAN interfaces. Each MANET has a mobile router, e.g. a cellular phone which can work as a root MR in DynaMoNET. The essential technologies for DynaMoNET include handover protocol in heterogeneous networks, network switchover algorithm considering multiple decision factors, root MR election process based on token-based algorithm, fast root MR discovery algorithm and fault avoidance mechanism to support reliable Internet connectivity. In addition DynaMoNET handles multiple HAs model, i.e. (n, n, n) model. Finally the system design of a mobile router is suggested in order to make relationship between the key components.

IPv6 stateless address auto-configuration scheme is very useful in mobile environment like as Mobile IPv6 network. Using address auto-configuration, mobile node can configure global address and access Internet when it moves away from home network. In some mobile environment, however, DAD procedure for auto-configuration might not be applicable. For example, in mobile ad-hoc network, DAD procedure using timeout cannot reliably detect the absence of a duplicate address. In this paper, we propose new IPv6 addressing mechanism and address auto-configuration scheme using proxy. A proxy, in our scheme, can allocate global addresses to other IPv6 nodes. Our solution uses a stateful approach, but the solution is lighter and faster than DHCPv6. Our scheme is flexible and efficient, as shown by simulation.

Developing an ideal routing protocol that satisfies various wireless sensor network applications is difficult due to their different requirements. In this paper, we propose a parametric routing protocol that considers performance parameters such as time, reliability, and energy to satisfy various circumstances of wireless sensor network applications. Based on a geographic algorithm, the proposed protocol calculates each node’s routing costs for three performance parameters and negotiates to select the next node. The framework supports adaptive service as well as scalability because the mechanism requires only neighboring nodes’ information. The experiment shows that the proposed protocol provides adaptive services for various sensor networks applications.

This paper evaluates the effect of a block Forward Error Correction (FEC) algorithm’s symbol size on power consumption in wireless sensor networks (WSN). The WSN channels exhibit frequent bursty errors with a high average Bit Error Rate (BER) due to low transmission power, random deployment, and moving intermediate objects obstructing WSN communications. For resisting against the bursty errors, WSN would adopt a block FEC algorithm that restores more tainted bits than other kinds of FEC algorithms as errors become burstier since it recovers errors symbol-by-symbol not bit-by-bit. Even when the same amount of bits are allocated for FEC code, different FEC symbol size meaning different number of FEC symbols vary the packet error rate and the transmission energy over a given WSN. They also affect the computational energy since their decoding and encoding complexities depend on the number of FEC symbols per packet. The analytical evaluation based on long-term sensor traffic traces indicates that the appropriate FEC symbol size saves a sensor node’s energy consumption by up-to 85 % comparing to other sizes.

One of the major requirements for new wireless sensor networks is to extend the lifetime of the network. Node scheduling techniques have been used extensively for this purpose. Some existing approaches rely mainly on location information through GPS devices for designing efficient scheduling strategies. However, integration of GPS devices with sensor nodes is expensive and increases the cost of deployment dramatically. In this paper we present a location-free solution for node scheduling. Our scheme is based on a graph theoretical approach using

minimum dominating sets

. We propose a heuristic to extract a collection of dominating sets. Each set comprises of a group of working nodes which ensures a high level of network coverage. At each round, one set is responsible for covering the sensor field while the nodes in other sets are in sleep mode. We evaluate our solution through simulations and discuss our future research directions.

An effective approach for energy conservation in wireless sensor networks is scheduling sleep intervals for redundant nodes, while guaranteeing continuous services with the remaining active nodes. For the sensor networks to operate successfully, the active nodes must maintain sensing coverage. In this paper, we present the design and analysis of a novel scheme that can adaptively configure a network to achieve entire coverage. This work differs from the existing coverage maintenance schemes in several ways: 1) We propose a novel scheme of configuring active nodes from all given nodes; 2) We provide simple coverage model to calculate the number of active nodes satisfying the coverage; and 3) Our scheme performs well without any location information, so it can be easily used in hostile environments which requires a fast network configuration. We demonstrate the capability of our scheme in terms of coverage configuration through simulations.

As the spectral recourses are limited, the spectrum efficiency is more important for the design of wireless sensor network. In this paper, a spectral correlation based detection method for spectrum sensing is proposed to increase the spectrum efficiency in wireless sensor network. The signal of interests is modeled as a cyclostationary random process, therefore, the spectral correlation function could be used, and a squared coherence function is used as the decision criterion. Simulations are made based on the 4-FSK signal. The results verify that the proposed method outperforms the conventional energy detection.

As key technologies of sensor network have been deployed to various applications, such as ubiquitous computing and mobile computing, the importance of sensor network were recognized. Because most sensors are battery operated, the constrained power of sensors is a serious problem. If data containing small error is tolerable to users, the sensor data can be sampled discretely. An efficient power conserving algorithm is presented in this paper. By observing the trend of the sensor data, it was possible to predict the time that exceeds the specified maximum error. The algorithm has been applied to various sensor data including synthetic data. Compared to the regular sensors which do not adapt the proposed algorithm, the proposed sensors in this paper shows that the sensor’s life time can be increased up to six folds within the range of 1% tolerable data error.

In this paper, we present a unique mechanism that enables seamless interconnection and scalability of the interface between the sensor platform layer and the middleware layer in pervasive computing. The disadvantages of using one-to-one hard coding for various device drivers and firmware for interfacing sensor platform and middleware layers are primarily a lack of flexibility and scalability when the system changes environments. The encapsulation of access point of the sensor platform layer provides developers and designers with an effective way of interconnecting and scaling up with diverse and various kinds of sensors and actuators. In this paper, we define encapsulation of the access point of the sensor platform layer. The module for encapsulating the detected sensor data is called ‘context representer’. The context representer converts detected sensor data to raw context. For interconnecting with the sensor platform layer and the middleware layer, ‘entity manager’ module in the middleware layer for each entity of context representer is defined. The entity manager is for transforming the raw context to a general context. Then, the middleware is able to interact with context-aware applications according to the reasoning with context in the middleware layer, upper layer requests and lower layer status, without being influenced by a change of sensor node and sensor platform environments. We present the encapsulation mechanism and entity manager for efficient interfacing and scalability of contextaware applications and compare them with other approaches.

We describe our data collection and results on activity recognition with wearable, coin-sized sensor devices. The devices were attached to four different parts of the body: right thigh and wrist, left wrist and to a necklace on 13 different testees. In this experiment, data was from 17 daily life examples from male and female subjects. Features were calculated from triaxial accelerometer and heart rate data within different sized time windows. The best features were selected with forward-backward sequential search algorithm. Interestingly, acceleration mean values from the necklace were selected as important features. Two classifiers (multilayer perceptrons and kNN classifiers) were tested for activity recognition, and the best result (90.61 % aggregate recognition rate for 4-fold cross validation) was achieved with a kNN classifier.

For the Ubiquitous computing system, a well designed computer interface should have the ability of making the user feel comfortable so as to encourage user performance. In order to do this, systems should be able to identify the emotional state of the user. Thus, the identification of the user’s emotional state is of importance in developing context-related devices, which offer the optimal feedback to the user depending on the user’s emotional state. Those devices require portability and continuous measurement for daily use, which is particularly essential for an ubiquitous healthcare device. In this paper, the portable device for bi-emotional state identification was designed using heart rate variability (HRV) extracted from beat-to-beat photoplethysmography (PPG) waveforms. A portable wrist-band type PPG measuring device was equipped with a Bluetooth communication interface to provide mobility. HRV was estimated from smoothed differentiated PPG waveforms by the absolute value of the successive beat-to-beat interval difference. Two emotional states are artificially induced by a composed video clip, and then validated by the self-assessment Manikin method. The designed device was then applied to a respiration training device to adjust the balance level of the autonomous nervous system throughout the respiration pager, whose level changes in proportion to the estimated ratio between negative and positive emotional states. Experimentation using 19 male and six female participants demonstrated the feasibility of a ubiquitous emotional feedback control device.

In 2003, Huang et al. proposed an authenticated key exchange (AKE) protocol for secure communications between a sensor and a security manager based on elliptic curve cryptography in a self-organizing sensor network. In 2005, Tian et al., however, showed that a security manager will learn the long-term private key of a sensor after having one normal run of the protocol with the sensor. Then they proposed an improvement which solves the problem. The current paper, however, demonstrates neither protocol provides perfect forward secrecy, and presents an improved protocol in order to address this problem. Not only is the computational cost of the proposed AKE protocol less than that of both protocols, the number of steps in communication is one fewer than conventional protocols and only needs three steps.