In recent years, Japan has grappled with a rapidly aging population, leading to pressing issues in emergency medical care and an uptick in ambulance services. Our research team, collaborating with the Kobe City Fire Department, aims to address this by devising a predictive model for strategic deployment of medical services. By analyzing extensive EMS data and population demographics, we aim to forecast long-term emergency transport demand accurately, bypassing machine learning algorithms. The results of the forecast revealed regional disparities in future emergency demand, providing indicators for optimal deployment of emergency resources and strategic planning for emergency services.

In pursuit of realizing Society 5.0, this study explores efficient development methods for services tailored to individual user needs. The rapid evolution of digital devices and the diversification of user demographics have led to swiftly changing service needs, necessitating increased personalization. This research focuses on developing technologies that enable service development based on a deep understanding of specific user needs. By utilizing a virtual agent (VA)-based interactive need extraction system developed in previous research, and leveraging the Sharing Up cycling Cases with Context and Evaluation for Efficient Software Development System (SUCCEED System), we aim to automate the extraction of user needs and provide insights to developers. We propose an interactive need extraction method for novel, undeveloped services and a method for obtaining development cases based on these needs, thereby contributing to the efficiency of personalized service development approaches.

The increase in single-person households and the onset of “COVID-19 frailty” due to COVID-19 have become problems. Our research group focuses on technology support and develops virtual agents using speech recognition technology. In this service, we examined how to promote self-care by having the agent listen to the older person and increase conversation opportunities. However, this service has the problem that it cannot increase opportunities for dialogue with family and friends, which leads to mutual support. In this study, we aim to propose and implement a service called “Easy Video Chat Service” that increases the opportunities for conversation with others through a virtual agent listening service for the elderly person at home.

Japan is facing the challenge of an aging society with a significant increase in the number of dementia patients, making it an urgent social issue. Ac-cording to the Ministry of Health, Labour and Welfare, it is estimated that by 2040, the population of elderly individuals aged 65 and above will reach 35.3 % of the total population, with over 7 million of them being dementia patients. Against this backdrop, measures for dementia prevention and support for elderly dementia patients are deemed necessary. Memory impairment is one of the core symptoms of dementia, and current strategies include recording and presenting information by caregivers and using memory aid tools such as notepads for rehabilitation. However, these methods have their limitations, prompting the need for new memory aid techniques that reduce caregiver burden and are sustainable for dementia patients. Therefore, recording daily activities and preserving them as docu-ments is proposed as a method for memory assistance. This study aims to verify the effectiveness of this technique by setting evaluation criteria, conducting self-recordings of daily activities, creating explanatory texts by multiple individuals, and subjectively eval-uating texts generated by recognition AI. Based on the results, the study examines the extent to which the texts generated by recognition AI are useful for memory assistance and provides insights accordingly.

In this study, we describe the design and implementation of a smart service for automated management of hybrid meeting spaces. Hybrid formats, where participants interact both online and offline, are becoming common in modern meeting rooms. To meet this new need, we have developed a remote task execution service using VNC and WebDriver. This service automates the launching of online services and website operations during meeting preparations, greatly reducing the time and effort required by the user. Specifically, the service provides functions to automatically execute multiple remote control tasks, such as launching Zoom and a web site for meeting minutes. In this paper, we confirm the actual operation of the system through a prototype implementation in our laboratory.

Software testing is being leveraged in programming education for automated assessment of programming assignments. When using software testing in programming education, program specifications are provided as unit or integration tests, and students create programs that pass these tests. Although this method has various advantages, such as ensuring objective program specifications and automating the operation check, it also has many disadvantages. For example, detecting innovations, such as original specifications and functional extensions by an individual student, is difficult. The purpose of this research is to automatically detect functional differences among student programs in programming education using tests. In our proposed method, automatic test generation is applied to student programs, and the generated tests are mutually executed for other student programs. Furthermore, we classify the tests based on the execution path to obtain sets of tests that are capable of detecting functional differences.

In this study, we aim to realize an assistive technology that can present necessary information to elderly people with cognitive concerns or dementia in a way that is adaptable to their home life. To achieve this goal, we propose ALPS (Assisted Living by Personalized Speaker), a system that presents appropriate information according to various locations and times in the home. We installs IoT speakers with motion sensors at key locations in the home, and by linking them to ECA(Event-Condition-Action) rules in the cloud, ALPS provides information based on location and time in voice. We implemented a prototype of the proposed ALPS and conducted a case study of two elderly people. As a result, it was found that by defining ECA rules for each problem, the system can present information according to the individual’s lifestyle.

In this study, we conducted a multifaceted analysis of heatstroke cases using the emergency transported big data in Kobe City, and discovered the characteristics of heatstroke incidents in Kobe City in 2020 that differed from previous years. As a result of the analysis, it was found that the peak period of WBGT in 2020 was later than usual, and it was found that the peak period of WBGT is later than usual in 2020, and the occurrences of heatstroke in 2020 is characterized by an increase in the occurrences of heatstroke in people over 65 years old and outdoors, and a decrease in the occurrences of heatstroke in people under 65 years old and indoors.

The personal adaptation of services, in which a system provides services according to the preferences and needs of individual users, is a key to the realization of the emerging Society 5.0. The personal adaptation of systems has been implemented through personal settings by users. However, it is very difficult for users who are not familiar with ICT to manually define the settings that meet their needs. In this paper, we therefore propose a new notion of smart system: Smart System with Interactive Personalization (SSIP). In SSIP, the system and the user have continuous and interactive conversations during the system operation. In the dialogue, the user tells his or her needs and the system introduces its functions. In this way, the user and the system understand each other and dynamically co-create personal settings. In this paper, in order to define SSIP, we present three functional requirements that the system must meet. We also characterize SSIP from the quality viewpoint by quality in use of the international standard SQuaRE. Finally, as a case study, we take the Mind Monitoring Service being developed by our group, and apply the proposed SSIP framework to individual adaptation of the incentive provision feature.

Currently, Japan is entering a super-aged society. It is difficult for the elderly at home to deal with their problems in their lives by self aid and mutual aid. So this study aims to propose and implement Compass4SL, a problem and solution sharing service for the elderly at home. We design and implement the system according to the following steps. First, we find the use cases and entities. Then, we design the overall system structure with a layered architecture. Finally, we implement the system as a Web application based on the obtained design using Java and Spring Boot framework.

We have been developing "Facial expression sensing service" for emotional analysis and quantitative evaluation of care based on subtle facial movements and conducted a preliminary experiment about its practicality. In this research, focusing both obtaining facial expression data and searching for an efficient care method, to elderly people can activate themselves and relieve their stress, we have developed "Video player service" that can easily play videos and automatically collect facial expression data. After developing the service, we have asked people who engage in elderly care to try it and obtained feedback. As a result, we received favorable comments for the usefulness of the service, and we were able to get facial expression data for four people.

To support sustainable in-home long-term care, it is essential to monitor mental states of elderly people at home, and to encourage their ability of self-care. However, many challenges exist in practice, including limitations of human interventions, sensor-based monitoring, the daily recording and externalization of mental states. In the previous research, we have proposed Mind Monitoring Service, which aims to monitor mental states and promote self-care of elderly people at home. In the proposed service, a chatbot asks a user specific questions to acquire his/her mental state. Based on the answers, the service assesses the mental state. In this research, we develop a new feature of weekly feedback. The feature automatically reviews answers of past one week, and sends advice to improve the current situation. We conduct an experiment to evaluate the effectiveness. Through the experiment, it was confirmed that the weekly feedback encouraged self-care consciousness and increased motivation of subjects.

To support in-home long-term care, we are studying techniques of Mind Sensing, which externalizes internal states of elderly people as words through conversations with agents or robots. We have previously developed a prototype system of Mind Sensing, integrated with an activity recognition system and an LINE chatbot. However, the system was tightly coupled with the fixed systems, it was difficult to add or change the setting of questions from the chatbot to individual elderly people.

In order to support sustainable in-home long-term care, it is essential to monitor mental states of elderly people at home, as well as to encourage their ability of self-care. However, the technical challenges include the limitations on human interventions and sensor-based monitoring, as well as daily recording and externalization of mental states. In this research, we propose Mind Monitoring Service, which aims to monitor mental states and promote self-care of elderly people at home. In the proposed service, an agent asks a user specific questions to acquire his/her mental state. Based on the answers, the service then assesses the mental state and sends feedback. We implement a prototype service, and evaluate the feasibility of the service through a preliminary experiment. The results show that data characterizing mental states of individual subjects was obtained successfully, and that some subjects externalized their minds by feedback from the service.

The emerging deep learning technology is a promising means for context recognition with multimedia data. We are interested in using the deep learning with images for context recognition in smart homes. In the home context recognition, the room layout, the environment, and the contexts to be recognized are different from one household to another. Therefore, a unique recognition model is required for every different household. For this, if we take a naive approach that uses the deep learning directly, a huge amount of labeled images are required, which is practically impossible for general households. The goal of this research is to develop an image-based context recognition method that is affordable at home. In the proposed method, we exploit a cognitive API which performs general image recognition, and retrieve the information within the image as text. By using the text as features, we classify the context with ordinal supervised machine learning. Compared with the expensive approach with deep learning, the proposed method uses generic image recognition of the cognitive API, and light-weight machine learning. As a result, the context recognition customized for every household can be achieved with much less effort.

It is known that the disturbance of daily life rhythm leads to chronic disease. Hence, it is important for everyone to keep a healthy rhythm. Owing to emerging technologies of smart phones and IoT, many studies and products recognizing personal daily activities exist. However, there is few research to evaluate if the life rhythm, characterized by the daily activities, is good or bad for the person. As a result, it is hard for individuals to understand what to be their own healthy life rhythms, and how to improve the current situation. To cope with the problem, we are developing a system that quantitatively assesses user's life rhythm based on the daily activity log and the self-assessment of QoL (Quality of Life). In this paper, we present a method of measuring user's life rhythm by analyzing sleep and eating log data. We then construct a personalized quantitative assessment model of life rhythm using regression analysis. We conduct a preliminary experiment in an actual apartment. Based on the derived model, we find personalized advice on daily activities to maintain healthy life rhythm of the resident.

In our previous research, we proposed a security information service, called PRISM, which personalizes the incident information based on living area of individual users. PRISM computes the severity of a given incidents based on distance, time, and type. It then visualizes the incident with the severity on a heat map. In this paper, we extend the functionality of PRISM, in order to analyze street crimes around living area in more details. More specifically, we add three new features to PRISM: showing a past heat map, showing a heat map focused on specified type of incidents, and showing statistics of incidents for every type. Using the extended features, we visualize the dynamic transition of street crimes in a specific area and the whole region. The visualization also reveals the ecology of wild boars. Finally, we also show that PRISM can be used to compare different districts by statistics of street crimes.

Recently, many studies about Activities of Daily Living (ADLs) recognition have been conducted, which can be applied to many real-life, human-centric problems such as eldercare and healthcare. In our previous work, we proposed an ADLs recognition system based on non-intrusive environment sensing for people in One-person Household (OPH). However, the proposed recognition system did not perform well, the micro-averaged and macro-averaged precision of most of the recognition models was only around 60%. In order to improve the quality of the system, in this article, we propose a new ADLs recognition system by integrating environment sensing and Bluetooth Low Energy (BLE) beacon technology and evaluate the new version of the ADLs recognition model by comparing the experimental data collected from a real resident in OPH.

Location information is expressed by various formats that depend on services. Location information is divided into two categories: absolute location information (such as latitude/longitude and address), and relative location information (such as distance and direction). Each service that utilizes location information defines location information system individually. Therefore, sharing location informations between the services is difficult. Consequently, reusability of location information decreases. Then, we consider new common location information system, which can be converted from location information of various systems and expresses location more flexible. In this study, we propose probabilistic location information, which expressed as a combination of area and existence probability. Moreover, we propose the algorithm which calculates probabilistic location information based on geographic coordinate information and pass-by information (PLPA-GP).

Local governments in Japan recently provide security information services for residents, which deliver regional incident information using Email or Web. However, since the conventional services usually provide “one-for-all” information. users tend to miss important incidents within the flood of information. In this paper, we propose a new security information service, called PRISM (Personalized Real-time Information with Security Map). For given incident information and user's living area, PRISM first computes severity of the incident, based on distance, time, and type of the incident. It then visualizes the incidents with the severity on a heat map. Thus, PRISM provides real-time personalized information adapted to individual situation of users. To illustlate the feasibility, we implement PRISM as a Web application using Hyogo Bouhan Net, and Kobe city facility open data.

Pervasive sensing technologies are promising for increasing one-person households (OPH), where the sensors monitor and assist the resident to maintain healthy life rhythm. Towards the practical use, the recognition of activities of daily living (ADL) is an important step. Many studies of the ADL recognition have been conducted so far, for real-life and human-centric applications such as eldercare and healthcare. However, most existing methods have limitations in deployment cost, privacy exposure, and inconvenience for residents. To cope with the limitations, this paper presents a new indoor ADL recognition system especially for OPH. To minimize the deployment cost as well as the intrusions to user and house, we exploit an IoT-based environment-sensing device, called Autonomous Sensor Box (SensorBox) which can autonomously measure 7 kinds of environment attributes. We apply machine-learning techniques to the collected data, and predicts 7 kinds of ADLs. We conduct an experiment within an actual apartment of a single user. The result shows that the proposed system achieves the average accuracy of ADL recognition with more than 88%, by carefully developing the features of environment attributes.

To realize indoor environmental sensing, which is a key technology of providing smart services in smart city, with low cost, our research group has proposed a small IoT device named sensor box. In the previous sensor box, however, it is difficult to deploy for the smart city with some problems. In this paper, we propose an indoor environment sensing service using autonomous sensor box to adapt the previous sensor box for the smart city. To confirm the effectiveness of proposed service, we deploy autonomous sensor boxes on practical indoor environments.

To fix various dependencies of application development using pass-by detection by a mobile device, we propose Pass-by Framework that handles data with standardization. In this study, we evaluate effects of performance of pass-by detection by differences in methods of implementation the sonar of Pass-by Framework. Therefore, we develop pass-by application using Bluetooth Low Energy as a first effort. We then conduct evaluation experiments for confirmation of change pass-by detection behavior depends on the difference of parameters.

In recently years, many care robots have received a lot of attention to help elderly people. However existing care robots have difficult to adopt personalization. For instance, some programmers have to customize robot program to meet needs of each elderly. Even if a care robot which has a feature of machine learning, it takes a long time to learn a preference for each elderly. In this paper, our goal is to deploy a smart care service integration agent that provides a personalization and integration for each elderly people. Our proposed service consists of three essential components, Virtual Care Giver (VCG), Virtual Care Personalizer (VCP) and Care Template. VCG is a robot agent, where executes care tasks in each home. The VCG is offered care tasks based on care template which Virtual Care Personalizer (VCP) generates. Virtual Care Personalizer (VCP) manages and generates personalization of care tasks the on cloud. Moreover, we deploy Care Template on the cloud which enables to provide the basic care tasks. To demonstrate the feasibility, we consider three kinds of usecase scenarios for two persona people.

As typically seen in Smart City, emerging technologies enable large-scale environment sensing using IoT devices deployed in a wide area. From the viewpoint of cost and efficiency, infrastructure of the large-scale environment sensing should be shared by multiple applications, with dynamically adapting the sensing behavior for different purposes. To achieve this, the infrastructure must implement a clever method that can command and control a lot of IoT devices in good order. To implement such multi-purpose large-scale environment sensing, we introduce an analogy of military system. Specifically, we propose a mission-oriented sensing with army hierarchy, where individual IoT devices and their dynamic purposes are regarded as soldiers and missions, respectively.

A number of reminder systems have been developed to help elderly people with dementia. However, the existing reminder systems lack the awareness of human context, the sympathetic human-machine interaction, and the flexibility of personal adaptation. To cope with the limitations, we are currently studying a new reminder service for people with dementia. Specifically, we exploit a BLE-based indoor positioning system to capture the current location and context of the patient. We then use a virtual agent system for rich interactions. Finally, we develop a schedule management system for personalized reminders. To integrate these heterogeneous systems, we re-design and deploy the systems as three services with Web-API: Location Service, Agent Service, and Schedule Service. These services are loosely integrated by Coordinator Service, based on the service-oriented architecture, In this paper, we first present the system architecture, and then discuss the key idea to implement the services. We also demonstrate “reminder at the entrance” as a practical scenario of the proposed services. In order to evaluate the Agent Service, which is a key component of proposed service, we have conducted the a preliminary experiment with 17 people with dementia.

The emerging indoor positioning systems (IPS) enable indoor location-aware applications (InL-App) within indoor space where GPS cannot reach. In most conventional systems, however, IPS and InL-App are tightly coupled, where one system cannot reuse location data or operation of other systems. This fact yields expensive development cost and effort of InL-App. To cope with the problem, this paper propose a cloud-based integration framework, called CIF4InL. With a common data model, CIF4InL integrates indoor location data obtained from heterogeneous IPS. It then provides application-neutral API for various InL-Apps. To evaluate the practical feasibility, we integrate two different IPS (RedPin and BluePin) using CIF4InL, where the applications transparently access the indoor locations gathered by two different IPS. Since CIF4InL allows the loose coupling between IPS and InL-Apps, it significantly improves reusability of indoor location information and operation.

This paper proposes a unified locating service, KULOCS, which horizontally integrates the heterogeneous locating services. Focusing on technology-independent elements [when], [where] and [who] in location queries, KULOCS integrates data and operations of the existing locating services. In the data integration, we propose a method where the time representation, the locations, the namespace are consolidated by Unix time, the location labels and the alias table, respectively. Based on possible combinations of the three elements, we then derive API for the operation integration.

Managing a home server within individual house is a major obstacle to practical use of home network system (HNS). Delegating the home server to a cloud is a promising approach. However, the conventional multitenant SaaS-based solution has the following risks among different households: security/privacy violation, failure propagation and insufficient resource demand. In this paper, we propose a novel cloud-based architecture for the home network system that achieves security isolation, fault isolation and resource isolation. Specifically, we first create a virtual home server for every household using IaaS. On top of every virtual home server, we then create additional virtual machines, each of which contains a single service or application. Finally, using dynamic scaling, we allocate resources needed for individual virtual home servers. Based on the idea, we construct the proposed architecture by three layers: HNS Lite, House Cloud and Service Provider.

The goal of this paper is to find an answer that how remote monitoring sensor should be accurate. To achieve the goal, we propose three methods, generalization by three-actor model, design the algorithm of the three-actor and development of RMS simulator. With the three-actor model, we can generalize RMS by interactions among three actors. As the second step, we design the algorithms that how to work the actor in RMS. So we could express how often the elderly become ill. Moreover, using the developed simulator, we could simulate with many patterns of conditions. The result of simulations shows that if the accuracy of the sensor is greater than 0.9990, then the RMS has much more detectionPower.

Modern cloud services and machine-to-machine (M2M) systems provide various kinds of data via various Web services. Implementing context-aware services integrating such global data are promising in various applications. However, it has been challenging to manage heterogeneous contexts and services defined in various Web services. To cope with this, we design a framework, called RuCAS, which systematically manages every context-aware service in form of ECA (Event-Condition-Action) rule. We also develop RuCAS platform, which publishes API of RuCAS as Web service. Using the RuCAS platform, users can define their own contexts with various Web services (e.g., information service, sensor services, networked appliances, etc.). Based on the defined contexts, they can create ECA rules to define custom context-aware services. To support users, We also implement a GUI front-end of RuCAS platform, called RuCAS.me. RuCAS.me supports users even if the users are non-expert. A case study in a real home network system demonstrates practical feasibility of RuCAS platform and RuCAS.me. The contribution of this paper is to provide design and implementation details of RuCAS, by which one can fully understand systematic management of context-aware services with Web services.

In order to achieve intuitive and easy operations for home network system (HNS), we have previously proposed user interface with virtual agent (called HNS virtual agent user interface, HNS-VAUI). The HNS-VAUI was implemented with MMDAgent toolkit. A user can operate appliances and services interactively through dialog with a virtual agent in a screen. However, the previous prototype heavily depends on MMDAgent, which causes a tight coupling between HNS operations and agent behaviors, and poor capability of using external information. To cope with the problem, this paper proposes a service-oriented framework that allows the HNS-VAUI to provide richer interaction. Specifically, we decompose the tightly-coupled system into two separate services: MMC Service and MSM service. The MMC service concentrates on controlling detailed behaviors of a virtual agent, whereas the MSM service defines logic of HNS operations and dialog with the agent with richer state machines. The two services are loosely coupled to enable more flexible and sophisticated dialog in the HNS-VAUI. The proposed framework is implemented in a real HNS environment. We also conduct a case study with practical service scenarios, to demonstrate effectiveness of the proposed framework.

A home network system (HNS) provides value-added services for home users by networking house-hold appliances and sensors. In the conventional architecture, the HNS appliances and services are tightly coupled. It is therefore difficult for users to freely choose their favorite appliances and services. In this paper, we propose a new HNS architecture that accommodates multi-vendor services by extensively using cloud technologies. The new architecture manages individual HNS operations and data as standard services within the cloud. The vendor services must go through the cloud to access the HNS. Thus, loose coupling among the HNS and services can be achieved. As a result, the proposed architecture realizes more flexible HNS beneficial for both users and vendors.

With the spread of Machine-to-Machine (M2M) systems and cloud services, various kinds of data are available through Web services. A context-aware service recognizes a real-world context from such data and behaves autonomously based on the context. However, it has been challenging to manage contexts and services defined on the heterogeneous and distributed Web services. In this paper, we propose a framework, called RuCAS, which systematically creates and manages context-aware service using various Web services (e.g. information services, sensor services, networked appliances, etc.). The framework describes every context-aware service by an ECA (Event-Condition-Action) rule. For this, an event is a context triggering the service, a condition is a set of contexts to be satisfied for execution, and the action is a set of Web services to be executed by the service. Thus, every context-aware service is simply managed in a uniform manner. Since the RuCAS is published as a Web service, it is easy for various applications to reuse and integrate created contexts and services. As a case study, RuCAS is applied to creating context-aware services in a real home network system.

Smart city provides various value-added services by collecting large-scale data from houses and infrastructures within a city. To use such large-scale raw data, individual applications usually take expensive computation effort and large processing time. To reduce the effort and time, we propose Materialized View as a Service (MVaaS). Using the MVaaS, each application can easily and dynamically construct its own materialized view, in which the raw data is cached in an appropriate format for the application. Once the view is constructed, the application can quickly access necessary data. In this paper, we design a framework of MVaaS specifically for large-scale house log, managed in our smart-city data platform Scallop4SC. In the framework, each application first specifies how the raw data should be filtered, grouped and aggregated. For a given data specification, MVaaS dynamically constructs a MapReduce batch program that converts the raw data into a desired view. The batch is then executed on Hadoop, and the resultant view is stored in HBase. We conduct an experimental evaluation to compare the response time between cases with and without the proposed MVaaS.

To support mashup of heterogeneous lifelog services, we have previously implemented the lifelog common data model (LLCDM). The previous LLCDM was implemented with MySQL, where various types of application-specific data (e.g., numeric values, text, JSON or XML) were all stored in a column in a schemaless text format. Any query with application-specific data had to be managed by individual applications. It had also a scalability issue as the data size grew.

On the software development PBL (SDPBL), the implementation of firmly-fused development environment for students and monitoring environment for teachers are required in order to succeed in education. We have proposed the service, named "DaaS BADER" in compliance with demands from practical teachers to decrease the cost for preparation and maintenance of unified exercise environment and to monitor the progress of projects by teachers. In this paper, we have reported knowledge and information obtained by practical SDPBL and feedback contents for a student or group given by monitoring environment. Then, we have discussed the effectiveness of DaaS BADER from these results.

Smart city provides various value-added services by collecting large-scale data from houses and infrastructures within a city. However, it takes a long time for individual applications to use and process the large-scale raw data directly. To reduce the response time, we use the concept of materialized view of database. For a given requirement of an application, the proposed method constructs a materialized view for caching the application-specific data. In this paper, we especially develop a method that uses MapReduce for large-scale power consumption data stored in HBase KVS. We conduct an experimental evaluation to compare the response time between cases with and without the materialized view. As a result, the proposed method with materialized view is effective especially when application repeatedly access the same data, or when the application-specific data is derived from a large set of raw data.

We have proposed a framework named SO-MSR: service-oriented mining software repository, which applied service oriented architecture to MSR. Following the SO-MSR, we have developed a web service, named MetricsWebAPI, for metrics calculation from a variety of software repositories and a variety source codes. In this paper, we develop and propose Metrics Viewer, which is client of Metrics Viewer and is a web application to support personal process improvement. Metrics Viewer provides an interactive user interface for repository file exploring. Moreover the Metrics Viewer visualizes change of source code metrics to support overhead view of personal process. End user can improve their development activities based on software repository data without MSR specific knowledge by using Metrics Viewer. We have conducted a pilot study to evaluate the effect of proposed system for personal process improvement.

This paper summarized experiences of practical software development exercise in PBL style activities from organizer perspective. The object of this PBL is nurturing advanced knowledge as advanced information and communication technology (ICT) engineers. A main pillar of this report is trace the 5-year history of three sub environments such as development, development support and teaching support environment which are badly need to hold our software development PBL, from problem and its solutions viewpoint.

In this paper, we propose an algorithm to improve the performance of the MU-LAW PNLMS algorithm (MPNLMS) for non-sparse impulse responses. Although the existing MPNLMS algorithm was recently proposed to achieve optimal proportionate step size for both large and small tap weights, it converges even slower than conventional NLMS algorithm for dispersive channels. The proposed approach adaptively estimates the sparsity of the impulse response to be identified. Then the estimation of this sparsity is incorporated into the IPNLMS algorithm to accordingly adjust its parameters. Simulation results verify the effectiveness of the proposed algorithm.

A novel structure with direct inversion of a multiple- input multiple-output (MIMO) system is proposed in this paper. Based on this structure, 2 adaptive algorithms, LMS-like and Affine Projection Algorithm (APA), are proposed to directly obtain an accurate estimate of the inverse of the plant system. The proposed algorithms have faster convergence speed than the FxLMS algorithm. Furthermore, this structure does not necessitate any a prior information of plant system and it can trace the fluctuation of the plant. It is shown through simulations that the proposed method over performs the FxLMS algorithm in convergence speed so it is much applicable to real world environments than the FxLMS algorithm.