Ebook: Behaviour Monitoring and Interpretation – BMI

The annual workshop on Behaviour Monitoring and Interpretation (BMI) was launched in 2007. The workshop is co-located with the German conference on Artificial Intelligence, and hence, receives much attention in the research community investigating intelligent means for behaviour monitoring and interpretation. The edition of the workshop in 2009 focused on the topic of well-being to reflect the significant interest in this research direction. The current volume consists of extended versions of selected contributions from this workshop as well as other invited articles covering major research themes in this field.

This volume aims to offer state-of-the-art contributions in the application area of well-being. The notion of well-being has been treated in diverse disciplines such as engineering, sociology, psychology and philosophy. With this book a perspective is offered to the latest trends in this field from a few different viewpoints. Well-being is indeed an omnipresent concept that reaches out to a myriad of aspects of our daily lives. In addition to supporting a healthy lifestyle, the concept of well-being extends to the selections involving the type of the environment we live in, the interactions we have with other humans, and the practices we engage in to achieve our plans for future. Well-being concerns us in our daily life, and hence, plays a fundamental role at all times and places. This fact in turn needs to be taken into account when designing ubiquitous computing technologies that pervade our life. With the presented articles the book provides a survey of different research projects that aim to address the many influential aspects of well-being that are considered in today's designs or play an essential role in the designs of the future.

The present volume is the second BMI edition that is published by IOS Press. We are looking forward to receiving feedback from our readership in order to better plan and prepare for future BMI workshop editions. We are thankful to all authors that have contributed to this volume and to the programme committee of the BMI workshop, who provided the authors with valuable reviews.

March 2011

Björn Gottfried and Hamid Aghajan

The notion of well-being has been treated in as diverse disciplines as engineering, sociology, and philosophy. This volume offers a perspective to the latest trends in research and development in well-being of humans across a subset of the involved disciplines. This introductory chapter provides an overview of the subsequent chapters and discusses the challenges shaping the future research agenda for technologies that support well-being.

This chapter provides an informal review of the opportunities provided by Information Communication Technology (ICT) for improving and maintaining well-being among older people. The manner in which ICT might improve well-being is considered across three domains; social networks (broadening and deepening connections, facilitating social contribution, providing indirect monitoring), cognitive function (maintaining and enhancing function, supporting cognitive impairment), and health (preventing and managing disorders, rehabilitation, enhancement of health, improvements in health care). The barriers and opportunities that exist are identified as falling into four categories; psychological, technological, costs (financial and human), and issues of privacy and confidentiality. Based on exploration of these barriers and opportunities, recommendations regarding possible solutions are made.

The paper discusses, from a philosophical perspective, the use of Pervasive Computing for supporting the well-being of humans. Philosophy can help computer scientists who are working in this area, to resolve conceptual problems, to take up a considered position concerning ethical issues, and to define clear technological principles. Drawing on the philosophical tradition of pragmatism, questions addressed in the paper are: How is the technological research programme of Pervasive Computing to be defined? How are humans and systems of Pervasive Computing to be modelled as intelligent agents? How do we understand the behaviour of agents as intentional? How is the interaction of humans and systems of Pervasive Computing to be analyzed? How can we define the well-being of intelligent agents? How should systems of Pervasive Computing be designed if they ought to support the well-being of humans?

Once a smart home system moves to a multi-resident situation, it becomes significantly more important that individuals are tracked in some manner. By tracking individuals the events received from the sensor platform can then be separated into different streams and acted on independently by other tools within the smart home system. This process improves activity detection, history building and personalized interaction with the intelligent space.

Historically, tracking has been primarily approached through a carried wireless device or an imaging system, such as video cameras. These are complicated approaches and still do not always effectively address the problem. Additionally, both of these solutions pose social problems to implement in private homes over long periods of time. This paper introduces and explores a Bayesian Updating method of tracking individuals through the space that leverages the CASAS platform of pervasive and passive sensors. This approach does not require the residents to maintain a wireless device, nor does it incorporate rich sensors with the social privacy issues.

As a result of the aging societies in the western world, the impact of dementia, with its characteristics like disorientation and obliviousness is becoming a significant problem to an increasing amount of persons and the health system. To enable such dementia patients to regain a self determined life, we have developed a mobile orientation system called KopAL that assists dementia patients in every day problems, like remembering appointments, keeping track within their familiar surroundings as well as informing caretakers in critical situations, with a focus on minimal operational costs and a speech based human computer interface.

While easy-to-use was one of KopALs requirements, the system itself uses and experiments with new technologies in the field of Mobile Ad Hoc Networks (MANETs), VoIP, and embedded systems. Further, KopAL is an interdisciplinary project. It is developed at Potsdam University within the Assisted Living Initiative of the Institute of Computer Science [3]. The working group Applied Computational Linguisics of Potsdam University cooperates to integrate speech generation and recognition on embedded systems. Psychologies from the University of Jena evaluate the project from the beginning: in helping to figure out the real user demands and a suited user-interface.

Chronic disease is identified as one of the main causes of all deaths worldwide and adversely affects the economy through huge healthcare costs and human capital losses. These outcomes can be ameliorated if patients and their healthcare providers adhere to the care plans developed for the patients' chronic condition. This paper proposes a framework for adherence support based on (1) continuous monitoring and recognition of possible deviations from plan; (2) determination of the cause of the possible deviation; and (3) generation of an appropriate to assist patients and their healthcare providers avoid plan failure. The framework and findings are generalised to include agents in any domain that adopts similar behaviour monitoring and intervention mechanisms. A cost/benefit analysis is performed under different settings using a number of theoretical and simulation studies. The theoretical analysis serves to: (1) establish the general principles of intervention; and (2) provide a basis for validating the simulation results. The agents are modelled as Belief Desire Intention (BDI) agents and different modes of failure (“deficits”) and interventions are characterised in these terms. The simulation code was shown to produce results compatible with the theoretical analysis and has the potential to be used for other settings and domains for which a theoretical analysis is difficult.

The growing number of people adopting a sedentary lifestyle these days creates a serious need for effective physical activity promotion programs. Often, these programs monitor activity, provide feedback about activity and offer coaching to increase activity. Some programs rely on a human coach who creates an activity goal that is tailored to the characteristics of a participant. Throughout the program, the coach motivates the participant to reach his personal goal or adapt the goal, if needed. Both the timing and the content of the coaching are important for the coaching. Insights on the near future state on, for instance, behaviour and motivation of a participant can be helpful to realize an effective proactive coaching style that is personalized in terms of timing and content. As a first step towards providing these insights to a coach, this chapter discusses results of a study on predicting daily physical activity level (PAL) data from past data of participants in a lifestyle intervention program. A mobile body-worn activity monitor with a built-in triaxial accelerometer was used to record PAL data of a participant for a period of 13 weeks. Predicting future PAL data for all days in a given period was done by employing autoregressive integrated moving average (ARIMA) models on the PAL data from days in the period before. By using a newly proposed categorized-ARIMA (CARIMA) prediction method, we achieved a large reduction in computation time without a significant loss in prediction accuracy in comparison with traditional ARIMA models. In CARIMA, PAL data are categorized as stationary, trend or seasonal data by assessing their autocorrelation functions. Then, an ARIMA model that is most appropriate to these three categories is automatically selected based on an objective penalty function criterion. The results show that our CARIMA method performs well in terms of PAL prediction accuracy (~9% mean absolute percentage error), model parsimony and robustness.

Mobile robots are already applied in factories and hospitals, merely to do a distinct task. It is envisioned that robots assist in households soon. Those service robots will have to cope with several situations and tasks and of course with sophisticated human-robot interactions (HRI). Therefore, a robot has not only to consider social rules with respect to proxemics, it must detect in which (interaction) situation it is in and act accordingly. With respect to spatial HRI, we concentrate on the use of non-verbal communication. This chapter stresses the meaning of both, machine movements as signals towards a human and human body language. Considering these aspects will make interaction simpler and smoother.

An observational study is presented to acquire a concept of spatial prompting by a robot and by a human. When a person and robot meet in a narrow hallway in order to pass by, they have to make room for each other. But how can a robot make sure that both really want to pass by instead of starting interaction? This especially concerns narrow, non-artificial surroundings. Which social signals are expected by the user and on the other side, can be generated or processed by a robot? The results will show what an appropriate passing behaviour is and how to distinguish between passage situations and others. The results shed light upon the readability of signals in spatial HRI.

In this chapter we discuss the system requirements and components of an adaptive smart-home service system. To achieve adaptivity in providing services to the user, the system needs to 1) sense the activity and state of the user and 2) customize service to the user's profile. To achieve this, three functional parts are developed and described. In the first part, we present behavior analysis of the user in a home environment based on multi-camera vision processing. In the second part, the concept of user profile is introduced and hierarchical reinforcement learning is employed as a technique to learn the user profile dynamically. The third part of the chapter discusses how to employ the user profile to control services to maximize user comfort and utility. Future work is discussed in conclusion.