People spend the majority of their time in buildings which are a major contributor to overall energy consumption. Both buildings and their occupants can therefore be viewed as an opportunity but also a genuine challenge to energy saving. A classical approach in this context is to build knowledge models which are based on expert knowledge to model the energy behavior of buildings. However, designing knowledge models requires a great deal of expert time, significant knowledge of building specifications and a thorough physical background. Once designed, the calibration phase based on sensor data from the building is also a long and complex task. Finally, the slightest change in the environment, such as adding a ventilation system, would necessitate a parameters adjustment or even a re-design of the model. At the same time, while the occupants are not able to analyze and apprehend the energetic behavior of their housing, due to these complexities, they can decide on their comfort and expectations. Unfortunately, human behavior models are likewise tough to build, maintain and are still a challenging task. This research presents a new occupant-centered approach to enhance his indoor comfort using historical data on both the building environment and the behavior of its occupant, thanks to case-based reasoning (CBR). We evaluate the effectiveness of our technique on over one and a half years of real-world data collected from an office building. Results show that the proposed approach improves the comfort of the occupant.

Movements are an important part of robot design and we need dedicated tools to design them. As previous research has shown, 3D animation techniques are of great use to animate a robot. However, most robots don't benefit from an animation tool and therefore from animation artists knowledge. We present an open-source robot animation software to address this issue and describe some principles for designing a robot animation software derived from the specificity of robot as a medium.

Mobile robot companions are service robots that are mobile and designed to share our living space. For such robots, mobility is essential and their coexistence with humans adds new aspects to the mobility issue: the first one is to obtain appropriate motion and the second one is interaction through motion. We encapsulate these two aspects in the term Human-Robot Motion (HRM) with reference to Human-Robot Interaction. The long-term issue is to design robot companions whose motions, while remaining safe, are deemed appropriate from a human point of view. This is the key to the acceptance of such systems in our daily lives. The primary purpose of this paper is to explore how the psychological concept of attention can be taken into account in HRM. To that end, we build upon an existing model of attention that computes an attention matrix that describes how the attention of each person is distributed among the different elements, persons and objects, of his/her environment. Using the attention matrix, we propose the novel concept of attention field that can be viewed as an attention predictor. Using different case studies, we show how the attention matrix and the attention field can be used in HRM.

This paper addresses the problem of learning situation models for providing context-aware services. Context for modeling human behavior in a smart environment is represented by a situation model describing environment, users, and their activities. A framework for acquiring and evolving different layers of a situation model in a smart environment is proposed. Different learning methods are presented as part of this framework: role detection per entity, unsupervised extraction of situations from multimodal data, supervised learning of situation representations, and evolution of a predefined situation model with feedback. The situation model serves as frame and support for the different methods, permitting to stay in an intuitive declarative framework. The proposed methods have been integrated into a whole system for smart home environment. The implementation is detailed, and two evaluations are conducted in the smart home environment. The obtained results validate the proposed approach.

Abstract: This paper deals with the problem of implementing a context model for a smart environment. The problem has already been addressed several times using many different data‐ or problem‐driven methods. In order to separate the modelling phase from implementation, we first represent the context model by a network of situations. Then, different implementations can be automatically generated from this context model depending on user needs and underlying perceptual components. Two different implementations are proposed in this paper: a deterministic one based on Petri nets and a probabilistic one based on hidden Markov models. Both implementations are illustrated and applied to real‐world problems.

This paper addresses the problem of detecting interaction groups in an intelligent environment. To understand human activity, we need to identify human actors and their interpersonal links. An interaction group can be seen as basic entity, within which individuals collaborate in order to achieve a common goal. In this regard, the dynamic change of interaction group configuration, i.e. the split and merge of interaction groups, can be seen as indicator of new activities. Our approach takes speech activity detection of individuals forming interaction groups as input. A classical HMM-based approach learning different HMM for the different group configurations did not produce promising results. We propose an approach for detecting interaction group configurations based on the assumption that conversational turn taking is synchronized inside groups. The proposed detector is based on one HMM constructed upon conversational hypotheses. The approach shows good results and thus confirms our conversational hypotheses.

This article lies within the interactive virtual stories telling scope and proposes the use of fuzzy cognitive maps as a tool to model emotional behavior of virtual actors improvising in free interaction within the framework of a "nouvelle vague" scenario. We show how fuzzy cognitive maps can be delocalized on each agent level to model autonomous agents within a virtual world. We describe the implementation carried out, starting from work in cognitive psychology and illustrate it by an improvisation between a shepherd, a dog and virtual sheep.

We describe our work on multiagent systems in a distributed virtual reality environment. Our main goal in the LI2 laboratory is to make interactive and co-operative prototyping for well defined systems. Users and software agents evolving in our virtual worlds can interact, communicate and modify their behavior. The article focuses on the way our agents communicate in a "natural language" by exchanging information and knowledge. For this purpose we have developed a communication language based on "speech acts theory" and on agent communication languages (ACL). The Knowledge Query Manipulation Language (KQML) is the ACL we have chosen to develop high level communication among "intelligent agents". To validate our developments and thus the KQML specifications, we have tested them in a contract net application. Contract net applications use communication mechanisms based on offers, bids, contracts and so on, which can be designed by KQML message exchanges.

This paper deals with an application of fuzzy-ART self-organizing neural classifier to adaptive categorization of the perceptual space of a mobile robot. The aim of the research is to develop a learning system for reactive locomotion control in an unknown, cluttered environment. A qualitative description of the proposed categorization technique for a trial-and-error learning paradigm is given. Experimental results show that the method of control is efficient, when learning starts from scratch, as well as after some major disturbances of an already experienced system.>