- Human-human relationships as relevant to companions (D.2.1)
- Report on first experiments on different modalities of interaction with companions (D.3.1)
- Framework for the perception of user actions (D.3.2)
- Design and implementation vision system (D.3.3)
- Limited Language and Speech Recognition and Synthesis System (D.3.4)
- Facial and body expressions for companions (D.3.5)
- Formulation of Initial Memory Model (D.4.1)
- Initial memory models and results (D.4.2)
- Enhanced Memory Model (D.4.3)
- Theories and mechanisms that support dynamic models of Social Relations (D.5.1)
- First specification of the components of the Companions’ Minds (D.5.2)
- Architecture development with action-selection mechanism (D.5.3)
- Foundations of Embodied Companions (D.6.1)
- Design of Embodied Companions (D.6.2)
- Ethogram of dog-human interaction (D.7.1)
- Communicative interaction in specific collaborative tasks (D.7.2)
- Test protocols for humans' attitudes toward robots and dogs (D.7.3)
- Test protocols for comparing dog and robot interaction with humans (D.7.4)
- Foundations of migrating companions (D.8.1)
- First skeleton companions (D.9.2)
- First version of integrated companions (D.9.3)
- Updated Integration Architecture (D.9.4)
- Preparatory studies and ethics for companion design (D.10.1)
- First generation companion issues and ethics (D.10.2)
Human-human relationships as relevant to companions (D.2.1)
In this deliverable psychological theories and models on human-human social relationships are reviewed regarding their relevance for human-companion relationships. Thereby, a special focus is set on long- term relationships. Moreover, studies conducted within the LIREC project/WP2 regarding psychological and ethological aspects of human-companion relationships are described and linked to theory.
Report on first experiments on different modalities of interaction with companions (D.3.1)
Overview of the different types of modalities for communication with companions from two perspectives: perceiving the user and expressions in the companion, and report on the first experiments with companions.
Read: First Experiments
Framework for the perception of user actions (D.3.2)
This document describes the work that LIREC partners have been conducting towards the development of a joint framework for the perception of the user actions using several interaction modalities. These include user’s face and body expression, limited speech recognition and contextual information of the interaction.
Design and implementation vision system (D.3.3)
This deliverable aims to provide a survey of the vision- based social perception abilities that are currently under investigation in the LIREC project. These are the result of the collaboration among different partners in LIREC and they are applied to different scenarios. We provide a description of the first prototypes of these abilities, highlighting relevant issues and open questions, and we discuss them with reference to different scenarios.
Limited Language and Speech Recognition and Synthesis System (D.3.4)
This report focuses on the tools available for speech interaction in environments with limited resources, especially portable devices or platforms such as those envisioned as embodiments for the agents in the LIREC scenarios. In addition, language issues and how they impact the project are also discussed. These include, but are not limited to, the choice and availability of resources for a given natural language, for instance, European Portuguese or Scottish English.
Facial and body expressions for companions (D.3.5)
Non-verbal communication serves very important functions in human social relationships (Patterson, 1990). As stated in D2.1, non-verbal communication is useful, for example, to regulate social interactions (e.g., by turn-taking or proximity mechanisms) and is considered a very efficient and genuine way of expressing emotions. For these reasons, non-verbal communication is also a very important requirement in artificial companions that interact with users in social settings. There is evidence that affective behaviour plays an important role in long-term human-computer relationships (Bickmore, 2005).
Formulation of Initial Memory Model (D.4.1)
Formulation of initial model linking long-term autobiographic memory and working memory to associated human processes of remembering and forgetting.
Initial memory models and results (D.4.2)
The objective of this report is twofold: describe the first experiment
of the initial memory model prototype proposed in the deliverable D4.1. and propose a refined memory model including remembering, forgetting and generalisation mechanisms taking into account migration, particular ethical issues and related ethological studies.
Enhanced Memory Model (D.4.3)
In this deliverable, we are going to provide more information on each of these aspects by describing the details of implemented mechanisms (mainly retrieval mechanisms) and giving justification/implementation proposals to other chosen but yet to be implemented mechanisms (forgetting and generalisation mechanisms). The link between WP4 and WP5 is more strongly defined through discussion on how the retrieval mechanisms are being used by the cognitive processes in the Agent Mind. Other issues such as means for data capture and storage; and ontology as well as database integration are also discussed.
Read: Enhanced Memory Model
Theories and mechanisms that support dynamic models of Social Relations (D.5.1)
Survey on theories and systems that support dynamic models of social relations: social intelligent agents, emotions and personality, theory of mind.
First specification of the components of the Companions’ Minds (D.5.2)
First specification of the companion’s mind architecture, focusing on building models of others and using knowledge about social relations and emotions in order to create and maintain long-term relations with users.
Architecture development with action-selection mechanism (D.5.3)
This deliverable fully describes the final specification of the architecture proposed for the companion. It extends the draft specification presented in Deliverable D5.2, by focusing on how the several internal components (Theory of Mind, Social Relations, Emotional State,
and Autobiographic Memory) influence the action-selection mechanism in order to create socially intelligent behaviour. One additional requirement was to enable all architectural components to deal with migration between embodiments without loss of consistency, thus
we will provide a brief description on how this is achieved.
Foundations of Embodied Companions (D.6.1)
This deliverable report presents a selective review and discussion of embodiment relating to LIREC definitions of artificial companions. This provides a rationale for WP6 partners respective Phase 1 plans and research aims and objectives regarding embodiment of LIREC companions. To facilitate these objectives new robots will be developed for use in studies and relevant showcase scenarios. Progress, main findings and outcomes for Phase 1 (M1-15) are summarised.
Design of Embodied Companions (D.6.2)
This deliverable report presents a review and progress report on the design of the two new companion robots, CHARLY and FLASH, currently under development by UH and WRUT for use in studies and relevant showcase scenarios. Design concepts and factors, progress, and outcomes for the first part of Phase 2 (M16 - M22) are summarised.
Ethogram of dog-human interaction (D.7.1)
The idea of robotic social companion is quite old, and various fictions have found their way to the popular media. Sometimes these ideas have been combined even with some form of extra-terrestrial intelligence which was also not the result of a carbon-based biological evolution. Getting in a communicative (and collaborative) contact with other “intelligent” creatures has always excited the public and the scientists. The famous experiments with the language-trained apes were also driven partly by such curiosity: If apes could learn human-like language then they may “tell” us about their (inner) life. Despite all such efforts no such cross-species talk has taken place. The apes did not acquire human-like linguistic skills and no encounter with “other intelligent” beings has taken place.
Communicative interaction in specific collaborative tasks (D.7.2)
To our knowledge LIREC is the first ICT project in which the research on robotic behaviour is paralleled by ethological research. Although this situation is very advantageous, there is no experience on ways how ethology can and should inform researchers working on artificial agents. In LIREC we have planned for three phases. In the 1 st phase we are concentrating on developing the basic tools for behavioural research, in the 2 nd phase we will provide sophisticated behaviour analysis which is believed to make companion robots more believable, and finally, in the 3 rd phase we will aim to compare the functioning robots with living companions under similar conditions.
After reviews on the attitude of humans towards robots and dogs, we present several sets of data. We compared university students' and dog owners' attitudes toward robots in comparison to dogs. People playing with AIBO and a puppy were interviewed and characteristics of the most popular dog breeds were also investigated. Results emphasise that detailed behavioural description is needed about dogs' faithfulness (attachment), individuality (personality) and emotions in order to enhance companionship between robots and humans. Further, we investigated personality traits of owner-dog dyads, and found several correlations. We developed a behavioural test battery for providing detailed description of more than one hundred dogs' behaviour in scenarios which are related to the showcases.
In the current deliverable we are starting with the presentation of a study on humans’ emotion-attribution to dogs and robots concluding that humans are able to attribute emotions to a robot based on simple action patterns borrowed from dog’s behaviour, and that the ethogram based on dogs’ social behaviour might be a useful tool for programming the expressive behaviour of robotic companions. We are also presenting new results regarding the role of personality in the behaviour test developed for providing detailed description of dogs' behaviour in scenarios which are related to the showcases. Finally, we provide a brief description of our plans on the adaptation of the above mentioned behaviour tests to
the studies of HRI.
Foundations of migrating companions (D.8.1)
This deliverable report presents a selective review, discussion, mythologies and results of migration companion studies relating and conducted for Phase 1 work in WP8 for the LIREC project.
First skeleton companions (D.9.2)
This document describes the first skeletons for the long-term companion agents and detailed communication interface for 3 level architecture.
First version of integrated companions (D.9.3)
This document demonstrates through the use of online videos, the first versions of integrated companions produced
Updated Integration Architecture (D.9.4)
After a review of the requirements for a LIREC integration architecture in D 9.1 and an architecture design specification in D 9.2, this deliverable 9.4 presents the actual implementation of the LIREC integration architecture. D 9.4 is primarily a software deliverable and this document is an accompanying overview of the delivered software. It outlines the motivation for creating an integrated architecture and the requirements for the architecture, before describing the architecture, its components and its application within the LIREC showcases.
Preparatory studies and ethics for companion design (D.10.1)
The objectives in Task 10.1.1 have been to conduct human-centred studies to explore lasting values of long-term relationships, existing human interests related to companion user experience, and ethical issues for companion design. D10.1 covers three main areas: preparatory studies of companions in everyday, natural and human-centred contexts; ethics issues specific to companion technology; and user-centred design of companions; all with the important focus on companion technology in an everyday, realistic and human-centred context.
First generation companion issues and ethics (D.10.2)
This deliverable reports on initial studies on interaction design for, and ethical issues in, long-term relationships with companion technologies. Described are studies to guide and understand companion technology, design guidelines and frameworks, as well as ethical considerations that arise from human- robot interaction.