Modelling Multimodal Dialogues for Social Robots Using Communicative Acts.

Social Robots need to communicate in a way that feels natural to humans if they are to effectively bond with the users and provide an engaging interaction. Inline with this natural, effective communication, robots need to perceive and manage multimodal information, both as input and output, and respond accordingly. Consequently, dialogue design is a key factor in creating an engaging multimodal interaction. These dialogues need to be flexible enough to adapt to unforeseen circumstances that arise during the conversation but should also be easy to create, so the development of new applications gets simpler. In this work, we present our approach to dialogue modelling based on basic atomic interaction units called Communicative Acts. They manage basic interactions considering who has the initiative (the robot or the user), and what is his/her intention. The two possible intentions are either ask for information or give information. In addition, because we focus on one-to-one interactions, the initiative can only be taken by the robot or the user. Communicative Acts can be parametrised and combined in a hierarchical manner to fulfil the needs of the robot's applications, and they have been equipped with built-in functionalities that are in charge of low-level communication tasks. These tasks include communication error handling, turn-taking or user disengagement. This system has been integrated in Mini, a social robot that has been created to assist older adults with cognitive impairment. In a case of use, we demonstrate the operation of our system as well as its performance in real human-robot interactions.

Social Robotics in Therapy of Apraxia of Speech.

Apraxia of speech is a motor speech disorder in which messages from the brain to the mouth are disrupted, resulting in an inability for moving lips or tongue to the right place to pronounce sounds correctly. Current therapies for this condition involve a therapist that in one-on-one sessions conducts the exercises. Our aim is to work in the line of robotic therapies in which a robot is able to perform partially or autonomously a therapy session, endowing a social robot with the ability of assisting therapists in apraxia of speech rehabilitation exercises. Therefore, we integrate computer vision and machine learning techniques to detect the mouth pose of the user and, on top of that, our social robot performs autonomously the different steps of the therapy using multimodal interaction.

Detecting and Classifying Human Touches in a Social Robot Through Acoustic Sensing and Machine Learning.

An important aspect in Human-Robot Interaction is responding to different kinds of touch stimuli. To date, several technologies have been explored to determine how a touch is perceived by a social robot, usually placing a large number of sensors throughout the robot's shell. In this work, we introduce a novel approach, where the audio acquired from contact microphones located in the robot's shell is processed using machine learning techniques to distinguish between different types of touches. The system is able to determine when the robot is touched (touch detection), and to ascertain the kind of touch performed among a set of possibilities: stroke, tap, slap, and tickle (touch classification). This proposal is cost-effective since just a few microphones are able to cover the whole robot's shell since a single microphone is enough to cover each solid part of the robot. Besides, it is easy to install and configure as it just requires a contact surface to attach the microphone to the robot's shell and plug it into the robot's computer. Results show the high accuracy scores in touch gesture recognition. The testing phase revealed that Logistic Model Trees achieved the best performance, with an F-score of 0.81. The dataset was built with information from 25 participants performing a total of 1981 touch gestures.

Augmented Robotics Dialog System for Enhancing Human-Robot Interaction.

Augmented reality, augmented television and second screen are cutting edge technologies that provide end users extra and enhanced information related to certain events in real time. This enriched information helps users better understand such events, at the same time providing a more satisfactory experience. In the present paper, we apply this main idea to human-robot interaction (HRI), to how users and robots interchange information. The ultimate goal of this paper is to improve the quality of HRI, developing a new dialog manager system that incorporates enriched information from the semantic web. This work presents the augmented robotic dialog system (ARDS), which uses natural language understanding mechanisms to provide two features: (i) a non-grammar multimodal input (verbal and/or written) text; and (ii) a contextualization of the information conveyed in the interaction. This contextualization is achieved by information enrichment techniques that link the extracted information from the dialog with extra information about the world available in semantic knowledge bases. This enriched or contextualized information (information enrichment, semantic enhancement or contextualized information are used interchangeably in the rest of this paper) offers many possibilities in terms of HRI. For instance, it can enhance the robot's pro-activeness during a human-robot dialog (the enriched information can be used to propose new topics during the dialog, while ensuring a coherent interaction). Another possibility is to display additional multimedia content related to the enriched information on a visual device. This paper describes the ARDS and shows a proof of concept of its applications.

A multimodal emotion detection system during human-robot interaction.

In this paper, a multimodal user-emotion detection system for social robots is presented. This system is intended to be used during human-robot interaction, and it is integrated as part of the overall interaction system of the robot: the Robotics Dialog System (RDS). Two modes are used to detect emotions: the voice and face expression analysis. In order to analyze the voice of the user, a new component has been developed: Gender and Emotion Voice Analysis (GEVA), which is written using the Chuck language. For emotion detection in facial expressions, the system, Gender and Emotion Facial Analysis (GEFA), has been also developed. This last system integrates two third-party solutions: Sophisticated High-speed Object Recognition Engine (SHORE) and Computer Expression Recognition Toolbox (CERT). Once these new components (GEVA and GEFA) give their results, a decision rule is applied in order to combine the information given by both of them. The result of this rule, the detected emotion, is integrated into the dialog system through communicative acts. Hence, each communicative act gives, among other things, the detected emotion of the user to the RDS so it can adapt its strategy in order to get a greater satisfaction degree during the human-robot dialog. Each of the new components, GEVA and GEFA, can also be used individually. Moreover, they are integrated with the robotic control platform ROS (Robot Operating System). Several experiments with real users were performed to determine the accuracy of each component and to set the final decision rule. The results obtained from applying this decision rule in these experiments show a high success rate in automatic user emotion recognition, improving the results given by the two information channels (audio and visual) separately.