Robota: Clever toy and educational tool

https://doi.org/10.1016/S0921-8890(02)00380-9Get rights and content

Abstract

Therapeutic and educational applications of robots have created a demand for robots showing a number of social skills. These skills include the capacity to imitate, to learn from demonstration, to interpret gestures and to recognize speech. Robot toys are an ideal platform to investigate the potential and limitations of human–robot social interactions.

This paper presents Robota, a mini-humanoid doll-shaped robot. Robota is used in an introductory robotics class at the undergraduate level. The class offers an introduction to different tools necessary for building human–robot social interactions. Through a series of hands-on projects, students learn how to use vision and speech processing and how to design learning algorithms. The goal of each project is to create an educational and entertaining game for normal and disabled children.

Introduction

During the past 10 years, researchers have identified a number of social skills that robots ought to be provided with. These include the ability to recognize others, to interpret gestures and verbal expressions, and to recognize and to express emotions [5]. One of the most fundamental and, perhaps, the most difficult skill to program is the ability to learn from others. This is known as social learning in the study of natural systems [17]. Social learning includes learning by imitation, that is learning through the observation of others’ behavior, and social facilitation, where the mere presence of others enhances learning of tasks.

Teaching skills, whether cognitive, verbal or motors, is the very motor of human society, and relies importantly on human ability for social learning. Human ability to learn from socially interacting with others is so much taken for granted that the few people who are impaired in this ability (such as people with autism [7]) have enormous difficulty to adapt to society. Providing robots with the ability to learn from observing human behavior has, thus, become an important topic of research in robotics. The rational is that the robot’s ability to imitate would offer a natural means of programming the robot, that would require none or minimal programming competence on the user’s part. Robot learning by imitation, also known as robot programming through demonstration, finds its use for a large number of tasks, such as object manipulation [13], [16], learning body motions [2], [8], [15], navigation [5], [10] and learning communication skills [4].

Robot toys are the most likely application of social robots in the near future [9]. Being cheap to build and likeable by a large part of the population, robot toys offer an ideal platform to test algorithms for social learning. Robota, a doll-shaped robot toy, was developed with the goal investigating what social skills would be required for a human-shaped robot to engage in believable social interactions [1]. Robota is currently being used in pilot studies with normal and cognitively impaired children, and with children with autism. The studies aim at developing Robota as a complementary educational tool that would guide the children through games that rehearse school material. In studies with autistic children, Robota is particularly interesting to educators as a means to systematically evaluate the child’s social competences.

While the above-mentioned studies have been reported elsewhere [3], [6], the present paper focuses on the novel application of Robota as a laboratory platform in an introductory class to robotics taught at the undergraduate level. The class uses a set of 10 Robota robots. The class revolves around lab sessions during which the students learn to program the robot’s micro-controller and the robot–PC interface, using both vision and speech synthesizing. The lab sessions aim at the development of games suitable for normal and disabled children.

Section 2 describes briefly the hardware of the Robota robots and the applications of Robota as a toy for children. Section 3 describes the application of Robota in an undergraduate robotics class. 4 Discussion: key features required for social robots, 5 Conclusion conclude this paper with a discussion of the choice of providing Robota with human-like features, vision and speech interfaces and how these can, in some cases, be important components for building social skills in a robot.

Section snippets

The Robota robots

Robota is a small humanoid robot, 45 cm high, with motors to drive its legs, arms and head.1 Robota electronics consists of 3 superimposed boards (PIC 16F870 4 MHz, PIC 16F870 16 MHz, and 68376 Motorola), that drive 6 motor outputs, 24 sensor entries (16 analog and 8 digital) and a set of sensors: 4 infra-red emitter/receivers, 2 light detectors, 2 pyroelectric sensors, 6 switches and 2 electro-magnets. Robota is

Robota course

This section focuses on the latest application of Robota in an undergraduate introductory class to robotics, at the Computer Science department at the University of Southern California (USC). The class is entitled “Mechatronics: Programming Humanoid Robots”. The class was given for the first time at USC during the spring semester 2001. It gathered the maximum allowed of 30 students: 28 males, 2 females. It is a 3-unit class. It will be given again in the 2003 fall semester, and then regularly

Discussion: key features required for social robots

Work on Robota is driven by the aim of designing a robot that fulfills a number of social tasks: education, entertainment, therapy. For each of these tasks, the robot is provided with competences for performing a specific social interaction, which is specified by the rules of the game. Throughout the design of Robota, a number of decision have been made concerning the physical appearance and capabilities of the robot. This section reviews and discusses these decisions.

Conclusion

Vision and speech are key elements of social interactions. Vision provides the means to interpret context-specific behaviors, such as gestures. Speech offers the means to seek and transmit information not readily available visually, such as past and future facts, as well as to provide explicit teaching, such as category differentiations. This paper presented the application of speech processing and machine vision to develop an educational toy robot, Robota, that is being used in pilot studies

Acknowledgements

The mechanical parts of the Robota robots were built by André Guignard, senior technician at the EPFL. The electronic parts of the Robota robots were created by DIDEL SA (http://www.didel.com). Some Robota robots also use the Kameleon board (K-Team SA). The equipment for the class (i.e. the Robota robots and PC’s) was sponsored by an innovative teaching grant from the Intel Corporation to Aude Billard. The research projects conducted with Robota are supported by a Professor Boursier award to

Aude Billard is SNF (Swiss National Science Foundation) Assistant Professor in the School of Engineering Sciences and Techniques at the Swiss Federal Institute of Technology/Lausanne and Adjunct Professor of Computer Science at the University of Southern California. She received her B.Sc. (1994) and M.Sc. (1995) in Physics from EPFL, with specialization in Particle Physics at the European Center for Nuclear Research (CERN). She received her M.Sc. in Knowledge-based Systems (1996) and her Ph.D.

References (17)

  • K Dautenhahn

    Getting to know each other—artificial social intelligence for autonomous robots

    Robotics and Autonomous Systems

    (1995)
  • A Whiten et al.

    On the nature and evolution of imitation in the animal kingdom: reappraisal of a century of research

    Advances in the Study of Behaviour

    (1992)
  • A Billard

    Drama, a connectionist architecture for on-line learning and control of autonomous robots: experiments on learning of a synthetic proto-language with a doll robot

    Industrial Robot

    (1999)
  • A Billard

    Learning motor skills by imitation: a biologically inspired robotic model

    Cybernetics Systems Journal

    (2001)
  • A. Billard, Play, dreams and imitation in Robota, in: K. Dautenhahn, B. Edmonds, L. Canamero (Eds.), Socially...
  • A. Billard, Imitation: a means to enhance learning of a synthetic proto-language in an autonomous robot, in: K....
  • K. Dautenhahn, A. Billard, Games children with autism can play with Robota, a humanoid robotic doll, in: S. Keates,...
  • G Dawson et al.

    Imitation and social responsiveness in autistic children

    Journal of Abnormal Child Psychology

    (1984)
There are more references available in the full text version of this article.

Aude Billard is SNF (Swiss National Science Foundation) Assistant Professor in the School of Engineering Sciences and Techniques at the Swiss Federal Institute of Technology/Lausanne and Adjunct Professor of Computer Science at the University of Southern California. She received her B.Sc. (1994) and M.Sc. (1995) in Physics from EPFL, with specialization in Particle Physics at the European Center for Nuclear Research (CERN). She received her M.Sc. in Knowledge-based Systems (1996) and her Ph.D. in Artificial Intelligence from the Department of Artificial Intelligence at the University of Edinburgh. She has worked as a Post-doctoral Fellow at IDSIA and LAMI (EPFL, 1998–1999), then as Research associate (1999–2000) and as Research Assistant Professor (2000–2002) in the Computer Science Department at the University of Southern California.

View full text