2B-1: Learning to Understand Spoken Commands through a Human-Robot Training Task -
| 発表番号 | 2B-1 |
| 発表タイトル | Learning to Understand Spoken Commands through a Human-Robot Training Task |
| 筆者氏名・所属 |
Anja Austermann (The Graduate University for Advanced Studies (SOKENDAI)) Seiji Yamada (The Graduate University for Advanced Studies (SOKENDAI)， National Institute of Informatics) |
| アブストラクト | We propose a method by which a robot can learn parameterized，naturally spoken commands，like “Please switch the TV on!”，“It’s too dark here” or “Can you bring me a coffee?” through natural interaction with a user in a training task. The goal of the training phase is to allow the user to give commands to a robot in his preferred way instead of learning predefined commands from a handbook. In the training phase，a simplified living room scene is shown on a screen. The robot can request the task server to show a situation，which requires an action from the robot. E.g. the light is switched off and the room is dark. The user instructs the robot by giving appropriate commands like “Please switch the light on”. The robot uses these utterances to learn to understand the user’s instructions. Learning is done in two successive steps. First the robot learns object names. Then it uses the known object names to learn parameterized command patterns and determine the position of parameters in a spoken command. The algorithm uses a combination of Hidden Markov Models and Classical Conditioning to handle alternative ways to utter the same command and integrate information from different modalities. |
| 論文 | PDFファイル |
