CONTROLLER, SPOKEN DIALOG DEVICE, SPEECH RECOGNITION SERVER, AND PROGRAM

To provide a controller for a spoken dialog device with which it is possible to improve speech recognition accuracy while suppressing a wasteful increase in communication cost and power consumption.SOLUTION: The controller includes an arithmetic unit for controlling a spoken dialog device having an utterance section detector that identifies an acquired speech by an identification level which is set to see whether or not the utterance of a subject is included and detects an utterance section. When an estimation result that the possibility of an utterance by a subject being included in the acquired speech is high is acquired by a speech recognition server, the arithmetic unit indicates to the spoken dialog device that the identification level of the utterance section detector be changed and communication with the speech recognition server be performed by utterance section detection that corresponds to an identification level after the change.SELECTED DRAWING: Figure 1 【課題】通信コスト及び電力消費が無駄に増大するのを抑えつつ、音声認識精度を向上させることができる、音声対話装置の制御装置を提供する。【解決手段】制御装置は、取得された音声について、対象者の発話を含むか否かを設定された識別レベルで識別して発話区間検出を行う発話区間検出器を有する音声対話装置を制御する演算部を備え、演算部は、音声認識サーバより、取得された音声に対象者による発話が含まれている可能性が高いとの推定結果を取得した場合に、音声対話装置に対し、発話区間検出器の識別レベルを下げるように設定変更して変更後の識別レベルに応じた発話区間検出で音声認識サーバとの通信を行うよう指示する。【選択図】図１