US 11,938,958 B2
Voice dialogue device, voice dialogue system, and control method for voice dialogue system
Takehito Teraguchi, Kanagawa (JP); Hirofumi Inoue, Kanagawa (JP); Jo Nishiyama, Kanagawa (JP); Shota Okubo, Kanagawa (JP); and Yu Shikoda, Kanagawa (JP)
Assigned to Nissan Motor Co., Ltd., Kanagawa (JP)
Appl. No. 17/266,401
Filed by Nissan Motor Co., Ltd., Kanagawa (JP)
PCT Filed Aug. 6, 2018, PCT No. PCT/JP2018/029470
§ 371(c)(1), (2) Date Feb. 5, 2021,
PCT Pub. No. WO2020/031241, PCT Pub. Date Feb. 13, 2020.
Prior Publication US 2021/0309241 A1, Oct. 7, 2021
Int. Cl. B60W 50/10 (2012.01); B60R 16/037 (2006.01); B60W 50/14 (2020.01); G10L 15/22 (2006.01)
CPC B60W 50/10 (2013.01) [B60R 16/0373 (2013.01); B60W 50/14 (2013.01); G10L 15/22 (2013.01); B60W 2420/54 (2013.01); B60W 2520/10 (2013.01); B60W 2540/21 (2020.02); B60W 2540/22 (2013.01); B60W 2552/05 (2020.02); B60W 2552/30 (2020.02); B60W 2554/406 (2020.02); B60W 2556/45 (2020.02); G10L 2015/227 (2013.01); G10L 2015/228 (2013.01)] 12 Claims
OG exemplary drawing
 
1. A voice dialogue device comprising:
a controller configured to execute a dialogue with an occupant in a vehicle using at least one sensor equipped in the vehicle, wherein
the at least one sensor detects at least one of a traveling state of the vehicle, an external environmental state of the vehicle, and a state of the occupant, and
the controller is configured to:
determine, on the basis of the detection result, a state of a load on the occupant; and
execute a dialogue with the occupant by executing a first dialogue program when the size of the load is larger than the predetermined standard, and executing a second dialogue program different from the first dialogue program when the size of the load is smaller than the predetermined standard,
wherein the first dialogue program is programmed to ask questions for which the expected occupant's answers include fewer words than with the second dialogue program,
wherein the first dialogue program uses discrete word recognition to acquire an answer to a question from the occupant, discrete word recognition recognizing one or more predetermined words in voice information of the occupant,
wherein the second dialogue program uses sentence recognition to acquire an answer to a question from the occupant, sentence recognition recognizing voice information of the occupant in a unit of a sentence.