公开(公告)号：US11158316B2

公开(公告)日：2021-10-26

申请号：US16558642

申请日：2019-09-03

Applicant: ALPINE ELECTRONICS, INC.

Inventor： Tomohiro Arikita ,  Masaaki Ohira

IPC: G10L15/22 ,  G10L15/08

Abstract: An in-vehicle device includes: a first speech recognition unit configured to recognize a word indicating the in-vehicle device and a word indicating a first instruction type by speech recognition, from first voice data input to the in-vehicle device, the first instruction type being one of instruction types for operating the in-vehicle device; and a second speech recognition unit configured, in response to the word indicating the in-vehicle device being recognized from the first voice data by the first speech recognition unit, to recognize a word indicating a second instruction type by speech recognition, from second voice data input to the in-vehicle device, the second voice data being input after the first voice data, and the second instruction type being one of the instruction types for operating the in-vehicle device.

公开(公告)号：US10915565B2

公开(公告)日：2021-02-09

申请号：US16263272

申请日：2019-01-31

Applicant: ALPINE ELECTRONICS, INC.

Inventor： Masaaki Ohira ,  Yuki Nakahara

IPC: G06F40/30 ,  G10L15/22 ,  G06F16/38 ,  G06F16/33 ,  G06F3/16 ,  G06F16/332 ,  G10L25/54 ,  G06F40/205 ,  G06F40/289 ,  G10L15/26

Abstract: A retrieval result providing device includes a retrieval unit which acquires an individual retrieval result having item values of a plurality of items by a retrieval based on a keyword, an unprocessed sentence generation unit which generates an unprocessed sentence where retrieval result corresponding sentences with the item values arranged are arranged, an output sentence generation unit which performs at least one of a first process of erasing some of the sentences on the basis of a commonality between a word and the keyword contained in the retrieval result corresponding sentences, a second process of erasing some of the sentences on the basis of a commonality of the item values of specific items in the different sentences, and a third process of erasing some of the sentences on the basis of a commonality of the item values of a plurality of different items in one sentence so as to generate an output sentence, and an output sentence output unit which outputs a voice based on the output sentence.