Back to results
US 9,870,770 B2Granted
Voice recognition device and method in vehicle
AssigneeHyundai Motor Company
InventorKyuSeop Bang
Life of this patent
Estimating the term…
Abstract
A voice recognition system in a vehicle includes: a first microphone mounted in the vehicle that collects voice data of an occupant of the vehicle; a second microphone provided in a mobile device of the occupant that collects voice data of the occupant; and a voice recognition device connected to the mobile device through local wireless communication including a noise elimination portion eliminating noise in the voice data collected by the first microphone or the second microphone and a voice recognition portion performing voice recognition using the voice data from which noise is eliminated by the noise elimination portion.
Claims9
A voice recognition system in a vehicle, comprising:
- a first microphone mounted in the vehicle that collects voice data of an occupant of the vehicle;
- a second microphone provided in a mobile device of the occupant that collects voice data of the occupant;
- and
- a voice recognition device connected to the mobile device through local wireless communication including a noise elimination portion eliminating noise in the voice data collected by the first microphone or the second microphone and a voice recognition portion performing voice recognition using the voice data from which noise is eliminated by the noise elimination portion,
- wherein the voice recognition portion further includes:
- a feature extraction portion extracting a reference parameter of the voice data from which noise is eliminated by the noise elimination portion;
- a parameter setting portion adjusting the reference parameter based on a recognition rate of the reference parameter;
- a storage portion storing voice recognition data;
- and
- a meaning extraction portion extracting meaning of the voice data from which noise is eliminated by the noise elimination portion by comparing the reference parameter to the voice recognition data,
- wherein the parameter setting portion calculates the recognition rate when a summation of a number of recognition and a number of misrecognition is greater than a predetermined value,
- wherein the parameter setting portion increases the reference parameter until the reference parameter reaches a predetermined maximum value when the recognition rate is less than a first threshold value, and then the meaning extraction portion performs voice recognition when the reference parameter reaches the predetermined maximum value, and
- wherein the parameter setting portion decreases the reference parameter until the reference parameter reaches a predetermined minimum value when the recognition rate is greater than a second threshold value, and then the meaning extraction portion performs voice recognition when the reference parameter reaches the predetermined minimum value.
10 more elements
A voice recognition method in a vehicle, comprising:
- collecting voice data of an occupant of the vehicle using a first microphone mounted in the vehicle and a second microphone provided in a mobile device of the occupant;
- eliminating, by a voice recognition device, noise in the voice data collected by the first microphone or the second microphone;
- performing, by the voice recognition device, voice recognition of the voice data from which noise is eliminated;
- extracting, by the voice recognition device, a reference parameter of the voice data from which noise is eliminated by the noise elimination portion;
- adjusting, by the voice recognition device, the reference parameter based on a recognition rate of the reference parameter;
- storing, by the voice recognition device, voice recognition data;
- and
- extracting, by the voice recognition device, meaning of the voice data from which noise is eliminated by the noise elimination portion by comparing the reference parameter to the voice recognition data,
- wherein the voice recognition device is connected to the mobile device through local wireless communication,
- wherein the parameter setting portion calculates the recognition rate when a summation of a number of recognition and a number of misrecognition is greater than a predetermined value,
- wherein the parameter setting portion increases the reference parameter until the reference parameter reaches a predetermined maximum value when the recognition rate is less than a first threshold value, and then the meaning extraction portion performs voice recognition when the reference parameter reaches the predetermined maximum value, and
- wherein the parameter setting portion decreases the reference parameter until the reference parameter reaches a predetermined minimum value when the recognition rate is greater than a second threshold value, and then the meaning extraction portion performs voice recognition when the reference parameter reaches the predetermined minimum value.
9 more elements
A non-transitory computer readable medium containing program instructions for performing a voice recognition method in a vehicle, in which voice data of an occupant of the vehicle is collected using a first microphone mounted in the vehicle and a second microphone provided in a mobile device of the occupant, the computer readable medium comprising:
- program instructions that eliminate noise in the voice data collected by the first microphone or the second microphone;
- program instructions that perform voice recognition of the voice data from which noise is eliminated;
- program instructions that extract a reference parameter of the voice data from which noise is eliminated by the noise elimination portion;
- program instructions that adjust the reference parameter based on a recognition rate of the reference parameter;
- program instructions that store voice recognition data;
- and
- program instructions that extract meaning of the voice data from which noise is eliminated by the noise elimination portion by comparing the reference parameter to the voice recognition data,
- wherein the voice recognition device is connected to the mobile device through local wireless communication,
- wherein the parameter setting portion calculates the recognition rate when a summation of a number of recognition and a number of misrecognition is greater than a predetermined value,
- wherein the parameter setting portion increases the reference parameter until the reference parameter reaches a predetermined maximum value when the recognition rate is less than a first threshold value, and then the meaning extraction portion performs voice recognition when the reference parameter reaches the predetermined maximum value, and
- wherein the parameter setting portion decreases the reference parameter until the reference parameter reaches a predetermined minimum value when the recognition rate is greater than a second threshold value, and then the meaning extraction portion performs voice recognition when the reference parameter reaches the predetermined minimum value.
8 more elements
Dependencies are inferred from the claim text.