Method, device, vehicle and machine-readable medium for waking up vehicle-mounted voice equipment

Abstract

Embodiments of the present invention provide a wake-up method, device, vehicle, and machine-readable medium for a vehicle-mounted audio device. The method includes: collecting image data of users in the vehicle, and then determining the user corresponding to the user from the image data. According to the user scenario, adjust the wake-up threshold of the vehicle-mounted voice device, and then wake up the vehicle-mounted voice device according to the wake-up threshold, so as to adjust the wake-up threshold of the vehicle-mounted voice device in real time by monitoring the user scene corresponding to the user in the vehicle in real time , adjust the sensitivity of the voice recognition of the vehicle-mounted voice device, and then in different driving scenarios, on the one hand, it can enable the user to wake up the vehicle-mounted voice device more accurately and conveniently, and improve the wake-up interaction efficiency between the user and the voice device; on the other hand By adjusting the wake-up threshold, the probability of false wake-up can be effectively avoided.

Description

technical field [0001] The present invention relates to the technical field of intelligent vehicles, in particular to a method, device, vehicle and machine-readable medium for waking up a vehicle-mounted voice device. Background technique [0002] The current human-computer wake-up interaction method of the vehicle-mounted intelligent voice device is to wake up the voice device after the user enters the voice "wake-up word" and the "voice recognition" is successful. However, only through the interaction mode of voice perception, not only the requirements for wake-up words are high (for example, it is necessary to pre-set the number of words to be 3-5 and avoid colloquial wake-up words, and at the same time require the syllables of the wake-up words to cover as much as possible), and secondly, there are It is difficult to wake up or has a high false wake-up rate. At the same time, when the user wakes up, he needs to input the same "wake-up word" every time, and the wake-up pr...

AI Technical Summary

Problems solved by technology

However, only through the interaction mode of voice perception, not only the requirements for wake-up words are high (for example, it is necessary to pre-set the number of words to be 3-5 and avoid colloquial wake-up words, and at the same time require the syllables of the wake-up words to cover as much as possible), and secondly, there are It is difficult to wake up or has a high rate of false wake-up. At the same time, when the user wakes up, he needs to input the same "wake-up word" every time, and the wake-up process is complicated.

Images