Electric vehicle low-speed alarm voice prompt system and method and vehicle

Through the combination of driving image recognition module and sound collection module, the preset training model is used to dynamically select alarm sounds, which solves the problem that electric vehicles are difficult to perceive when driving at low speeds, and realizes dynamic response to ambient audio and more effective pedestrian warnings.

CN119058542APending Publication Date: 2024-12-03FAW VOLKSWAGEN AUTOMOTIVE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310627627.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-31
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Due to the silent state of electric vehicles when driving at low speed, it is difficult to be perceived by pedestrians, especially blind people, and there are major safety hazards. The existing low-speed alarm system fails to effectively process external sounds, resulting in the preset sound stored in advance that cannot meet the actual ambient audio. change.

Method used

The driving image recognition module is used to collect environmental images, and the sound collection module is used to collect environmental audio. The audio features are extracted through preset training models, and the most relevant target sound is selected from multiple sound libraries for playback. The speaker is installed under the bottom guard plate to balance it. Sound pressure level warns pedestrians.

Benefits of technology

It realizes dynamic adjustment of alarm sounds based on environmental audio, enhances the warning effect for pedestrians, especially when driving at low speeds, which improves safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119058542A_ABST
    Figure CN119058542A_ABST
Patent Text Reader

Abstract

The invention discloses an electric vehicle low-speed alarm voice prompt system and method and a vehicle, and the system comprises a driving image recognition module which is used for collecting an environment image, recognizing obstacles according to the driving image, and outputting a starting instruction when detecting a target obstacle in the recognized obstacles; the sound collection module is electrically connected with the driving image recognition module, and the sound collection module responds to the starting instruction output by the driving image recognition module to collect environment audio; the controller is electrically connected with the sound collection module, receives the environment audio, outputs environment audio characteristics based on the environment audio, comprises a preset training model, and inputs the environment audio characteristics into the preset training model to obtain an environment audio type; the controller selects a target sound library corresponding to the environment audio type from the multiple sound libraries based on the environment audio type, and selects target sound from the target sound library; and the loudspeaker receives and plays the target sound through the controller.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle low-speed alarm voice prompting, and in particular to an electric vehicle low-speed alarm voice prompting system, method and vehicle. Background Art

[0002] In recent years, with the continuous development of the electric vehicle industry, the use of electric vehicles has become more and more common. Because electric vehicles do not have engines inside, they are almost silent when traveling at low speeds. Therefore, pedestrians, especially blind people walking on the roadside, cannot sense the existence of electric vehicles, which poses a major safety hazard.

[0003] The prior art discloses a low-speed warning prompt system for an intelligent automobile, comprising: a low-speed warning controller, a 360° vehicle-mounted camera, a detection radar, a power amplifier module and a speaker; the input end of the low-speed warning controller is respectively connected to the 360° vehicle-mounted camera and the detection radar signal, the first output end of the low-speed warning controller is connected to the input end of the power amplifier module, and the output end of the power amplifier module is connected to the input end of the speaker; the low-speed warning controller obtains the surrounding conditions of the vehicle through the 360° vehicle-mounted camera and the detection radar, and controls the speaker to play a preset voice through the power amplifier module when there are pedestrians or other vehicles within a set distance around the vehicle to give an alarm prompt. However, in the prior art, the surrounding conditions of the vehicle are only identified and transmitted to the low-speed warning controller for processing and decision-making, and it is judged whether the sound is within the alarm distance. If it is within the alarm distance, the preset sound is played. It can be seen that the sound in the patent is stored in advance, and there is no operation such as processing the external sound. Summary of the invention

[0004] To solve at least one aspect of the above problems, the present invention provides an electric vehicle low-speed alarm voice prompt system, comprising: a driving image recognition module, the driving image recognition module is used to collect environmental images and identify obstacles according to the driving images, and the driving image recognition module outputs a start instruction when a target obstacle is detected in the identified obstacles; a sound collection module, the sound collection module is electrically connected to the driving image recognition module, and the sound collection module collects environmental audio in response to the start instruction output by the driving image recognition module; a controller, the controller is electrically connected to the sound collection module, the controller receives the environmental audio, the controller outputs environmental audio features based on the environmental audio, the controller includes a preset training model, the environmental audio features are input into the preset training model to obtain the environmental audio type, the controller includes a plurality of sound libraries, the plurality of sound libraries correspond to a plurality of audio types respectively, the controller selects a corresponding target sound library from the plurality of sound libraries based on the environmental audio type, and selects a target sound from the target sound library; a speaker, the speaker is connected to the controller, the speaker receives the target sound through the controller, and plays the target sound.

[0005] Preferably, the environmental audio characteristics include frequency components, amplitude, loudness, sharpness and roughness.

[0006] Preferably, among the plurality of sound libraries, the plurality of sounds in the target sound library have the smallest coherence with the ambient audio.

[0007] Preferably, the driving image recognition module includes a vehicle speed detection unit and an obstacle detection unit, the vehicle speed detection unit determines the vehicle speed based on the environmental image, and the vehicle speed detection unit outputs an obstacle recognition instruction when the vehicle speed is less than a preset speed, and the obstacle recognition instruction is used to start the obstacle detection unit, and the obstacle detection unit is used to identify obstacles.

[0008] Preferably, the obstacle detection unit outputs the start instruction when detecting the pedestrian.

[0009] On the other hand, a method for low-speed alarm voice prompt of an electric vehicle is provided, which adopts any of the above-mentioned low-speed alarm voice prompt systems of an electric vehicle, including: when the driving image recognition module detects that the vehicle speed is lower than the preset speed, the obstacle detection unit is turned on to judge whether there is a pedestrian in front of the vehicle; when the driving image recognition module detects the pedestrian, the driving image recognition module sends a start instruction to the sound collection module; the sound collection module collects ambient audio in response to receiving the start instruction, and sends the ambient audio to the controller; the controller receives and accepts the ambient audio, and determines the ambient audio characteristics based on the ambient audio, the controller includes a preset training model, the ambient audio characteristics are input into the preset training model to obtain the ambient audio type, the control includes multiple sound libraries, and the multiple sound libraries correspond to multiple audio types one by one, the controller selects its corresponding target sound library from the multiple sound libraries based on the ambient audio type, and selects the target sound from the target sound library; the target sound is played through a speaker.

[0010] Preferably, the environmental audio characteristics include frequency components, amplitude, loudness, sharpness and roughness.

[0011] Preferably, the step of obtaining a preset training model includes: collecting multiple ambient audio samples and preprocessing the multiple ambient audio samples, wherein the preprocessing includes deleting sudden sounds in the ambient audio samples and unifying the ambient audio format; performing feature extraction on the multiple ambient audio samples after preprocessing to obtain ambient audio features of the multiple ambient audio samples; performing neural network model training on the ambient audio features of the multiple ambient audio samples and their corresponding audio types to determine the correspondence between ambient audio samples and sound types to obtain a preset training model.

[0012] On the other hand, a vehicle is provided, comprising an electric vehicle low-speed warning voice prompt system as described above, wherein the speaker is installed in a middle position below the bottom guard plate.

[0013] The electric vehicle low-speed warning voice prompt system and method of the embodiment of the present invention have the following beneficial effects: after the vehicle collects ambient sound during driving, the collected ambient sound will be pre-processed. After pre-processing, the five parameters of the ambient sound, namely frequency component, amplitude, loudness, sharpness and roughness, are extracted, and the five parameters are input into a pre-trained model for matching, so as to output sound to warn pedestrians; the low-speed warning sound device is installed in the middle position under the bottom guard plate, so that the sound pressure levels on the left and right can be more balanced, the sound emitted by the speaker outside the vehicle can be the largest, and the sound heard inside the vehicle can be the smallest, which can better warn pedestrians, especially the blind. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to better understand the above and other purposes, features, advantages and functions of the present invention, reference may be made to the embodiments shown in the accompanying drawings. The same reference numerals in the accompanying drawings refer to the same components. It should be understood by those skilled in the art that the accompanying drawings are intended to schematically illustrate the preferred embodiments of the present invention and have no limiting effect on the scope of the present invention, and the components in the drawings are not drawn to scale.

[0015] Figure 1 A schematic diagram of the range of environmental images collected by the driving image recognition module of the electric vehicle low-speed warning voice prompt system according to an embodiment of the present invention;

[0016] Figure 2 The present invention is a schematic diagram of an application scenario of a sound collection module of a low-speed warning voice prompt system for an electric vehicle according to an embodiment of the present invention.

[0017] Reference numerals:

[0018] 1. Vehicle; 2. Driving image recognition module; 3. Sound collection module. DETAILED DESCRIPTION

[0019] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0020] As used herein, the term "including" and its variations mean open inclusion, i.e., "including but not limited to". Unless otherwise stated, the term "or" means "and / or". The term "based on" means "based at least in part on". The terms "an example embodiment" and "an embodiment" mean "at least one example embodiment". The term "another embodiment" means "at least one additional embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0021] In order to at least partially solve one or more of the above problems and other potential problems, an embodiment of the present disclosure proposes a low-speed warning voice prompt system for an electric vehicle, comprising: a driving image recognition module, the driving image recognition module is used to collect environmental images and identify obstacles based on the driving images, and the driving image recognition module outputs a start instruction when a target obstacle is detected in the identified obstacles; a sound collection module, the sound collection module is electrically connected to the driving image recognition module, and the sound collection module collects environmental audio in response to the start instruction output by the driving image recognition module; a controller, the controller is electrically connected to the sound collection module, the controller receives the environmental audio, the controller outputs environmental audio features based on the environmental audio, the controller includes a preset training model, the environmental audio features are input into the preset training model to obtain the environmental audio type, the controller includes multiple sound libraries, the multiple sound libraries correspond to multiple audio types one by one, the controller selects its corresponding target sound library from the multiple sound libraries based on the environmental audio type, and selects the target sound from the target sound library; a speaker, the speaker is connected to the controller, the speaker receives the target sound through the controller, and plays the target sound.

[0022] Specifically, the driving image recognition module includes a driving image recorder, which is installed at the rearview mirror inside the vehicle and shared with the driving recorder to collect environmental images in the direction of the vehicle's advance in real time, and identify obstacles based on the environmental images, so as to determine whether there are pedestrians among the obstacles. In some embodiments, the obstacle detection unit outputs a start command when a pedestrian is detected. The selection parameters of the driving image recorder are: a wide-angle range greater than 140 degrees, a resolution of more than 2k, and an aperture requirement of less than F1.6, which can effectively improve night monitoring and prediction capabilities, such as Figure 1 As shown, the shooting range of the driving image recorder is a 140-degree sector within a preset distance in front of the vehicle. In another real-time example, the obstacle detection unit outputs a start command when detecting an animal or a pedestrian.

[0023] The sound collection module is used to collect ambient audio around the vehicle. The sound collection module includes a first microphone and a second microphone. Figure 2 As shown, in order to collect the ambient sound around the vehicle, the first microphone is installed on the left side of the vehicle, and the second microphone is installed on the right side of the vehicle, and the positions of the first microphone and the second microphone are symmetrical. The first microphone and the second microphone of the sound collection module start collecting ambient audio in response to the start command issued by the driving image recognition module.

[0024] The controller receives the ambient audio by connecting with the sound collection module, and the controller preprocesses the acquired ambient audio, and the preprocessing method includes data validity and data format, etc. The ambient audio features are extracted from the preprocessed ambient audio, and the extracted ambient audio features include five parameters: frequency component, amplitude, loudness, sharpness and roughness. The controller includes a preset training model, and the preset training model is used to determine the ambient audio type according to the ambient audio features, wherein different ambient audio types characterize the sounds of different environments. The controller includes multiple sound libraries, each of which corresponds to an ambient audio type, and each of which includes multiple sounds. The controller selects a sound library corresponding to the ambient audio type from the multiple sound libraries according to the ambient audio type output by the preset training model, and selects the sound library as the target sound library. The controller arbitrarily selects a sound from the target sound library as the target sound.

[0025] The speaker receives the target sound by connecting to the controller, and the speaker is used to play the target sound to prompt pedestrians.

[0026] In some embodiments, the plurality of sounds in the target sound library have the least coherence with the ambient audio among the plurality of sound libraries.

[0027] Specifically, the coherence of the multiple sounds in the target sound library with the ambient audio collected by the sound collection module is the smallest among all the sound libraries of the controller. The greater the difference in frequency components, amplitudes, loudness, sharpness, and roughness between the sounds in the target sound library and the ambient audio, the easier it is to hear the sounds in the target sound library when played in the sound environment corresponding to the ambient audio.

[0028] In some embodiments, the driving image recognition module includes a vehicle speed detection unit and an obstacle detection unit. The vehicle speed detection unit determines the vehicle speed based on the environmental image. The vehicle speed detection unit outputs an obstacle recognition instruction when the vehicle speed is less than a preset speed. The obstacle recognition instruction is used to start the obstacle detection unit, and the obstacle detection unit is used to identify obstacles.

[0029] Specifically, the vehicle speed detection unit of the controller determines the current vehicle speed through the environmental image, and starts the obstacle detection unit when the current vehicle speed is less than the preset vehicle speed (for example, 30 km / h). The obstacle detection unit is used to identify obstacles and determine whether there are pedestrians among the obstacles. Alternatively, in another embodiment, the driving image recognition module receives the current vehicle speed by connecting to the vehicle controller, and starts the obstacle detection unit when the received current vehicle speed is less than the preset vehicle speed.

[0030] On the other hand, a method for low-speed alarm voice prompt of an electric vehicle is provided, which adopts any of the previous low-speed alarm voice prompt systems of an electric vehicle, including: when the driving image recognition module detects that the vehicle speed is lower than the preset speed, the obstacle detection unit is turned on to judge whether there is a pedestrian in front of the vehicle; when the driving image recognition module detects the pedestrian, the driving image recognition module sends a start instruction to the sound collection module; the sound collection module collects ambient audio in response to receiving the start instruction, and sends the ambient audio to the controller; the controller receives the received ambient audio and determines the ambient audio characteristics based on the ambient audio, the controller includes a preset training model, the ambient audio characteristics are input into the preset training model to obtain the ambient audio type, the controller includes multiple sound libraries, the multiple sound libraries correspond to multiple audio types one by one, the controller selects its corresponding target sound library from the multiple sound libraries based on the ambient audio type, and selects the target sound from the target sound library; the target sound is played through a speaker.

[0031] Specifically, when the driving image recognition module detects that the current vehicle speed is lower than 30 km / h, the obstacle detection unit is turned on to identify obstacles based on the collected environmental images, and to determine whether there are pedestrians in front of the vehicle based on the identified obstacles.

[0032] When the driving image recognition module detects a pedestrian, the driving image recognition module sends a start command to the sound collection module. After the sound collection module receives the command, the first microphone and the second microphone of the sound collection module will collect the current ambient sound, which refers to the sound around the current vehicle, and send the collected ambient sound to the controller.

[0033] After receiving the ambient audio, the controller will pre-process the ambient audio, the purpose of which is to clean the collected ambient audio. The pre-processing methods include data validity (for example, judging that the sound whose volume exceeds the average volume of the ambient audio is invalid audio), data format (for example, unifying different audio formats), etc. The ambient audio features of the ambient audio after pre-processing are obtained, including five parameters: frequency component (including sounds of different frequencies), amplitude, loudness, sharpness, and roughness. The five parameters of frequency component, amplitude, loudness, sharpness, and roughness of the ambient audio features are input into a preset training model that has been trained in advance to output the audio type of the sound information. Among them, different audio types represent sounds of different environments, and each audio type corresponds to an E-Sound sound library; the E-Sound sound library corresponding to each audio type is the E-Sound sound library with the least coherence with the sound of the audio type.

[0034] According to the audio type output by the preset training model, a target E-Sound sound library corresponding to the audio type is determined, and a sound is randomly selected from the target E-Sound sound library and input into the speaker, which is played by the speaker.

[0035] In some embodiments, the step of obtaining a preset training model includes: collecting multiple ambient audio samples and preprocessing the multiple ambient audio samples, the preprocessing including deleting mutant sounds in the ambient audio samples and unifying the ambient audio format; performing feature extraction on the multiple ambient audio samples after preprocessing to obtain ambient audio features of the multiple ambient audio samples; training a neural network model with the ambient audio features of the multiple ambient audio samples and their corresponding audio types to determine the correspondence between the ambient audio samples and the sound types to obtain a preset training model.

[0036] Specifically, firstly, a plurality of environmental audio samples are collected, for example, audio of a road surface such as an intersection or a crosswalk at any time is collected as an environmental audio sample;

[0037] The sound in multiple environmental audio samples is preprocessed to clean the collected samples. The preprocessing methods include data validity, data format, etc.

[0038] Feature extraction is performed on the multiple environmental audio samples after preprocessing to obtain frequency components, amplitudes, loudness, sharpness, roughness, etc. of the multiple environmental audio samples. The feature extraction method includes empirical mode decomposition or Fourier transform.

[0039] The frequency components, amplitude, loudness, sharpness and roughness of multiple environmental audio samples and their corresponding sound types are used to train a neural network model, the correspondence between multiple environmental audio samples and sound types is determined, and a GBDT model (i.e., a preset training model) is obtained.

[0040] The electric vehicle low speed warning voice prompt method also includes, after obtaining the preset training model, calculating the coherence between the sound corresponding to each audio type and the multiple sounds included in each E-Sound sound library, and selecting the E-Sound sound library with the smallest coherence value with each sound type as the E-Sound sound library corresponding to each audio type, thereby determining the E-Sound sound library corresponding to each audio type. The smaller the coherence between the sound corresponding to the sound type and the sound included in each E-Sound sound library, the greater the difference in frequency components, amplitude, loudness, sharpness and roughness between the two, and the easier it is to hear the sound in the E-Sound sound library when played in the sound environment corresponding to the audio type.

[0041] On the other hand, a vehicle is provided, the vehicle includes a low-speed warning voice prompt system for an electric vehicle as described above, and a speaker is installed in the middle position below the bottom guard plate.

[0042] Specifically, the electric vehicle low speed warning voice prompt system includes a driving image recognition module, a sound collection module, a controller and a speaker. Figure 2 As shown, the sound collection module includes a first microphone and a second microphone. In order to collect the ambient sound around the vehicle, the first microphone is installed on the left side of the vehicle, and the second microphone is installed on the right side of the vehicle, and the positions of the first microphone and the second microphone are symmetrical. The driving image recorder of the driving image recognition module is installed at the rearview mirror inside the vehicle and shared with the driving recorder to collect in real time whether there are pedestrians in the direction of the vehicle's advance. The selection parameters of the driving image recorder are: wide-angle range greater than 140 degrees, resolution above 2k, and aperture requirement below F1.6, which can effectively improve night monitoring and prediction capabilities. The speaker is installed in the middle position under the bottom guard plate, which can make the left and right sound pressure levels more balanced, so that the sound emitted by the speaker outside the car is the largest, and the sound heard inside the car is the smallest, which can better warn pedestrians, especially the blind.

[0043] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other persons of ordinary skill in the art to understand the present invention.

Claims

1. An electric vehicle low speed alarm voice prompt system, characterized in that: include: A driving image recognition module, the driving image recognition module is used to collect environmental images and recognize obstacles based on the driving images, and the driving image recognition module outputs a start instruction when a target obstacle is detected among the recognized obstacles; A sound collection module, the sound collection module is electrically connected to the driving image recognition module, and the sound collection module collects ambient audio in response to the start instruction output by the driving image recognition module; A controller, wherein the controller is electrically connected to the sound collection module, the controller receives the ambient audio, the controller outputs ambient audio features based on the ambient audio, the controller includes a preset training model, the ambient audio features are input into the preset training model to obtain the ambient audio type, the controller includes a plurality of sound libraries, the plurality of sound libraries correspond to a plurality of audio types respectively, the controller selects a corresponding target sound library from the plurality of sound libraries based on the ambient audio type, and selects a target sound from the target sound library; A speaker is connected to the controller, and the speaker receives the target sound through the controller and plays the target sound.

2. The system according to claim 1, characterized in that The environmental audio characteristics include frequency components, amplitude, loudness, sharpness, and roughness.

3. The system according to claim 2, characterized in that Among the multiple sound libraries, the multiple sounds in the target sound library have the smallest coherence with the environmental audio.

4. The system according to claim 1, characterized in that The driving image recognition module includes a vehicle speed detection unit and an obstacle detection unit. The vehicle speed detection unit determines the vehicle speed based on the environmental image. The vehicle speed detection unit outputs an obstacle recognition instruction when the vehicle speed is less than a preset speed. The obstacle recognition instruction is used to start the obstacle detection unit, and the obstacle detection unit is used to identify obstacles.

5. The system according to claim 4, characterized in that The obstacle detection unit outputs the start instruction when detecting the pedestrian.

6. A method for low-speed alarm voice prompt of an electric vehicle, using the low-speed alarm voice prompt system of an electric vehicle as claimed in any one of claims 1 to 5, characterized in that: include: When the driving image recognition module detects that the vehicle speed is lower than the preset speed, the obstacle detection unit is turned on to determine whether there are pedestrians in front of the vehicle; When the driving image recognition module detects a pedestrian, the driving image recognition module sends a start instruction to the sound collection module; The sound collection module collects ambient audio in response to receiving the start instruction, and sends the ambient audio to the controller; The controller receives and accepts the ambient audio, and determines the ambient audio features based on the ambient audio, the controller includes a preset training model, the ambient audio features are input into the preset training model to obtain the ambient audio type, the controller includes a plurality of sound libraries, the plurality of sound libraries correspond to a plurality of audio types respectively, the controller selects a corresponding target sound library from the plurality of sound libraries based on the ambient audio type, and selects a target sound from the target sound library; The target sound is played through a speaker.

7. The method according to claim 6, characterized in that The environmental audio characteristics include frequency component amplitude, loudness, sharpness and roughness.

8. The method according to claim 7, characterized in that The steps to obtain a preset training model include: Collecting multiple environmental audio samples, and preprocessing the multiple environmental audio samples, wherein the preprocessing includes deleting sudden sounds in the environmental audio samples and unifying the environmental audio format; Performing feature extraction on the plurality of environmental audio samples after preprocessing to obtain environmental audio features of the plurality of environmental audio samples; The environmental audio features of the plurality of environmental audio samples and their corresponding audio types are subjected to neural network model training to determine the correspondence between the environmental audio samples and the sound types, thereby obtaining a preset training model.

9. A vehicle, characterized in that: The vehicle includes an electric vehicle low-speed warning voice prompt system as described in any one of claims 1-5, and the speaker is installed in the middle position below the bottom guard plate.