RU200043U1 - Wearable speech aid - Google Patents

Abstract

The utility model relates to the field of wearable electronics, in particular, medical technology and is designed to help people with speech and other disorders of speech and hearing (dyslalia, dysarthria, aphonia, dumbness, deafness) in speech, as well as for all those who want to expand opportunities own speech and hearing due to synthesized voice and recognized speech, for example, when practicing rhetoric, training public speaking. The useful model consists of a housing with a speaker, a microphone, a Bluetooth radio module and a control unit, as well as a set of sensors located in half arcs and attached to the body devices that simultaneously act as fixators of the utility model on a person's neck. The utility model allows a stuttering person to reproduce a speech utterance using artificially synthesized speech created in a cloud service and transmitted to a utility model in the form of an audio file using an input / output device, for example, using a smartphone with Bluetooth technology, pa working in 3G / 4G / 5G mobile networks, for further playback in front of the interlocutor. 4 ill.

Description

The field of technology.

The utility model relates to the field of wearable electronics, in particular, medical technology and is designed to help people with speech and other disorders of speech and hearing (dyslalia, dysarthria, aphonia, dumbness, deafness) in speech, as well as for all those who want to expand opportunities own speech and hearing due to synthesized voice and recognized speech, for example, when practicing rhetoric, training in public speaking.

Prior art.

Known device, described in patent RU 2230532, related to medical technology and designed to eliminate stuttering, speech and hearing correction caused by wounds and traumas of the brain. The device contains an audio signal receiver, a signal amplifier, two delay lines, two switching units, a low-pass and high-pass filter, a filter block and frequency settings, which has a direct channel, two band-pass filters, low-pass and high-pass filters, and the apparatus also contains a summing signal amplifier, headphones and tactile vibrator. The device's action is based on the delayed speech effect. The time of sound perception is extended: stuttering causes a change in the acoustic perception of speech. When the return acoustic signal is delayed by a certain amount, a positive effect is detected, manifested in the complete or partial normalization of speech activity.

A device for eliminating stuttering in simplex speech is built on a similar principle of separating pronunciation and speech perception (patent RU 126592). A feature is the complete noise suppression of the speech of a stuttering person for their own perception (transition to simplex speech mode). The device contains an acoustic noise generator, an audio amplifier, headphones, a controlled closing switch with a delayed opening and a laryngophone, the output of which is connected through an audio amplifier to the control input of the switch, the input of which is connected to the output of the acoustic noise generator, and the output of the controlled switch is connected to the inputs of the head phones, which ensures a successful result of prompt (immediate) elimination of stuttering in simplex speech.

The disadvantages of the above devices include the normalization of speech on an outpatient basis: during classes with a speech therapist, independent speech training. In the process of real communication during study or work, the stutter hears himself in full duplex mode and without delays in perception. As a result, hesitation in speech returns.

Known patent RU 2312646 "Device for partial replacement of speech and hearing functions" related to medical technology and aimed at the rehabilitation of disabled people and people with disabilities, such as speech and hearing impairment, to carry out communication actions, verbally communicate with others and receive the necessary messages and commands via telephone and radio communications. The device protected by this patent contains a control unit (signal processor unit) connected by bidirectional multi-bit buses with a plurality of units: keyboards, signal coding, generation of signals of an optical indicator, optical indicator, speech synthesizer, sound indication, wire interface with a telephone line, contact interface with external digital devices, memory units for indication of optical signals, telephone addresses, codes of prepared words and references, codes for transmitting digital information, memory of device algorithms, electromechanical indicator of Braille alphabet, radio receiver, decoding of radio receiver signals and other auxiliary units.

The disadvantages of this invention include a wide scope of technical problems solved by the invention and, at the same time, the presence of functional limitations to practical use, significantly reducing the overall efficiency of using the device by disabled people and persons with disabilities, including stuttering people.

A known system for reducing stuttering by detecting vibrations of the vocal cords (US 2015/0305920). The system consists of two components: a unit located on the stutter's neck and headphones with a wireless interface. The unit is equipped with a piezoelectric sensor that detects the vibration of the vocal cords, an operational amplifier, an analog-to-digital converter, a processor and a wireless transmitter. When a spasm of the vocal cords occurs, which responds to the stuttering stutter, the unit sends a signal to the earpiece, which immediately turns on the music and disrupts the connection between the subconscious and the spasmodic vocal cords. In other words, it switches attention from stammering, and also does not provide an opportunity to stutter constantly, which gives a person the opportunity to relax and continue communication.

For accurate operation of the system and timely switching on of music, a universal calibration mechanism is required for sensors that determine the vibration of the vocal cords, regardless of the anthropometric parameters of a person. The quality of such a mechanism directly determines the possibility of using the device in practice and is a vulnerability in the system under consideration.

The two above-described devices were selected as the closest prototypes of the utility model: "Device for partial replacement of speech and hearing functions", "System for reducing stuttering by detecting vibrations of the vocal cords."

Synthesis and recognition systems for artificial speech.

Technologies for synthesizing speech from text (TTS, Text-To-Speech) have been developing for several decades and are currently widely represented, both in the form of open source software with the possibility of unlimited use (for example, Gnuspeech), and in the form of proprietary technologies based on the cloud services of the largest technology companies: Google Cloud Speech, Microsoft Cognitive Services, Amazon Polly and others. In order to synthesize speech, it is enough to use the programming interface of these services (API), available via the Internet: transmit the text in a web request and receive an audio recording in response.

Disclosure of the essence of the utility model.

There is an approach in which stuttering is considered as a violation of the stage of readiness for speech (cited by E.N. Sadovnikova "Stuttering from the perspective of a logopsycho-corrective approach." M .: LOGOMAG, 2019, pp. 14-15):

"N.Yu. Abeleva identifies four stages of generating a speech utterance:

1) communicative intention;

2) inner speech;

3) readiness to speak;

4) sound implementation of speech.

It is at the third stage, the stage of readiness for speech, that the stutter anticipates the onset of usual hesitation for him, which is expressed in a state of fear of varying severity. N.Yu. Abeleva relies on the theoretical provisions of P.K. Anokhin and N.A. Bershtein, who proved that the motor response is anticipated and constantly corrected by the “model of the future result”. In other words, the state of expectation of stuttering in the forthcoming speech causes fear and affects the sound realization of speech with stuttering. Fear prevents a person from entering into a conversation, and often the stutter prefers not to do this, leaving his thought unspoken, communication not begun, the problem not resolved.

Helping people with stuttering start a conversation (according to N.Yu. Abeleva, go to the fourth stage - sound realization of speech) is an important task on the way to the goal of free communication between the stutter and a person without speech defects. The proposed useful model of a wearable speech aid device is considered as one of the ways to solve this problem.

The utility model is based on replacing the transition from internal speech (the second stage of generating an utterance) to the fourth stage using a method that functions as well for most stuttering people as for people without speech defects - written speech. Having written an inner speech in any convenient way (in a smartphone, on a keyboard), the actions of the stutter by the generation of a speech utterance end, and the rest (to utter a speech utterance using high-quality synthesized speech) is done for him by a useful model worn around his neck. Thus, the third and fourth stages of utterance generation move from the inevitable components that generate fear to the optional ones. The person himself chooses how in the current situation it is more comfortable for him to form a speech utterance - independently or using the proposed device. Similarly, one can act in the case of dyslalia, dysarthria, aphonia, dumbness, deafness and other disorders, when the use of synthesized speech may be more comfortable for person A than the capabilities of his own speech apparatus.

Thus, the technical result is achieved, which consists in providing assistance in speech for stuttering people, as well as for people with other speech and hearing impairments.

The utility model in its version can be equipped with sensors for measuring the physiological parameters of a person, which register the state before speech, during speech and after speech, a microphone for recording one's own speech and the speech of others, which can be converted into text using cloud speech recognition and output on the screen of a smartphone or in memory, as well as a video camera, which together makes it possible to collect comprehensive data about the stutter and his environment during the onset of hesitation for further study of this speech disorder.

As a result, an additional technical result is achieved, which consists in collecting data about the stuttering person and his environment.

Without limiting generality, the technical results achieved by a wearable speech aid device using speech generation and recognition functionality can be extended not only to stuttering people, but also to anyone who wants to use synthesized voice or recognized speech in everyday life instead of their own speech capabilities. and hearing or in addition to them. For example, in the case of rhetoric and public speaking training, you can synthesize reading at the desired speed and then focus on it as a model. Since the device is portable and designed to be worn around the neck, it is convenient to always keep it nearby during the day.

Thus, an additional technical result is achieved, which consists in expanding the capabilities of a person's speech and hearing in everyday life.

Brief description of the drawings.

Figure 1 shows a schematic representation of variants of a wearable device for speech assistance (hereinafter referred to as wearable device 100) in a basic arrangement without a sensor unit 140-145 (version 1), as well as in an arrangement with a sensor unit 140-145 (option 2).

Figure 2 shows a schematic representation of variants of a wearable device 100 in an arrangement with a camera unit 130-132 and without a sensor unit 140-146 (option 3), as well as in an arrangement with a camera unit 130-132 and a sensor unit 140-146 (option execution 4).

Figure 3 is a block diagram of a wearable device 100 for embodiment 4 with the largest number of components. Component diagrams for options 1-3 are obtained by excluding blocks from the diagram that are not involved in the work of the corresponding option.

Figure 4 shows a diagram of a wearable device 100 in a situation of dialogue between two people: A - the user of the device (stuttering person), B - his interlocutor.

Implementation of the invention.

For clarification, we will use Figures 1 and 3 for versions 1 and 2 and Figures 2 and 3 for versions 3 and 4.

The wearable device 100 is located on the neck of the stuttering person A and consists of a central unit 101 and two half-arcs 102, 103. In embodiment 1, the central unit contains a speaker 120, a device control unit 150, a data transmission unit 170, a switch unit 180, a memory unit 190 . In the left half-arc 102 are: microphone 110, switch on / off device 181, switch for wireless synchronization via Bluetooth 182 with I / O device 300, Bluetooth radio 171. In the right half-arc 103 are: microphone 111, volume switches 183, 184, switch stop / resume recording from sensor unit 185, battery pack 160.

Option 2 includes all of the above in Option 1, plus an optical heart rate monitor 141, an electromyography sensor 142, a temperature sensor 143, an accelerometer 144, a gyroscope 145, and a sensor control unit 140.

Embodiment 3 builds on Embodiment 1 and further includes a video camera 131, a wide-angle camera 132, a light sensor 146, a display 191 and a camera control unit 130, a stop / resume switch for camera 186.

Embodiment 4 builds on Option 2 and further includes a video camera 131, a wide angle camera 132, a light sensor 146, a display 191 and a camera control unit 130, a stop / resume switch for camera 186.

Next, we will use Figure 4. In order for the wearable device 100 to reproduce the speech utterance 220, a number of actions are performed:

1. The wearable device 100 connects via the Bluetooth wireless protocol 230 to the I / O device 300 (smartphone as an example);

2. The stuttering person A types the utterance using the input / output device 300 in the form on the web page containing the graphical interface to the cloud speech generation and recognition service 500, and sends it for generation.

3. The request is transmitted over the wireless mobile network 400 of the 3G / 4G / 5G standard to the speech generation and recognition service 500;

4. In response to the request to the I / O device 300, a cloud-generated utterance in the audio file format is sent over the 3G / 4G / 5G wireless mobile network 400;

5. The audio file is transmitted over Bluetooth 230 to the wearable device 100, which then reproduces the speech utterance 220.

The wearable device 100 can also recognize the speech 210 spoken by the interlocutor B. For this, the person A uses the input / output device (smartphone) to send the audio received from the microphone of the device 100 and transmitted to the smartphone via the Bluetooth protocol 230 to the cloud speech generation and recognition service 500, and in return receives a text representation on I / O 300.

The registration of the readings of sensors 141-146, cameras 131, 132 is carried out in a continuous mode with storage in the memory of the device 190 and further sending via Bluetooth 230 to an input-output device (smartphone) for subsequent saving to a removable storage medium or a cloud service. At the request of the user of the device, registration of readings of sensors and cameras can be stopped / resumed using switches 185.186.

The display 191 in versions 3 and 4 is made using OLED technology and is used for accompanying visualization during the reproduction of a speech utterance, for example, displays a spectrogram or pulsations in order to further draw the attention of the interlocutor B to the sound source of the interlocutor A.

Industrial applicability.

A useful model of a wearable speech aid device is applicable in everyday life to help stuttering people to speak, as well as any other people, regardless of whether they have speech or hearing problems or want to use synthesized voice or recognized speech in everyday life instead of their own. speech and hearing capabilities or in addition to them.

Claims (1)

A wearable speech aid device containing a central unit and left and right half arcs attached to it with the possibility of wearing around the user's neck, while the central unit contains a device control unit, associated data transmission unit, configured to receive an audio recording of a speech utterance, a unit memory, a control unit for switches, a speaker for playing audio recording, in the left half-arc there are a microphone and a device on / off switch, a switch for wireless synchronization via Bluetooth and a Bluetooth radio module, connected to the device control unit, and in the right half - a microphone and a battery pack connected to the device control unit. batteries, and volume switches.

Images