A stuffed toy-type voice dialogue device with speech response and play functions

Abstract

To provide a stuffed toy type voice dialogue device that can respond flexibly and naturally to a user's speech and has functions such as a word chain game and music playback, and can be used independently of a communication environment. [Solution] This device is a voice dialogue device that is housed in a stuffed toy-shaped housing and equipped with a voice recognition means, a voice output means, and a dialogue control means, and is configured to selectively output recorded voice or synthesized voice.It also has a built-in word chain response function, music playback function, LED display, battery, and Qi charging structure, and is capable of operating completely offline.

Description

[Technical Field]

[0001] This invention relates to a plush toy-type voice dialogue device, primarily intended to support dialogue with and monitor elderly people and children. More specifically, it relates to a configuration that can operate completely offline, responding to user utterances with recorded or synthesized voice, and also equipped with game and music playback functions. [Background technology]

[0002] Until now, robots and devices capable of voice interaction have relied mainly on cloud-based voice recognition and AI responses, and most have relied on a communication environment. Even stuffed animal toys have been limited to playing standard recorded voices, with few offering flexible responses or learning capabilities. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2025-000882 [Patent Document 2] Patent Publication No. 2024-55866 [Patent Document 3] Patent Publication No. 2023-155861 [Non-patent literature]

[0004] [Non-Patent Document 1] "https: / / robostudy.jp / feature1.html" (Robosta - Special article on monitoring robots) Summary of the Invention [Problem to be solved by the invention]

[0005] The objective of this invention is to provide a voice dialogue device that can operate even in an unstable communication environment, and to enable natural dialogue and games (such as Shiritori) with the elderly, thereby contributing to the maintenance of emotional support and cognitive function in daily life. [Means for solving the problem]

[0006] The device according to the present invention includes the following components: -Audio output section that can switch between pre-recorded audio and synthesized audio A voice recognition unit that responds to user utterances Response selection structure based on dictionary (wav_map) according to speech recognition results A control unit that performs interactive games such as "Shiritori" in response to voice input -Music playback function that responds to the user's mood Completely offline operation (with Raspberry Pi) This device can use the Qi method as a wireless power supply method, but it can also be powered by a wired method using a USB cable. [Effects of the Invention]

[0007] This invention provides the following effects: -Provides a highly functional interactive device that can be used anywhere without the need for communication · Language stimulation is possible through games such as Shiritori. - BGM playback allows for a performance that matches the user's emotions -The learning function allows you to build a continuous relationship with users. Furthermore, the stuffed toy-type housing of this invention is intentionally designed without facial expressions (eyes, mouth, etc.), allowing users to freely project their emotions and personalities onto it, and can be used as a universal interactive device that is not dependent on a specific character. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a diagram showing the external configuration of a stuffed toy-type interactive device according to the present invention; [Figure 2] Cross-sectional view showing the internal structure of the device DETAILED DESCRIPTION OF THE INVENTION

[0009] The device is a stuffed toy measuring approximately 35-40cm in height, and contains a Raspberry Pi, microphone, speaker, battery unit, charging coil, LED, and various control boards. The exterior is made of soft cloth, giving it a reassuring appearance. Voice responses are provided by switching between pre-recorded WAV files and synthesized voices generated by VOICEVOX. The interactive stuffed toy device of the present invention can employ a non-contact power supply method as a power supply means. For example, a wireless charging coil conforming to the Qi standard can be built into the back of the device, and charging can begin automatically when the device is placed on a compatible power supply base. In addition to the wireless power supply described above, it is also possible to use a wired power supply method via a general USB Type-C terminal. The control unit 8 in this embodiment is configured using a Raspberry Pi, and comprehensively executes voice recognition, response selection, voice synthesis, recording of user status, control of external input and output, and the like. [Example]

[0010] When the user says "hello," the recorded voice file greet_konnichiwa.wav is played. When the user says "let's play shiritori," the device starts shiritori using the built-in dictionary, converts the user's voice into hiragana, and responds accordingly. Various settings of this device can be configured to be performed via a dedicated application installed on an external device such as a smartphone. This application connects to the device via communication means such as Bluetooth or Wi-Fi, and allows you to register a user name, adjust the voice volume, select a response pattern, specify the music to play, limit usage time, and make the following individual settings. - **The name the user wants to be called** (e.g., "Mom," "Takashi-kun," etc.) - **Plushie's name** (e.g. "Masaru", "Lily", etc.) - **Setting the gender of the stuffed animal** (e.g. male character / female character / neutral character) - **Select the voice tone of your stuffed toy** (e.g., energetic boyish, gentle feminine, cool masculine, etc.) This device has the ability to switch between multiple voice synthesis styles (such as voice libraries such as VOICEVOX), making it easy to give the user "voice individuality" according to their preferences. This allows for a stronger attachment to be formed, and increases user satisfaction. [Industrial Applicability]

[0011] This invention can be applied to a wide range of fields, including monitoring, elderly care, educational toys, and interactive promotional tools. [Explanation of symbols]

[0012] 1: Plush toy case 2: Raspberry Pi 3: Audio input section (microphone) 4: Audio output section (speaker) 5: Qi charging coil 6:LED (nose) 7: Switch (head) 8: Control unit (including software control and dictionary) 9: Battery

Claims

1. An interactive stuffed toy device comprising a body of the stuffed toy, which comprises a voice recognition means, a response generation means, a voice output means, and a control means for selecting between pre-recorded voice and voice synthesis, and wherein the control means selects one of the voice output means to respond based on a user's utterance, and controls an LED provided on the body of the stuffed toy to visually display the response status.

2. 2. The interactive stuffed toy device according to claim 1, further comprising a power supply means built into the back surface thereof, and configured to be rechargeable when placed on the base.

3. 2. The interactive stuffed toy device according to claim 1, further comprising a Shiritori function for normalizing a user's voice input into hiragana characters, selecting a word corresponding to the ending of the previous word, and outputting it by voice.

4. 2. The interactive stuffed toy device according to claim 1, further comprising a music playback function for storing music data corresponding to a preset mood or genre and playing corresponding music based on a user's input.

5. 4. The interactive stuffed toy device according to claim 3, further comprising a storage means for recording words obtained by the user's speech in the Shiritori function, and making the recorded words available for subsequent Shiritori responses.

Images