Mobile phone with intoxication detection and call control functions and method for operating the same

Abstract

The present invention relates to a technology that detects a state of intoxication through real-time voice analysis of a mobile phone and provides multi-stage warnings, call quality control, automatic notification sending, and dynamic call blocking functions. Changes in the foremand frequency of a voice signal and other voice parameters are analyzed by an AI module, and said module is implemented as an embedded mobile phone or a downloadable application. It is characterized by preventing social and personal risks by warning the user upon detection of an intoxication state and controlling calls when necessary.

Description

Technology Field

[0001] The present invention relates to real-time voice analysis and call control technology for mobile phones, and in particular to the field of technology combining alcohol detection and an automatic response system. Background Technology

[0002] 1. Limitations of existing technology

[0003] Alcohol detection technology: Low accuracy due to reliance on user behavior patterns instead of voice analysis.

[0004] Voice AI: Existing voice assistant services lack the ability to determine alcohol consumption.

[0005] Drunk driving prevention device: Vehicle-specific, lacking portability and versatility.

[0006] Alcohol prevention app: Requires manual settings and cannot detect alcohol in real time.

[0007] 2. Necessity of Solution

[0008] Increase in social problems (drunk driving, conflict) caused by talking on the phone while intoxicated.

[0009] Absence of voice-based real-time detection and automatic response technology. The problem to be solved

[0010] 1. The problem of being unable to accurately detect the state of intoxication in real time.

[0011] 2. Absence of a call restriction function when intoxicated.

[0012] 3. Lack of an automatic notification system for user safety.

[0013] 4. Limitations of portability and scalability of existing technology. means of solving the problem

[0014] 1. AI Voice Analysis Module

[0015] Complex analysis of voice signal formant frequency changes, vocal jitter, formant bandwidth expansion, etc.

[0016] Implemented as a mobile phone built-in chipset or a downloadable app.

[0017] 2. Multi-stage warning system

[0018] Provides visual (popup), auditory (voice notification), and tactile (vibration) warnings.

[0019] 3. Call Quality Control

[0020] Application of quality degradation techniques such as frequency band limitation, signal delay, and voice distortion.

[0021] 4. Send automatic notifications

[0022] Send a message including GPS location information if the intoxicated state persists for longer than the set time.

[0023] 5. Dynamic Call Blocking

[0024] Automatically adjust blocking time based on the number and pattern of user warnings ignored. Effects of the invention

[0025] 1. Accuracy: Over 95% alcohol detection accuracy through multi-voice parameter analysis.

[0026] 2. Stability: Prevent dangerous situations with emergency alerts and call blocking.

[0027] 3. Scalability: Various implementation options, such as embedded hardware, apps, and cloud integration.

[0028] 4. Security: Prevents personal information leakage through TEE and encryption. Brief explanation of the drawing

[0029] Drawing 1: Overall system block diagram Diagram 2: Internal structure of the AI ​​voice analysis module Figure 3: App-based voice analysis system Diagram 4: Call Blocking Flowchart Drawing 5: Example of a user interface Specific details for implementing the invention

[0030] Figure 1 is an overall system block diagram, Figure 2 is an internal structure of an AI voice analysis module, Figure 3 is an app-based voice analysis system, Figure 4 is a call blocking flowchart, and Figure 5 is an example of a user interface.

[0031] [Figure 1] Overall System Block Diagram

[0032] 110 (microphone)

[0033] function:

[0034] - An input device that collects the user's voice in real time.

[0035] particular:

[0036] - Removes unnecessary frequencies by applying high-frequency (8kHz or higher) and low-frequency (80Hz or lower) filtering.

[0037] - Improves speech signal clarity through an echo reduction algorithm.

[0038] 120 (AI Voice Analysis Module)

[0039] function:

[0040] - Analyze changes in the foremant (F1, F2, F3) frequencies of the collected voice data.

[0041] particular:

[0042] - Determines alcohol status within 0.2 seconds using a CNN+LSTM hybrid model.

[0043] - Voice data is processed using encryption and deleted immediately after analysis is complete.

[0044] 130 (Alcohol Detection Unit)

[0045] function:

[0046] - Calculate the probability of drinking and classify risk levels (low / medium / high) by synthesizing the analysis results of the AI ​​module.

[0047] particular:

[0048] - Activate the warning system (140) when the probability of drinking is 80% or higher.

[0049] - Risk level can be adjusted according to user settings.

[0050] 140 (Warning System)

[0051] function:

[0052] - Provides 3-stage warning (Popup → Voice notification → Screen flashing).

[0053] particular:

[0054] - Warning intensity can be adjusted according to user settings.

[0055] - Warning messages are written in friendly and clear language.

[0056] 150 (Call Quality Control Module)

[0057] function:

[0058] - Call quality is degraded by frequency band limitation (300Hz~3kHz), latency (500ms), and **distortion (echo)**.

[0059] particular:

[0060] - Gradual reduction applied (1 minute difference: 50% reduction → 3 minutes difference: 10% reduction).

[0061] - Users can select the call quality degradation intensity as low / medium / high.

[0062] 160 (Emergency Contact Notification Module)

[0063] function:

[0064] - If intoxication persists for 2 minutes or longer, a message including GPS location information is automatically sent to pre-registered contacts.

[0065] particular:

[0066] - Users can customize message templates and sending conditions (e.g., probability of drinking 90% or higher).

[0067] - Provides a cancel button within 10 seconds before sending the message.

[0068] 170 (User Interface)

[0069] function:

[0070] - Display alcohol detection results and adjust settings (warning intensity, cutoff time).

[0071] particular:

[0072] - Implemented with a touch screen-based GUI and features a user-friendly design.

[0074] [Figure 2] Internal structure of the AI ​​voice analysis module

[0075] 210 (Voice Input Unit)

[0076] function:

[0077] - Preprocess voice data received from the microphone (110).

[0078] particular:

[0079] - 90% noise removal based on Wavelet Transform.

[0080] - Voice data is stored and processed using AES-256 encryption.

[0081] 220 (AI Analysis Module)

[0082] function:

[0083] - Analyze forward patterns by dividing voice frames into 1ms units.

[0084] particular:

[0085] - Processes 5,000 frames per second via GPU acceleration.

[0086] - Analysis results are securely managed in a **TEE (Trusted Execution Environment)**.

[0087] 230 (Feature Extraction Unit)

[0088] function:

[0089] - Extraction of characteristics during drinking (voice jitter ≥1.5%, formant bandwidth ≥1.3x).

[0090] particular:

[0091] - Detect variability by calculating the standard deviation of the frequency spectrum.

[0092] - Extracted features are encrypted and transmitted.

[0093] 240 (Alcohol Detection Unit)

[0094] function:

[0095] - Final determination based on a combination of AI analysis data and extracted features.

[0096] particular:

[0097] - Quantify and convey the probability of drinking as **0~100%**.

[0098] - The judgment result is transmitted via a **low-latency communication protocol (e.g., MQTT)**.

[0099] 250 (Result Delivery Unit)

[0100] function:

[0101] - The judgment result is transmitted to the warning system (140) and the call control module (150).

[0102] particular:

[0103] - Cyclic Redundancy Check (CRC) is applied to ensure data integrity during transmission.

[0105] [Figure 3] App-based voice analysis system

[0106] 310 (Downloadable App)

[0107] function:

[0108] - Perform voice collection and analysis using an application installed by the user.

[0109] - Local AI models can be used in offline mode.

[0110] particular:

[0111] - Developed for Android and iOS platforms, and operates as a background service.

[0112] - You can choose whether to integrate with the cloud based on user settings.

[0113] 320 (Cloud Server)

[0114] function:

[0115] - Voice data analysis using high-performance AI models (optional).

[0116] particular:

[0117] - Operates on AWS EC2 instances and returns analysis results within 0.5 seconds.

[0118] - Voice data is transmitted after anonymization and is deleted immediately after analysis is complete.

[0119] 330 (Local Encrypted Storage)

[0120] function:

[0121] - Store voice data with AES-256 encryption.

[0122] particular:

[0123] - Cloud transmission is possible only with user consent, and data retention is limited to a maximum of 24 hours.

[0124] - Encryption keys are managed in the **TEE (Trusted Execution Environment)**.

[0125] 340 (TEE security zone)

[0126] function:

[0127] - Store contact information and encryption keys in a hardware isolation area.

[0128] particular:

[0129] - Blocks external access by applying Qualcomm Snapdragon's TrustZone technology.

[0130] - Accessible only through biometric authentication (fingerprint / face recognition).

[0132] [Diagram 4] Call Blocking Flowchart

[0133] 410 (Alcohol Detected)

[0134] function:

[0135] - AI module detects over 80% probability of intoxication.

[0136] particular:

[0137] - The judgment threshold can be adjusted by the user within the range of 70–95%.

[0138] - Voice data is processed using encryption and deleted immediately after analysis is complete.

[0139] 420 (1st Warning)

[0140] function:

[0141] - Warning with screen pop-up and vibration.

[0142] particular:

[0143] - Display a friendly and clear message such as "Intoxication detected. Please end the call."

[0144] - Warning messages can adjust language and voice speed according to user settings.

[0145] 430 (Degraded call quality)

[0146] function:

[0147] - Gradually reduce the frequency band.

[0148] particular:

[0149] - 1-minute difference: Frequency band reduced by 50%.

[0150] - 2-minute difference: Frequency band reduced by 30% and echo added.

[0151] - 3-minute difference: 10% reduction in frequency band and voice interruptions at 3-second intervals.

[0152] 440 (Force Termination)

[0153] function:

[0154] - Call terminated and blocked if 3 warnings are ignored.

[0155] particular:

[0156] - Blocking time is automatically adjusted from 1 to 24 hours based on user behavior logs (e.g., detection of 3 or more occurrences within 24 hours).

[0157] - Users can adjust the blocking time in the settings menu.

[0159] [Figure 5] Example of user interface

[0160] 1. Components:

[0161] Warning message popup: "Intoxication detected. Please end the call." Includes text.

[0162] Call quality degradation indicator bar: Graphically displays the degree of frequency band reduction.

[0163] Settings menu screen: Displays alert intensity, blocking time, and contact settings options.

[0164] 2. Design:

[0165] Represents the smartphone screen similarly to the real thing.

[0166] Each element is designed to be intuitive and user-friendly.

[0168] Integrated Implementation Scenario

[0169] 1. Voice Collection:

[0170] The microphone (110) collects the user's voice → preprocessing (210).

[0171] Voice data is stored and processed using AES-256 encryption.

[0172] 2. AI Analysis:

[0173] Analysis of changes in foremand (F1, F2, F3) (220) → Alcohol determination (240).

[0174] Analysis results are safely managed in a **TEE (Trusted Execution Environment)**.

[0175] 3. Execute Warning:

[0176] After providing a 3-step warning (140), call quality degradation (150) is applied.

[0177] Warning messages can be adjusted in language and voice speed according to user settings.

[0178] 4. Urgent Alert:

[0179] If intoxication persists for 2 minutes or longer, a message including GPS location information is automatically sent to pre-registered contacts.

[0180] Provides a cancel button within 10 seconds before sending the message.

[0181] 5. Call Blocking:

[0182] Call terminated and blocked if 3 warnings are ignored.

[0183] Blocking time is automatically adjusted from 1 to 24 hours based on user behavior logs (e.g., detection of 3 or more occurrences within 24 hours).

[0184] This allows for the systematic management of all risks associated with making calls while intoxicated.

[0185] As such, while specific embodiments have been described in the detailed description of the present invention, it is understood that various modifications are possible within the scope of the invention.

[0186] Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims set forth below as well as equivalents thereof. Explanation of the symbols

[0187] 110: Microphone 120: AI voice analysis module 130: Alcohol detection unit 140: Warning system 150: Call quality control module 160: Emergency contact notification module 170 : User interface 210 : Voice input unit 220: AI Analysis Module 230: Feature Extraction Unit 240: Alcohol Detection Unit 250: Result Delivery Unit 310: Downloadable App 320: Cloud Server 330: Local Encrypted Storage 340: TEE Security Zone 410: Alcohol Detected 420: First Warning 430: Poor call quality 440: Force close

Claims

Claim 1 A mobile phone for detecting a state of intoxication through real-time voice analysis, comprising an AI module that analyzes multiple forward frequency changes of a voice signal and a control module that determines a state of intoxication based on the analysis results of the AI ​​module, wherein the AI ​​module is embedded in the operating system of the mobile phone or implemented as a downloadable application. Claim 2 A mobile phone that detects a state of intoxication through real-time voice analysis comprises an AI module that analyzes multiple forward frequency changes of a voice signal and a control module that determines a state of intoxication based on the analysis results of the AI ​​module, wherein the control module provides a multi-stage warning to a user upon detection of a state of intoxication, the multi-stage warning includes a visual notification, an auditory notification, and a tactile notification, and wherein the control module degrades the call quality in a state of intoxication by limiting the frequency band, signal delay, or voice distortion in one or more ways. Claim 3 A portable phone according to claim 1 or claim 2, wherein the AI ​​module further analyzes the vocal jitter and formant bandwidth of a voice signal. Claim 4 A portable phone according to claim 1 or claim 2, wherein the downloadable application analyzes voice data in conjunction with a cloud server. Claim 5 A portable phone according to claim 1 or claim 2, wherein the AI ​​module converts voice data into the frequency domain and analyzes it.

Images