Description
The LU-ASR01 is a compact offline voice recognition development board powered by the TW-ASR ONE microcontroller, designed for speech-controlled applications that don’t require an internet connection. With its built-in microphone, 3 W audio amplifier, and temperature sensor interface, this board enables local speech recognition, voice response, and I/O control — all from a single, highly integrated platform.
It supports custom wake words and voice commands, GPIO and PWM outputs, and direct programming via USB Type-C, making it ideal for smart home, robotics, and IoT voice interaction projects.
Key Features
- Offline Speech Recognition: Operates entirely without a network or cloud connection.
- High Accuracy: Up to 98 % recognition rate with a 10 m wake-up range.
- Custom Commands: Define up to 5 wake-up words and 200 recognition commands (expandable to 300 entries).
- Built-in Audio System: Integrated microphone, 3 W amplifier, and speaker connector (supports 4 Ω/3 W speaker).
- 12 through holes, including 8 GPIOs (6 PWM-capable)
- 1 UART Tx, 1 DHT11/DS18B20 temperature sensor interface
- Flexible I/O:
- Plug-and-Play Programming: Powered and flashed via USB Type-C, featuring an onboard CH340C USB-to-TTL converter.
- Visual Development: Supports Blockly-based TwenBlock IDE for drag-and-drop programming and direct entry modification.
- FreeRTOS Operating System: Ensures multitasking and responsive real-time performance.
- No Coding Required: Voice entries can be configured and burned directly through the provided GUI.
Specification
- Microcontroller: TW-ASR ONE (Twen ASR ONE)
- CPU: 32-bit RISC core with BNPU (Voice Neural Processing Unit)
- Clock Speed: Up to 240 MHz
- RAM: 512 – 640 KB SRAM
- Flash Memory: 4 MB (on-chip)
- Operating Voltage: 3.6 – 5.0 V DC
- Current Requirement: ≥ 500 mA
- Interfaces: UART / I²C / PWM / SPI / GPIO
- Audio I/O: Built-in microphone, 2-pin speaker header with 3 W amplifier (4 Ω/3 W speaker supported)
- Recognition Latency: < 0.1 s
- Recognition Distance: Up to 10 m (open area)
- Recognition Accuracy: ≈ 98 %
- Serial Speed: 9600 bps (default, configurable)
- Operating Temperature: –20 °C to +80 °C
- Dimensions: 37 mm × 37 mm
- Weight: ≈ 35 g
- Power & Programming Port: USB Type-C
Application
- Voice-activated smart-home devices
- IoT and embedded AI voice interfaces
- Robotics and automation control
- Offline personal assistants and voice toys
- Educational & STEM learning projects
- Custom DIY voice command systems
Software & Development
- Supported IDE: TwenBlock (Blockly-based visual environment)
- Programming Method: Drag-and-drop voice definition and command mapping
- Firmware Flashing: One-click USB upload (via Type-C)
- Operating System: FreeRTOS
- Language Support: Currently Mandarin (English support undocumented)
- Note: English support is limited