Low power offline speech recognition module

Basic Parameters

Dimensions: 16*18*2.4mm

Package Type: SMD16

Peripheral Interfaces: UART/GPIO/PWM/SPI/IPC/I²S/ADC

Number of IO Ports: Up to 8 GPIO

Operating Voltage: 3.3V (±0.3V)

Power Consumption: Standby current 0.9mA, maximum current 8mA

Operating Temperature: -20℃ to +85℃

Storage Temperature: -40℃ to +85℃

Operating Humidity: 10% to 95% (non-condensing)

Product Description

The low-power offline voice recognition module is designed for battery-powered and portable products, as well as various products with stringent power consumption requirements. Leveraging next-generation voice recognition technology and algorithms, it deeply integrates offline recognition algorithms with chip architecture, providing customers with a comprehensive voice recognition solution. It can be widely and rapidly applied to various battery-powered small appliances, wearables, toys, and voice-controlled products like single-wire 86 boxes.

The module uses an MCU + voice recognition dedicated NPU core architecture, adopting ultra-low power process technology. It has a built-in high-precision voice detection module and a multi-level system startup mode, reducing standby power consumption to sub-milliwatt levels and operating power consumption to a few milliwatts. With built-in SRAM and FLASH, it requires only a few peripheral components to form a complete solution.

Product Features:

Single Microphone Pickup
- Single microphone solution.
- Supports a 3-5 meter range in home environments.
Voice Wake-up
- High-performance wake-up engine.
- Low power consumption.
- Supports Mandarin with accents.
- Low false wake-up rate (< 1 false wake-up in 24 hours).
Offline Recognition
- Supports recognition of 50 local control commands.
Multi-turn Conversation
- Continuous conversation with a single wake-up, making voice operation more convenient and natural.
Customization Options
- Custom wake-up words.
- Custom command words.
- Custom response playback.
- Personalized acoustic models.
- Voice actor tone selection.
- Other product-specific custom configurations.
Rich Interfaces
- 2 UART communication interfaces.
- Audio interfaces: I2S, ADC, DAC.
- GPIO, I2C, SPI, PWM and other expansion interfaces.

SOC Functional Diagram

Product Size

Application Scenarios:

Smart Home Appliances Adjusts the working mode of appliances (on, standby, off) in real-time based on the presence of people in the room, achieving smart appliance management.
Human Sensing Lighting Detects whether a space is occupied and automatically controls lighting, applicable in public spaces, offices, and various sensing lights.
Smart Home Scenarios For homes, hotels, offices, bathrooms, and other locations, it detects real-time occupancy or presence of people, enabling security, appliance control, and personnel monitoring. It can be integrated with IoT platforms to effectively manage these environments.

Product Applications

The module can serve as a main controller in scenarios such as voice-controlled LED lights and voice-controlled relays. It can be applied in the following settings: voice-controlled chandeliers, wall lamps, bathroom heaters, switches, spotlights, ceiling lamps, desk lamps, wall switches, hotel control panels, LED desk lamps, panels, clothes dryers, electric curtains, fans, smart door locks, robotic vacuum cleaners, smart desk lamps, smart air conditioners, smart kettles, story machines, smart curtains, smart fans, voice-controlled speakers, and in-car voice control.