The design of speech recognition system based on AVR microcontroller, the system takes AVR microcontroller as the control core to realize the recognition and control of human voice. The main control chip of the system is atmegal28 of atreel company, and the speech recognition function is ld3320 of ICR oute company. Ld3320 is integrated with speech recognition algorithm, without external flash and ram resources, and can complete speaker independent speech recognition tasks well. At the same time, the chip integrates MP3 playing function, supports MPEG and other formats, and can realize voice prompt or MP3 song playing function. Because it contains 16 bit a / D, D / a converter and power amplifier circuit, it can produce clear sound without external power amplifier circuit. The system has reserved various interfaces and has good expansibility.
ATmega128 chip produced by ATMEL company is used in the controller. Advanced RISC structure is adopted, and 128 KB flash, 4 KB SRAM, 4 KB E2PROM and other rich resources are built in. The chip is an 8-bit microprocessor with high performance and low power consumption in the industry, and is widely used in the 8-bit MCU market.
Ld3320 speech recognition circuit
Ld3320 is a special chip for speech recognition. The chip integrates speech recognition processor and some external circuits, including a / D, D / a converter, microphone interface, voice output interface, etc., and can play MP3. Without any external auxiliary chips such as flash, ram, etc., it can be directly integrated into the product to realize the functions of speech recognition, voice control and man-machine dialogue. Figure 3 is the schematic diagram of ld3320 circuit. SPI bus mode is adopted for communication with MCU, and the clock can not exceed 1.5MHz.
The working circuit of microphone is shown in the figure. The audio output only needs to connect the speaker to spop and spon. When using SPI bus mode, MD of ld3320 should be set to high level, and SPIs should be set to low level. SPI bus pins include SDI, SDO, sdck and SCS. Intb is the interrupt port. When there is recognition result or MP3 data is insufficient, it will trigger the interrupt and inform MCU to process. Rstb pin is ld3320 reset end, low level is valid. LED1 and LED2 are used as power on indicators.
The feasibility of speech recognition system based on AVR MCU is discussed, and the design scheme is given. The test results show that the system has the advantages of stable operation, high speech recognition rate and low cost. At the same time, with the help of the MP3 playing function of ld3320, the system has certain interactivity and entertainment. In the aspect of portability, ld3320 driver can be easily transplanted to various embedded systems through simple modification. With the demand of artificial intelligence, speech recognition technology will be more and more concerned by people. It is believed that speech recognition will have a broader application in the near future.