A star flash voice collection module for LLM robot task planning

Abstract

The utility model discloses a kind of star flash voice collection module for LLM robot task planning, comprising: DC power module, for providing stable DC voltage;Main control chip module, for the voice audio data collected is sent by star flash wireless communication;Voice collection module, for collecting voice audio data;LED lamp output display module, for the state of wireless data transmission is shown;Star flash wireless communication module, for the wireless transmission of data;The utility model will analog microphone, low-power audio codec chip and star flash wireless communication chip are highly integrated, utilize the microsecond level delay and super anti-interference ability of star flash technology, to solve the instruction recognition error and planning failure problem caused by delay and packet loss of traditional wireless scheme, provide stable and reliable auditory perception guarantee for LLM.

Description

Technical Field

[0001] This utility model relates to the technical field of LLM robot task planning, and in particular to a StarFlash voice acquisition module for LLM robot task planning. Background Technology

[0002] In current robot task planning technologies based on Large Language Models (LLMs), accurate understanding of natural language instructions is crucial for achieving advanced autonomous decision-making, which heavily relies on the real-time performance and reliability of front-end voice acquisition. However, existing wireless voice transmission solutions commonly used in robots, such as Wi-Fi or classic Bluetooth, exhibit significant shortcomings in complex real-world application scenarios: their high and unstable transmission latency easily leads to delayed instruction response and LLM parsing errors, while their weak anti-interference capabilities cause data packet loss, resulting in incomplete or inaccurate instructions. These factors collectively restrict the fluency of human-computer interaction and the accuracy of task planning, becoming a bottleneck that urgently needs to be overcome in this technological field. Utility Model Content

[0003] The purpose of this invention is to overcome the shortcomings and deficiencies of the prior art and provide a StarFlash voice acquisition module for LLM robot task planning. It highly integrates an analog microphone, a low-power audio codec chip and a StarFlash wireless communication chip, providing stable and reliable auditory perception for LLM.

[0004] To achieve the above objectives, the technical solution provided by this utility model is: a StarFlash voice acquisition module for LLM robot task planning, comprising:

[0005] DC power supply module, used to provide stable DC voltage;

[0006] The main control chip module is used to transmit the collected voice audio data via StarScan wireless communication;

[0007] The voice acquisition module is used to collect voice audio data;

[0008] The LED light output display module is used to display the status of wireless data transmission;

[0009] The StarScan wireless communication module is used for wireless data transmission.

[0010] The DC power supply module is powered by a lithium battery (3.7-4.2V, 1200mAh), which is connected to the main control chip module. It generates 3.3V through a voltage regulator chip (LM1117-3.3V) to power the main control chip module. The main control chip module uses a StarSpark WS63E, which has a standard digital audio interface and integrates a StarSpark wireless communication interface for wireless transmission of acquired voice audio data. The voice acquisition module connects to a single-channel stereo audio analog-to-digital converter (CS5341) via an analog microphone. This converter is connected to the main control chip module and transmits voice audio data to the WS63E through the digital audio interface. The LED output display module is connected to the main control chip module and features a single LED circuit. It directly generates high and low levels to drive the LEDs to turn on and off by connecting a current-limiting resistor. The StarSpark wireless communication module connects to an onboard antenna on the PCB to achieve wireless data transmission based on the WS63E.

[0011] Preferably, the voice acquisition module is connected to the VINL pin of the CS5341 via the output pin of the analog microphone to transmit voice audio data; the VINR pin of the CS5341 is grounded and configured as a single-ended input mode; the MD0 pin and MD1 pin of the CS5341 are connected to a high level of 3.3V, configured as a slave mode, automatically detecting the master clock frequency and sampling rate; the VA pin of the CS5341 is connected to 3.3V to obtain analog power; the VD pin of the CS5341 is connected to 3.3V to obtain digital power; the GND pin of the CS5341 is connected to ground; the REF_GND pin of the CS5341 is connected to ground; the VQ pin of the CS5341 is connected to a first resistor. The pull-down pin is connected to ground (GND); the RST pin of the CS5341 is connected to the GPIO00 pin of the WS63E, and the GPIO00 pin of the WS63E generates a low-level pulse to reset the CS5341; the SDOUT pin of the CS5341 is connected to the I2S_DI pin of the WS63E to realize the acquisition of voice audio data by the WS63E; the I2S_MCLK pin of the WS63E is connected to the MCLK pin of the CS5341 as the master clock; the I2S_SCLK pin of the WS63E is connected to the SCLK pin of the CS5341 as the bit clock; the I2S_LRCLK pin of the WS63E is connected to the LRCLK pin of the CS5341 as the left and right channel clocks.

[0012] Preferably, the main control chip module uses a star-shaped flash chip, which connects to the corresponding pins of the CS5341 via the I2S_DI, I2S_MCLK, I2S_SCLK, and I2S_LRCLK pins of the WS63E to acquire voice audio data; the VDD pin of the WS63E is connected to a 3.3V power supply, and the GND pin of the WS63E is connected to ground GND to realize the power supply circuit; the GPIO001 pin of the WS63E directly generates high and low levels to drive the LED to turn on and off by connecting a current-limiting second resistor; the RFIO pin of the WS63E is connected to the onboard antenna of the PCB through a third resistor.

[0013] Compared with the prior art, this utility model has the following advantages and beneficial effects:

[0014] This invention highly integrates an analog microphone, a low-power audio codec chip, and a StarFlash wireless communication chip. By utilizing the microsecond-level latency and exceptional anti-interference capabilities of StarFlash technology, it solves the problems of instruction recognition errors and planning failures caused by latency and packet loss in traditional wireless solutions, providing stable and reliable auditory perception assurance for LLM. Attached Figure Description

[0015] Figure 1 This is a simplified structural diagram of the present invention.

[0016] Figure 2 This is a circuit diagram related to the voice acquisition of this utility model. Detailed Implementation

[0017] The present invention will be further described below with reference to specific embodiments.

[0018] like Figure 1 and Figure 2 As shown, this embodiment discloses a StarFlash voice acquisition module for LLM robot task planning, including:

[0019] DC power supply module 1 is used to provide a stable DC voltage;

[0020] The main control chip module 2 is used to transmit the collected voice audio data via StarScan wireless communication;

[0021] Voice acquisition module 3 is used to acquire voice audio data;

[0022] LED light output display module 4 is used to display the status of wireless data transmission;

[0023] The StarScan wireless communication module 5 is used for wireless data transmission;

[0024] The DC power supply module is powered by a lithium battery (3.7-4.2V, 1200mAh), which is connected to the main control chip module. It generates 3.3V through a voltage regulator chip (LM1117-3.3V) to power the main control chip module. The main control chip module uses a StarSpark WS63E, which has a standard digital audio interface and integrates a StarSpark wireless communication interface for wireless transmission of acquired voice audio data. The voice acquisition module connects to a single-channel stereo audio analog-to-digital converter (CS5341) via an analog microphone. This converter is connected to the main control chip module and transmits voice audio data to the WS63E through the digital audio interface. The LED output display module is connected to the main control chip module and features a single LED circuit. It directly generates high and low levels to drive the LEDs to turn on and off by connecting a current-limiting resistor. The StarSpark wireless communication module connects to an onboard antenna on the PCB to achieve wireless data transmission based on the WS63E.

[0025] Specifically, the analog microphone transmits voice to the CS5341 by connecting its output pin to the VINL pin (pin 10) of the CS5341, and connecting the VINR pin (pin 12) to ground (GND).

[0026] For the CS5341 to function properly, its VA pin (pin 13) needs to be connected to 3.3V to obtain analog power; the VD pin (pin 6) needs to be connected to 3.3V to obtain digital power; the GND pin (pin 5) and REF_GND pin (pin 14) need to be connected to ground (GND) to obtain ground signal; and the VQ pin (pin 11) needs to be pulled down to ground (GND) through the first resistor.

[0027] The MCLK pin (pin 2) of the CS5341 is connected to the I2S_MCLK pin (pin 15) of the StarShine WS63E; the SCLK pin (pin 7) of the CS5341 is connected to the I2S_SCLK pin (pin 17) of the StarShine WS63E; the LRCK pin (pin 8) of the CS5341 is connected to the I2S_LRCLK pin (pin 16) of the StarShine WS63E; the M0 pin (pin 1) and M1 pin (pin 1) of the CS5341 are both connected to 3.3V, setting the CS5341 chip to slave mode, so that its related clock pins MCLK, SCLK, and LRCK receive the clock signals sent from the master StarShine WS63E pins I2S_MCLK, I2S_SCLK, and I2S_LRCLK.

[0028] The GPIO00 pin (pin 11) of the StarShine WS63E is connected to the RST pin (pin 9) of the CS5341 to initialize the CS5341; the I2S_DI pin (pin 39) of the StarShine WS63E is connected to the SDOUT pin (pin 4) of the CS5341 to acquire voice audio data from the CS5341; the GPIO01 pin (pin 6) of the StarShine WS63E is connected to the LED and the second resistor to ground (GND) to control the LED's on / off state; the RFIO pin (pin 2) of the StarShine WS63E is connected to the third resistor to the onboard antenna on the PCB to enable wireless data transmission via the StarShine WS63E.

[0029] The VDD pins (pins 9 and 22) of the StarShine WS63E are connected to a 3.3V power supply, and the GND pins (pins 1, 3, 20, 29, 31, 36, and 42) of the StarShine WS63E are connected to ground (GND) to enable the power supply circuit to function properly.

[0030] The above-described embodiments are merely preferred embodiments of this utility model and are not intended to limit the scope of implementation of this utility model. Therefore, all changes made in accordance with the shape and principle of this utility model should be covered within the protection scope of this utility model.

Claims

1. A StarFlash voice acquisition module for LLM robot task planning, characterized in that, include: DC power supply module, used to provide stable DC voltage; The main control chip module is used to transmit the collected voice audio data via StarScan wireless communication; The voice acquisition module is used to collect voice audio data; The LED light output display module is used to display the status of wireless data transmission; The StarScan wireless communication module is used for wireless data transmission. The DC power supply module is powered by a lithium battery (3.7-4.2V, 1200mAh), which is connected to the main control chip module. It generates 3.3V through a voltage regulator chip (LM1117-3.3V) to power the main control chip module. The main control chip module uses a StarSpark WS63E, which has a standard digital audio interface and integrates a StarSpark wireless communication interface for wireless transmission of acquired voice audio data. The voice acquisition module connects to a single-channel stereo audio analog-to-digital converter (CS5341) via an analog microphone. This converter is connected to the main control chip module and transmits voice audio data to the WS63E through the digital audio interface. The LED output display module is connected to the main control chip module and features a single LED circuit. It directly generates high and low levels to drive the LEDs to turn on and off by connecting a current-limiting resistor. The StarSpark wireless communication module connects to an onboard antenna on the PCB to achieve wireless data transmission based on the WS63E.

2. The StarFlash Voice Acquisition Module for LLM Robot Task Planning according to claim 1, characterized in that, The voice acquisition module connects to the VINL pin of the CS5341 via the output pin of the analog microphone to transmit voice audio data. The VINR pin of the CS5341 is grounded and configured for single-ended input mode. The MD0 and MD1 pins of the CS5341 are connected to a high level of 3.3V, configured as slave mode, automatically detecting the master clock frequency and sampling rate. The VA pin of the CS5341 is connected to 3.3V to obtain analog power. The VD pin of the CS5341 is connected to 3.3V to obtain digital power. The GND pin of the CS5341 is connected to ground (GND). The REF_GND pin of the CS5341 is connected to ground (GND). The VQ pin of the CS5341 is pulled down through a first resistor. Connect the CS5341 to ground (GND); connect the CS5341's RST pin to the WS63E's GPIO00 pin, which generates a low-level pulse to reset the CS5341; connect the CS5341's SDOUT pin to the WS63E's I2S_DI pin to enable the WS63E to acquire voice audio data; connect the WS63E's I2S_MCLK pin to the CS5341's MCLK pin as the master clock; connect the WS63E's I2S_SCLK pin to the CS5341's SCLK pin as the bit clock; connect the WS63E's I2S_LRCLK pin to the CS5341's LRCLK pin as the left and right channel clocks.

3. The StarFlash Voice Acquisition Module for LLM Robot Task Planning according to claim 2, characterized in that, The main control chip module uses a star-shaped flash chip. It connects to the corresponding pins of the CS5341 via the I2S_DI, I2S_MCLK, I2S_SCLK, and I2S_LRCLK pins of the WS63E to acquire voice audio data. The VDD pin of the WS63E is connected to a 3.3V power supply, and the GND pin of the WS63E is connected to ground (GND) to implement the power supply circuit. The GPIO001 pin of the WS63E directly generates high and low levels to drive the LED to turn on and off by connecting a current-limiting second resistor. The RFIO pin of the WS63E is connected to the onboard antenna of the PCB through a third resistor.

Images