Control method of multidirectional sound receiver based on chassis type robot

Abstract

The invention relates to the technical field of robot control, in particular to a control method of a multidirectional sound receiver based on a chassis type robot, which comprises the following steps of: S10, controlling the sound receiver to perform rotary scanning, recording the angle of the sound source when the sound source is detected, selecting a target sound source from a plurality of sound sources, and recording the angle r1 of the target sound source; s20, calculating an optimal rotation path and angle based on the angle r1, controlling the sound receiver to rotate, and correcting a rotation error; and step S30, controlling the sound receiver directly facing the sound source to carry out audio acquisition at a preset sampling frequency Fs and sampling precision Ps. Through active scanning and optimal path rotation, the sound receiver is rapidly and accurately aligned with a main sound source, the sound pickup effect is remarkably improved, the fidelity of audio collection is effectively ensured by adopting the design of an arc-shaped sound pickup surface and combining high sampling rate and sampling precision, the sound receiver is integrated on a chassis which is generally not shielded by clothes, sound wave attenuation is avoided, and the sound pickup efficiency is improved. And the sound source characteristics are enriched.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot control, in particular to a control method of a multi-directional sound receiver based on a chassis robot. BACKGROUND

[0002] Robots are a new industry in recent years, and robots for various industries have developed rapidly. Among them, chassis robots have gained great application in travel and tourism interpretation, character interaction, etc. due to their stability and safety.

[0003] In most existing chassis robots, the sound receiver is mostly fixedly installed, which leads to poor sound collection effect for specific angle communication in a multi-subject interaction scenario. The robot can only realize alignment by passive angle adjustment, which has a strong sense of fragmentation in human-computer interaction experience and leads to poor user experience.

[0004] Therefore, we propose a control method of a multi-directional sound receiver based on a chassis robot. SUMMARY

[0005] The present application aims to provide a control method of a multi-directional sound receiver based on a chassis robot, which has the advantage of good interactive experience and solves the problem of multi-person interaction of chassis robots in complex and noisy scenarios.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a control method of a multi-directional sound receiver based on a chassis robot, comprising the following steps: Step S10: control the sound receiver to rotate and scan, record the angle when detecting the sound source, and select a target sound source from multiple sound sources and record its angle r1; Step S20: calculate the optimal rotation path and angle based on the angle r1, control the sound receiver to rotate and correct the rotation error; Step S30: control the sound receiver facing the sound source to collect audio at a preset sampling frequency Fs and sampling accuracy Ps; Step S40: convert the collected audio signal into a digital audio file; Step S50: convert the digital audio file into text and perform interactive processing, and finally reply through voice playback.

[0007] Preferably, in step S10, the specific way of selecting a target sound source from multiple sound sources is to select the sound source with the largest loudness as the target sound source.

[0008] Preferably, in the step S20, the specific method for calculating the optimal rotation path and angle is that the initial angle of the sound receiver is 0, and the shortest angle Q to be rotated is min(r1, 360-r1).

[0009] Preferably, in the step S20, the sound receiver is controlled to rotate at a constant angular velocity w, and the theoretical rotation time Ttheory is Q / w.

[0010] Preferably, in the step S20, the specific method for correcting the rotation error is that the actual rotation time Tturn is recorded, when Tturn≠Ttheory, the current angle is updated according to the formula R=Tturn×w, and the deviation |Q-R| from the target angle Q is calculated. If Tturn<Ttheory, the sound receiver is controlled to continue rotating in the original direction by |Q-R| degrees. If Tturn>Ttheory, the sound receiver is controlled to rotate in the opposite direction by |Q-R| degrees.

[0011] Preferably, the error correction is single correction.

[0012] Preferably, in the step S30, the sampling frequency Fs is at least 40000Hz.

[0013] Preferably, in the step S30, the sampling precision Ps is at least 16bit.

[0014] Preferably, in the step S40, the size of the digital audio data after analog-to-digital conversion is S_data=Ps×Fs×T2, wherein T2 is the conversion time.

[0015] Preferably, in the step S10, when no sound source is detected, the sound receiver enters a silent state and automatically restarts scanning after an interval time t1.

[0016] Compared with the prior art, the present application has the following advantages: 1. The present application actively scans and rotates along the optimal path to quickly and accurately align the sound receiver with the main sound source, significantly improving the sound pickup effect.

[0017] 2. The present application adopts a circular arc-shaped sound pickup surface design, combined with high sampling rate and sampling precision, effectively ensuring the fidelity of audio collection.

[0018] 3. The present application integrates the sound receiver on the chassis which is usually not covered by clothes, avoiding sound wave attenuation and enriching the sound source characteristics.

[0019] 4. The present application has a clear and reliable control method flow, and is suitable for multi-subject dialogue scenes such as tourism and education. BRIEF DESCRIPTION OF DRAWINGS

[0020] Fig. 1This is the flowchart of the method of the present invention; Fig. 2 This is the control method diagram of the multi-directional microphone of the chassis-type robot of the present invention; Fig. 3 This is the audio information processing link diagram of the present invention. Specific implementation manner

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention.

[0022] Please refer to Figs. 1-3 As shown, the present invention provides a technical solution: a control method for a multi-directional microphone based on a chassis-type robot, including the following steps: Step S10: Sound source detection and angle recording.

[0023] Control the microphone to perform a 360° rotation scan; During the scan, continuously detect whether there is a sound source at the current angle; If there is no sound source, the microphone enters the silent state and restarts the scan after an interval time t1; If there is a sound source, collect the sound wave vibration through the pickup surface of the arc-shaped vertical plate and convert it into an electrical signal containing loudness characteristics, and at the same time record the current angle r when the sound source is detected; When there are multiple sound sources, select the sound source with the largest loudness as the target sound source and record its corresponding angle r1.

[0024] Step S20: Optimal path rotation and error correction.

[0025] Set the initial angle of the microphone to 0, and calculate the shortest angle Q to be rotated = min(r1, 360 - r1); According to the rotation direction and angle Q, and the preset constant angular velocity w, calculate the theoretical rotation time T theory = Q / w; Control the microphone to start rotating and continuously record the actual rotation time T rotation; When T rotation = T theory, it is determined that the rotation is in place; When T rotation ≠ T theory, update the current angle according to the formula R = T rotation × w and calculate the angle deviation |Q - R|; If T rotation < T theory, control the microphone to continue rotating |Q - R| degrees in the original direction; If T rotation > T theory, control the microphone to rotate |Q - R| degrees in the reverse direction to complete a single error correction.

[0026] Step S30: High-fidelity audio collection.

[0027] Control the microphone facing the sound source to start the audio collection process; The sampling frequency Fs is set to at least 40000 Hz, and the sampling precision Ps is set to at least 16 bits, and sampling is performed according to the Nyquist theorem.

[0028] Step S40: audio signal processing and storage.

[0029] The collected analog audio signal is subjected to analog-digital conversion, the conversion time is T2, and digital audio data S_data=PsxFsXT2 is generated. The S_data and the sampling parameters are packaged into a WAV format file for local storage.

[0030] Step S50: interactive implementation.

[0031] The stored WAV file is sent to an audio recognition module to convert it into text information. The text information is transmitted to a voice interaction model to generate a reply text. The reply text is converted into an audio signal through a voice synthesis technology, and played through a loudspeaker, completing the interaction.

[0032] The technical solution: taking a chassis robot for exhibition tour as an example, a radio module driven by a stepping motor and capable of 360° horizontal rotation is installed on the chassis, the pickup surface of the radio is arc-shaped, the central angle θ is 90°, the radius r is 40mm, and the height h is 30mm.

[0033] The implementation process of the control method is as follows: 1. Sound source detection and angle recording: a. After the robot starts, the radio begins to rotate counterclockwise at an angular velocity of 90° / s.

[0034] b. During the scanning process, the system listens in real time at a sampling rate of 44.1 kHz. When it rotates to the 45° direction, it detects a sound source A with a loudness of 70dB; when it rotates to the 300° direction, it detects a sound source B with a loudness of 65dB.

[0035] c. The system compares the loudness of the two and selects sound source A as the target sound source, recording its angle r1=45°.

[0036] 2. Optimal path rotation and error correction: a. Assuming that the initial angle of the radio is 0°, the shortest rotation angle Q is calculated as min(45, 360-45)=45°. The direction is counterclockwise.

[0037] b. The preset angular velocity w=90° / s, and the theoretical rotation time Ttheoretical=45 / 90=0.5s.

[0038] c. Control the motor to start rotating. Due to the slight fluctuations of the mechanical load, when the actual rotation time T_rotation = 0.48 seconds, the controller considers it to be in place.

[0039] d. The system performs error correction: Update the current angle R = 0.48 × 90 = 43.2°. The angle deviation is |45 - 43.2| = 1.8°. Since T_rotation (0.48 s) < T_theory (0.5 s), the system controls the radio receiver to rotate another 1.8° in the original direction (counterclockwise). Finally, the radio receiver stabilizes at the 45° direction and completes the alignment.

[0040] 3. High - fidelity audio acquisition and processing: a. After alignment, the system picks up the sound of sound source A continuously for 5 seconds with parameters Fs = 44.1 kHz and Ps = 16 bit.

[0041] b. After analog - to - digital conversion, the data volume S_data = 16 × 44100 × 5 = 3,528,000 bits (about 441 KB).

[0042] c. Write this data together with the parameters into a WAV file and store it locally on the robot.

[0043] 4. Interaction implementation: a. This WAV file is sent to the cloud voice recognition service and converted into the text "Excuse me, where is the space exhibition area?"

[0044] b. The text is input into the large - language model and the reply text "Please go straight for 50 meters and then turn left at the space exhibition area." is obtained. c. After the reply text is synthesized into speech by the TTS engine, it is played through the robot's speaker to guide the audience.

[0045] Thus, a complete human - machine interaction loop is completed. After this interaction ends, the radio receiver stops working for 3 seconds (silent state, t1 = 3 s), and then starts 360° scanning again, waiting for the next interaction.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the protection scope of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the essence and scope of the technical solutions of the present invention.

Claims

1. A control method for a multi-directional microphone based on a chassis-type robot, characterized in that, It includes the following steps: Step S10: Control the radio to perform rotational scanning. When a sound source is detected, record its angle, and select a target sound source from multiple sound sources and record its angle r1; Step S20: Calculate the optimal rotation path and angle based on the angle r1, control the radio to rotate, and correct the rotation error; Step S30: Control the radio facing the sound source to collect audio at a preset sampling frequency Fs and sampling precision Ps; Step S40: Perform analog-to-digital conversion on the collected audio signal and store it as a digital audio file; Step S50: Convert the digital audio file into text and perform interactive processing, and finally reply through voice playback.

2. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S10, the specific method of selecting a target sound source from multiple sound sources is: Select the sound source with the largest loudness as the target sound source.

3. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S20, the specific method of calculating the optimal rotation path and angle is: Assume the initial angle of the radio is 0, then the shortest rotation angle Q = min(r1, 360 - r1).

4. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S20, control the radio to rotate at a constant angular velocity w, and the theoretical rotation time T_theory = Q / w.

5. The control method for a multi-directional microphone based on a chassis-type robot according to claim 4, characterized in that, In the step S20, the specific method of correcting the rotation error is: Record the actual rotation time T_rotation. When T_rotation ≠ T_theory, update the current angle according to the formula R = T_rotation × w, and calculate the deviation |Q - R| from the target angle Q; If T_rotation < T_theory, control the radio to continue rotating |Q - R| degrees in the original direction; If T_rotation > T_theory, control the radio to rotate |Q - R| degrees in the reverse direction.

6. The control method for a multi-directional microphone based on a chassis-type robot according to claim 5, characterized in that, The error correction is a single correction.

7. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S30, the sampling frequency Fs is 40000Hz.

8. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S30, the sampling precision Ps is 16bit.

9. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S40, the size of the digital audio data after analog-to-digital conversion S_data = Ps × Fs × T2, where T2 is the conversion time.

10. The control method for a multi-directional microphone based on a chassis-type robot according to claim 1, characterized in that, In the step S10, when no sound source is detected, the radio enters the silent state and automatically restarts scanning after an interval time t1.

Images