An interactive TR robot and its interaction method

Abstract

This invention discloses an interactive TR robot and its interaction method, relating to the field of robot interaction technology. It includes a workbench and a voice acquisition device. A lower microphone is located at the top of the outer wall of the workbench, and the lower microphone has a wider bottom and narrower top structure. A voice acquisition device is located at the top of the outer wall of the lower microphone, and eight upward-tilted acquisition holes are located on the side of the outer wall of the voice acquisition device. An upper microphone is located at the top of the outer wall of the voice acquisition device, and the upper microphone has a wider top and narrower bottom structure. This invention, by incorporating the upper microphone, lower microphone, voice acquisition device, eight sets of tilted acquisition holes, and a directional microphone, achieves converged sound wave acquisition and directional noise reduction from eight groups, suppressing noise from non-target directions, significantly improving the strength and accuracy of voice signal acquisition, and clearly capturing operator voice even in noisy environments.

Description

Technical Field

[0001] This invention relates to the field of robot interaction technology, specifically to a TR robot that facilitates interaction and its interaction method. Background Technology

[0002] With the rapid development of industrial automation and intelligent manufacturing, TR robots are increasingly widely used in material handling, assembly and precision operation. The convenience and accuracy of human-robot interaction have become core indicators for measuring the efficiency and practicality of robot operations. Traditional TR robots' voice acquisition components mostly use a single microphone or two sets of symmetrical microphones, which have limited sound pickup range and lack acoustic convergence structure. When the operator is in a noisy environment, the microphone is prone to weak signal acquisition and incomplete filtering of environmental noise, resulting in missed or misrecognized voice commands. Operators need to repeat the commands repeatedly, which seriously reduces work efficiency.

[0003] Patent CN107278302B discloses a robot interaction method and an interactive robot. The above patent realizes the provision of the most relevant response to the question with the fewest voice question-and-answer sessions.

[0004] The aforementioned patents proactively interact with users through various interactive methods to collect various information and user preferences, continuously improving the user feature information set to support subsequent user questions and requests, and providing the most relevant response to the question with the fewest voice Q&A sessions. However, there is still room for optimization in voice acquisition.

[0005] Therefore, this application proposes a robot interaction method with strong voice signal acquisition and an interactive robot. Summary of the Invention

[0006] The purpose of this invention is to provide an interactive TR robot and its interaction method to solve the technical problems of weak signal acquisition and incomplete environmental noise filtering mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an interactive TR robot, comprising a workbench, a voice acquisition device, and a data processing module. A lower microphone is located at the top of the outer wall of the workbench, and a rubber base is located at the bottom of the outer wall of the lower microphone. The lower microphone has a wider bottom and narrower top structure for receiving sound. A voice acquisition device is located at the top of the outer wall of the lower microphone, and a acquisition hole is located on the side of the outer wall of the voice acquisition device. A data processing module is located inside the voice acquisition device, and the data processing module contains a weighted fusion algorithm. An access terminal, an output terminal, and eight signal input terminals are located on the outside of the data processing module. The output terminal of the data processing module is connected to a connecting cable, which passes through the lower microphone and connects to a controller. A mechanical rod is located at the top of the outer wall of the controller, and an operation screen is located at the top of the outer wall of the mechanical rod. The controller is connected to the TR robot via a conduit cable. An upper microphone is located at the top of the outer wall of the voice acquisition device, and the upper microphone has a wider top and narrower bottom structure for receiving sound.

[0008] Preferably, a directional microphone is installed inside the acquisition hole. The receiving end of the directional microphone is arranged coaxially with the acquisition hole and the distance between them is 5mm. The output end of the directional microphone is connected to the data processing module through a data cable. The output ends of the eight directional microphones are respectively connected to the eight independent signal input ends of the data processing module. The data processing module is connected to the controller through a connecting cable. The controller is connected to the operation screen through the data cable inside the mechanical rod.

[0009] Preferably, the top of the outer wall of the workbench is provided with a circular groove, the bottom of the outer wall of the workbench is provided with four support rods, the bottom of the inner wall of the circular groove is provided with a sound-insulating base, the side of the inner wall of the circular groove and the side of the outer wall of the sound-insulating base are fitted together, the top of the outer wall of the sound-insulating base is provided with a sound-insulating ring, the side of the outer wall of the sound-insulating ring and the side of the inner wall of the circular groove are fitted together, the bottom of the outer wall of the sound-insulating ring and the top of the outer wall of the sound-insulating base are in contact, the side of the inner wall of the sound-insulating ring and the side of the outer wall of the rubber seat are fitted together, and the bottom of the outer wall of the rubber seat and the top of the outer wall of the sound-insulating base are in contact.

[0010] Preferably, the sound-insulating base and the sound-insulating ring are made of closed-cell rubber and each has an independent sealed air bubble inside. The closed-cell structure of the sound-insulating base and the sound-insulating ring blocks the mechanical vibration noise of the workbench itself, and the densely packed independent sealed air bubbles inside absorb the remaining mechanical sound wave energy of the workbench through the vibration of the bubble walls.

[0011] Preferably, a noise reduction mesh is provided on the outer side of the upper and lower radio stations. Four second screw holes are provided at the top of the outer wall of the noise reduction mesh. The second screw holes are concentrically aligned with the first screw holes. The first screw holes are located at the top of the outer wall of the fixing plate. The first screw holes pass through the top of the outer wall of the fixing plate and emerge from the bottom of the outer wall of the fixing plate. Bolts pass through the first and second screw holes to fix the fixing plate to the top of the outer wall of the noise reduction mesh. A connecting rod is provided at the bottom of the outer wall of the fixing plate.

[0012] Preferably, the noise reduction mesh is cylindrical with a hollow interior. Noise reduction holes are provided on the outer side of the noise reduction mesh. The inner wall of the noise reduction holes is made of fiber composite material, which consists of an outer layer and an inner layer. The outer layer is a dense fiber layer and the inner layer is a loose fiber layer. The noise reduction holes have a regular hexagonal structure. A total of twenty sets of noise reduction holes are provided on the outer side of the noise reduction mesh, with six noise reduction holes in each set. The noise reduction holes in each set are evenly arranged from top to bottom on the outer side of the noise reduction mesh.

[0013] Preferably, the outer side of the connecting rod and the inner side of the connecting groove are fitted together. The connecting groove is located at the top of the outer wall of the upper radio console. There are a total of four connecting grooves at the top of the outer wall of the upper radio console, and the four connecting grooves are fitted together with the four connecting rods respectively.

[0014] Preferably, the bottom end of the outer wall of the noise reduction mesh is provided with a cable inlet groove, which is a rectangular structure. The connecting wire passes through the cable inlet groove into the lower radio station. The bottom end of the outer wall of the noise reduction mesh is in contact with the top end of the outer wall of the sound insulation ring, and the side of the outer wall of the noise reduction mesh and the side of the inner wall of the circular groove are interlocked.

[0015] Preferably, the interaction method includes the following steps:

[0016] After the S1 and TR robots are started, the sound-insulating base made of closed-cell rubber and the sound-insulating ring in the circular groove at the top of the workbench form a double sound barrier through the interlocking structure. The sound-insulating base uses its closed-cell structure to block the transmission of the mechanical vibration of the workbench itself to the voice collector. The densely packed independent sealed bubbles inside absorb the remaining mechanical sound wave energy through the vibration of the bubble walls. The noise reduction mesh set on the outside of the upper and lower microphones is fixed to the connecting groove of the upper microphone through the connecting rod. The cylindrical hollow structure of the noise reduction mesh, together with the twenty sets of regular hexagonal noise reduction holes on the outer wall side, forms an acoustic filter layer. The dense fiber layer on the surface reflects human voice, and the loose fiber layer on the inner surface absorbs low-frequency noise.

[0017] S2. When the operator issues a voice command, the sound wave first enters through the noise reduction hole of the noise reduction mesh, and undergoes preliminary noise reduction and filtering through the double-layer structure of the fiber composite material. Then, the acoustic convergence channel formed by the lower and upper receivers reflects and converges the sound wave, causing the sound wave to concentrate on the eight sets of upward-tilted acquisition holes on the side of the voice acquisition device. The directional microphones in the acquisition holes receive the converged voice signal and convert the sound wave vibration into an analog electrical signal.

[0018] Preferably, the interaction method further includes the following steps:

[0019] The eight sets of analog electrical signals output by the S3 directional microphone are respectively connected to the independent signal input terminals of the data processing module. The data processing module uses a weighted fusion algorithm to assign weights according to the signal-to-noise ratio (SNR) of each signal, and generates a fused signal by weighted superposition of the signals. This weakens the interference of low-quality signals. At the same time, the soundproof base and soundproof ring continuously block the noise generated by the mechanical vibration of the workbench, preventing it from interfering with the signal during the data processing process and ensuring the stability of the fused signal.

[0020] S4. The data processing module converts the fused signal into a digital signal via analog-to-digital conversion and then transmits it to the controller via a connection line. The controller compares and identifies the instruction with the preset semantic library and displays it on the operation screen. After the operator confirms the instruction, the controller transmits the instruction to the TR robot for execution via a summary line.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. This invention, by setting up an upper and lower receiver, a voice acquisition unit, eight sets of tilted acquisition holes and directional microphones, achieves eight sets of sound wave convergence and directional noise reduction. The upper and lower receivers form an acoustic channel, reflecting and converging sound waves to the acquisition holes. The eight omnidirectional acquisition holes, combined with a 60-degree acquisition angle, can suppress noise from non-target directions, greatly improving the strength and accuracy of voice signal acquisition. It can clearly capture the operator's voice even in noisy environments, solving the problems of limited voice acquisition range and susceptibility to environmental noise interference leading to interaction failure in traditional TR robots.

[0023] 2. This invention, through the design of a data processing module and a weighted fusion algorithm, achieves dynamic weighted fusion of eight sets of analog electrical signals. The data processing module allocates weights according to SNR levels, and generates a fused signal by weighted superposition of the eight sets of microphone signals. While filtering high-quality signals, it weakens the interference of low-quality signals, and outputs a stable fused signal with a high signal-to-noise ratio, providing reliable input for subsequent digital processing and semantic recognition. This solves the problem of signal distortion and weak anti-interference ability caused by simple superposition in traditional multi-microphone signal processing, which in turn affects the accuracy of semantic recognition.

[0024] 3. This invention achieves dual physical isolation and energy absorption of the mechanical vibration noise of the workbench itself by setting up a circular groove, a sound-insulating base, and a sound-insulating ring. The circular groove provides a fitting space for the sound-insulating base and the sound-insulating ring, so that the sound-insulating base and the sound-insulating ring form an all-round wrapping structure. The closed-cell structure of the closed-cell rubber of the sound-insulating base and the sound-insulating ring directly cuts off the vibration transmission path. The independent sealed air bubbles inside the sound-insulating base and the sound-insulating ring convert the remaining mechanical vibration energy into heat energy through elastic deformation, which significantly reduces the interference of the mechanical vibration of the workbench on the voice acquisition device, greatly improves the signal-to-noise ratio of the voice signal, and solves the problem that the voice signal is easily mixed in by the mechanical vibration of the workbench in traditional TR robots, resulting in the voice command being drowned out or distorted by mechanical noise.

[0025] 4. This invention, by incorporating a noise-reducing mesh and noise-reducing holes, achieves directional filtering of low-frequency environmental noise and efficient convergence and transmission of human voice signals. The cylindrical structure of the noise-reducing mesh and twenty sets of regular hexagonal noise-reducing holes form a 360-degree acoustic filtering space. The dense fiber layer on the inner wall of the noise-reducing holes reflects human voice and guides it to the acquisition hole, while the loose fiber layer absorbs low-frequency environmental noise. This ensures that the high transmission efficiency of human voice signals can be maintained even in noisy environments, improving the absorption efficiency of low-frequency environmental noise and ensuring that the voice acquisition device prioritizes the capture of valid commands. This solves the problem that traditional TR robots have difficulty distinguishing between human voices, mechanical buzzing, and electronic noise in complex environments, leading to missed or misrecognized voice commands. Attached Figure Description

[0026] Figure 1 This is a three-dimensional structural diagram of the present invention;

[0027] Figure 2 This is a schematic diagram of the workbench structure of the present invention;

[0028] Figure 3 This is a schematic diagram of the soundproof base and soundproof ring pull-out workbench structure of the present invention;

[0029] Figure 4 This is a schematic diagram of the noise reduction network-pulled speech acquisition device structure of the present invention;

[0030] Figure 5 This is a schematic diagram of the noise reduction mesh structure pulled out by the connecting rod of the present invention;

[0031] Figure 6 This is a schematic diagram of the voice acquisition device structure of the present invention;

[0032] Figure 7 This is a schematic diagram of the bolt pull-out fixing plate structure of the present invention;

[0033] Figure 8 This is a partial structural diagram of the connection rod and the connection groove of the present invention.

[0034] In the diagram: 1. Workbench; 2. Lower microphone; 3. Voice acquisition device; 4. Acquisition hole; 5. Upper microphone; 6. Rubber base; 7. Connecting cable; 8. Controller; 9. Mechanical rod; 10. Operation panel; 11. Support rod; 12. Conduit cable; 13. Circular groove; 14. Soundproof base; 15. Soundproofing ring; 16. Connecting groove; 17. Connecting rod; 18. Fixing plate; 19. First screw hole; 20. Bolt; 21. Noise reduction mesh; 22. Second screw hole; 23. Noise reduction hole; 24. Cable inlet groove; 25. TR robot. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," "outer," "front end," "rear end," "both ends," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0037] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0038] Please see Figure 4 , Figure 5 , Figure 7 and Figure 8 An embodiment of the present invention provides: a TR robot that is easy to interact with, wherein the top of the outer wall of the noise reduction net 21 is provided with four second screw holes 22, the second screw holes 22 and the first screw holes 19 are concentrically aligned, and the bolts 20 pass through the first screw holes 19 and the second screw holes 22 to fix the fixing plate 18 to the top of the outer wall of the noise reduction net 21. The bottom of the outer wall of the fixing plate 18 is provided with a connecting rod 17, and the connecting rod 17 and the connecting groove 16 are interlocked.

[0039] Furthermore, before starting the equipment, the operator first installs the four connecting rods 17 on the top of the outer wall of the noise reduction net 21, then lifts the noise reduction net 21 and places it directly above the upper radio station 5. The operator rotates the noise reduction net 21 to align the inlet slot 24 with the connecting line 7, and then aligns the four connecting rods 17 with the connecting slot 16 at the top of the outer wall of the upper radio station 5. Then, the operator slowly moves the noise reduction net 21 down along the outer wall of the upper radio station 5 until the four connecting rods 17 and the connecting slot 16 are completely in contact with the bottom. At this time, the inlet slot 24 is stuck directly above the connecting line 7, and the bottom of the outer wall of the noise reduction net 21 is in contact with the top of the outer wall of the sound insulation ring 15.

[0040] Please see Figure 1 , Figure 2 , Figure 3 and Figure 6 The present invention provides an embodiment of a TR robot that facilitates interaction. The bottom of the outer wall of the workbench 1 is provided with four support rods 11. The inner side of the circular groove 13 and the outer side of the soundproof base 14 are interlocked. The top of the outer wall of the soundproof base 14 is provided with a soundproof ring 15. The outer side of the soundproof ring 15 and the inner side of the circular groove 13 are interlocked. The inner side of the soundproof ring 15 and the outer side of the rubber seat 6 are interlocked. The bottom of the outer wall of the rubber seat 6 and the top of the outer wall of the soundproof base 14 are in contact. The soundproof base 14 and the soundproof ring 15 are made of closed-cell rubber and each has an independent sealed air bubble inside. The soundproof base 14 and the soundproof ring 15 block the mechanical vibration noise of the workbench 1 itself.

[0041] Furthermore, when the equipment starts up, the low-frequency mechanical noise generated by the contact between the workbench 1 and the ground is transmitted to the workbench 1 through the support rod 11, thereby spreading to the area of ​​the circular groove 13. At this time, the soundproof base 14 and the soundproof ring 15 work together to filter the low-frequency noise. The bottom end of the outer wall of the soundproof base 14 is in contact with the bottom end of the inner wall of the circular groove 13, and the top end of the outer wall of the soundproof base 14 is in contact with the bottom end of the outer wall of the rubber seat 6. The soundproof base 14 filters the mechanical noise at the bottom of the circular groove 13, preventing the noise from entering the voice collector 3 through the rubber seat 6. The side of the outer wall of the soundproof ring 15 is in contact with the side of the inner wall of the circular groove 13, and the side of the inner wall of the soundproof ring 15 and the side of the outer wall of the rubber seat 6 are in contact. The sound insulation ring 15 and the sound insulation base 14 are closely fitted together, preventing mechanical noise from entering the voice collector 3 from the side of the circular groove 13. The closed-cell structure of the sound insulation ring 15 and the sound insulation base 14 directly cuts off the continuous transmission path of the vibration of the workbench 1 to the voice collector 3 through the solid medium. When the remaining unblocked low-frequency vibration energy is transmitted to the sound insulation base 14 and the sound insulation ring 15, the bubble walls of the densely packed independent bubbles will undergo elastic deformation with the vibration. For low-frequency vibrations of 20 to 100 Hz, the bubble walls will produce a reciprocating motion of stretching and contraction, converting the mechanical energy of the vibration into frictional heat energy inside the bubble wall material, thereby filtering out the remaining low-frequency vibration noise.

[0042] Please see Figure 3 , Figure 4 , Figure 5 and Figure 6 The present invention provides an embodiment of a TR robot that facilitates interaction. The noise reduction net 21 is cylindrical and hollow inside. Noise reduction holes 23 are provided on the outer side of the noise reduction net 21. The inner wall of the noise reduction hole 23 is made of fiber composite material, which consists of an outer layer and an inner layer. The outer layer is a dense fiber layer and the inner layer is a loose fiber layer. The noise reduction hole 23 has a regular hexagonal structure. A total of twenty sets of noise reduction holes 23 are provided on the outer side of the noise reduction net 21. Each set of six noise reduction holes 23 are evenly arranged from top to bottom on the outer side of the noise reduction net 21.

[0043] Furthermore, after the equipment is started, the low-frequency noise to be filtered, such as mechanical vibration noise and electronic humming, is concentrated between 20 and 100 Hz, with a relatively long wavelength, typically 1.7 to 17 m. In contrast, the human voice frequency is mainly distributed between 100 and 3400 Hz, with a shorter wavelength, typically between 0.1 and 1 m. The long wavelength characteristic of low-frequency noise allows it to bypass the dense fiber barrier and enter the noise reduction hole 23 through diffraction. Diffraction refers to the phenomenon where waves deviate from a straight propagation path and go around behind an obstacle or pore when they encounter it during propagation. Its significance is directly related to the ratio of wavelength to obstacle size. The longer the wavelength and the smaller the obstacle size, the more obvious the diffraction phenomenon. The wavelength of low-frequency noise is 1.7 to 17 m, while the size of the fibers and pores in the dense fiber layer is much smaller than the wavelength of low-frequency noise. The sound wave can easily bypass the fiber edge and enter the noise reduction hole 23 through the surface layer in a curved propagation manner.

[0044] Human voice has a short wavelength and strong directionality, making it impossible to bypass dense fibers. The directionality of sound waves is closely related to their wavelength. The shorter the wavelength, the closer the wave propagates to a straight line. The stronger the directionality, the longer the wavelength, and the easier it is to spread in all directions during propagation. Human voice is a short wavelength with strong directionality. When the operator's voice encounters a dense fiber layer, due to the strong directionality of short waves, it propagates in a straight line and impacts the fiber layer. At this time, the dense structure of the fiber layer forms a physical barrier to the human voice sound wave. Moreover, the straight-line propagation characteristic of short wavelengths makes it difficult for short wavelengths to bypass the fibers and enter the noise reduction hole 23. The diffraction phenomenon is extremely weak, and more human voice sound waves are reflected by the fiber surface. The reflected human voice sound waves are guided inward and converge, eventually concentrating and transmitting towards the acquisition hole 4 of the voice acquisition device 3, thus improving the directional transmission efficiency of the human voice signal.

[0045] Please see Figure 1 , Figure 4 , Figure 7 and Figure 8The present invention provides an embodiment of an interactive TR robot. A lower microphone 2 is located at the top of the outer wall of the workbench 1. The lower microphone 2 has a wider bottom and narrower top sound-receiving structure. A voice collector 3 is located at the top of the outer wall of the lower microphone 2. Eight vertically arranged collection holes 4 are located on the side of the outer wall of the voice collector 3. A data processing module is located inside the voice collector 3. The data processing module has an input end, an output end, and eight signal input ends. A connecting cable 7 passes through the lower microphone 2 and connects to a controller 8. An upper microphone 5 is located at the top of the outer wall of the voice collector 3. The upper microphone 5 has a wider top and narrower bottom sound-receiving structure. A directional microphone is located inside the collection holes 4. The receiving end of the directional microphone is coaxially arranged with the collection hole 4 and the distance between them is 5mm. The output end of the directional microphone is connected to the data processing module via a data cable. The output ends of the eight directional microphones are respectively connected to the eight independent signal input ends of the data processing module.

[0046] Furthermore, when the operator interacts with the TR robot 25 via voice, the voice, which has been initially filtered by the noise reduction net 21, is transmitted to the lower receiver 2 and the upper receiver 5. The lower receiver 2, which is wider at the bottom and narrower at the top, and the upper receiver 5, which is wider at the top and narrower at the bottom, form an acoustic converging channel. The sound waves are reflected by the inclined outer walls of the lower receiver 2 and the upper receiver 5 and converge toward the direction of the acquisition hole 4, thus enhancing the incident sound waves. The eight sets of vertically arranged acquisition holes 4 cover a voice acquisition area ranging from 0.5 to 2 meters in height. Each set of acquisition holes 4 has three holes, and the sound acquisition angle of the eight sets of acquisition holes 4 is 60 degrees, which can effectively suppress the remaining environmental noise in non-target directions.

[0047] During voice signal acquisition, the directional microphone converts the received sound wave vibrations into analog electrical signals. The directional microphone uses a moving coil for conversion. Inside the directional microphone, there is a constant magnetic field formed by a permanent magnet. A coil rigidly connected to the diaphragm is placed in the magnetic field. The conversion principle is that when sound wave vibrations are transmitted, the diaphragm is driven by changes in air pressure to generate mechanical vibrations, which drive the coil to move synchronously in the magnetic field, cutting magnetic field lines. According to the law of electromagnetic induction, an alternating electromotive force is induced in the coil, which is consistent with the vibration law of the diaphragm. The law of electromagnetic induction states that when a conductor in a closed circuit moves in a magnetic field and cuts magnetic field lines, or when the magnetic flux through the closed circuit changes, an induced electromotive force is generated in the circuit. The alternating electromotive force is the analog electrical signal. The amplitude of the analog electrical signal corresponds to the sound wave intensity, and the frequency corresponds to the sound wave tone. Then, the directional microphone transmits the analog electrical signal to the data processing module.

[0048] The data processing module then uses a weighted fusion algorithm to obtain a fused signal from the eight sets of analog electrical signals. The weighted fusion algorithm divides the eight signals into three levels based on their real-time signal-to-noise ratio (SNR): signal groups with an SNR greater than or equal to 30 dB are assigned a weight of 0.15; signal groups with an SNR less than 30 dB but greater than or equal to 20 dB are assigned a weight of 0.12; signal groups with an SNR less than 20 dB but greater than or equal to 10 dB are assigned a weight of 0.08; and signal groups with an SNR less than 10 dB are assigned a weight of 0.05. For example, at a certain moment... The signal-to-noise ratios of the eight signal groups are 42dB, 38dB, 30dB, 25dB, 22dB, 18dB, 8dB, and 6dB, respectively, with corresponding weights of 0.15, 0.15, 0.15, 0.12, 0.12, 0.08, 0.05, and 0.05. At this time, the instantaneous amplitudes of the eight signal groups are 0.9V, 0.8V, 0.7V, 0.6V, 0.5V, 0.4V, 0.2V, and 0.1V, respectively. The module multiplies each amplitude by its corresponding weight and then sums them to obtain the fused signal amplitude of 0.539V.

[0049] Then, the analog-to-digital converter inside the data processing module converts the fused analog signal into a digital signal. The analog-to-digital converter realizes the conversion of analog electrical signals into digital signals through three steps: sampling, quantization, and encoding. First, sampling is performed, and the instantaneous voltage value of the analog signal is intercepted at fixed intervals higher than twice the highest frequency of the signal, discretizing the continuous time domain signal into time-discontinuous sampled values. Next, quantization is performed, and the continuous voltage range obtained by sampling, such as 0 to 3.3V, is divided into discrete levels, with each sampled voltage value corresponding to the closest level. Finally, encoding is performed, and the quantized levels are represented by binary numbers, such as 16-bit levels corresponding to 16-bit binary codes, outputting a digital signal that can be recognized by the processor. Finally, the data processing module transmits the digital signal to the controller 8 through connection line 7.

[0050] Please see Figure 1 , Figure 4 and Figure 5 The present invention provides an embodiment of a TR robot that is easy to interact with. The output end of the data processing module is connected to the connecting line 7. The connecting line 7 passes through the lower radio station 2 and is connected to the controller 8. The top of the outer wall of the controller 8 is provided with a mechanical rod 9. The top of the outer wall of the mechanical rod 9 is provided with an operation screen 10. The controller 8 is connected to the TR robot 25 through the conduit 12.

[0051] Furthermore, the controller 8 compares the input digital signal with its internally preset semantic recognition library. The semantic recognition library contains voice commands such as "start robotic arm 0001", "adjust gripping angle 0010", and "stop running 1000". When the input digital signal is 0010, the controller 8 recognizes the command as "adjust gripping angle" and transmits the recognized command to the operation screen 10. After the operator confirms that it is the voice command, the controller 8 converts the recognized voice command into mechanical motion execution. The conversion process involves the controller 8 calling the built-in command parameter mapping table. The voice command parameters corresponding to the mapping table are: the first servo motor speed is 30 r / min and the rotation angle is 30 degrees; the second servo motor speed is 30 r / min and the rotation angle is 45 degrees; the third servo motor speed is 30 r / min and the rotation angle is 45 degrees. The first servo motor controls the left and right rotation angles, the second servo motor controls the rotation angle of the rear robotic arm, and the third servo motor controls the rotation angle of the front robotic arm. Then, the controller 8 transmits the corresponding voice command parameters to the TR robot 25 through the aggregation cable 12.

[0052] Working principle: After the equipment is started, the operator issues a voice command. The sound insulation base 14 and the sound insulation ring 15 block the mechanical vibration of the workbench 1 through the closed-cell rubber structure. The independent air bubbles inside the sound insulation base 14 and the sound insulation ring 15 absorb the low-frequency mechanical noise energy, forming a double sound insulation barrier. At the same time, the noise reduction net 21 is fixed to the upper receiver 5 through the connecting rod 17. The dense fiber on the surface of the regular hexagonal noise reduction hole 23 reflects human voice, and the loose fiber on the inner layer absorbs low-frequency noise, thus completing the acoustic noise reduction pretreatment.

[0053] Then, after being filtered by the noise reduction network 21, the sound waves are reflected and converged through the acoustic channel formed by the lower receiver 2 and the upper receiver 5, and are directionally transmitted into the eight sets of acquisition holes 4 of the voice acquisition device 3. The directional microphones in the acquisition holes 4 convert the sound wave vibrations into analog electrical signals, which are transmitted to the data processing module through the data line. The data processing module uses a weighted fusion algorithm to assign weights to the signals according to the signal-to-noise ratio (SNR), and generates a fused signal by superposition. Then, it is converted from analog to digital signal and transmitted to the controller 8. The controller 8 compares the recognition command with the preset semantic library and displays it on the operation screen 10.

[0054] After the operator confirms, the controller 8 calls the instruction parameter mapping table and transmits the corresponding voice instruction parameters to the TR robot 25 through the aggregation line 12, driving the mechanical structure to perform the corresponding operation and achieving efficient human-machine interaction.

[0055] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An interactive TR robot, comprising a workbench (1), a voice acquisition unit (3), and a data processing module, characterized in that: The workbench (1) has a lower microphone (2) at the top of its outer wall and a rubber seat (6) at the bottom of its outer wall. The lower microphone (2) has a microphone structure that is wider at the bottom and narrower at the top. The lower microphone (2) has a voice acquisition device (3) at the top of its outer wall and a acquisition hole (4) on the side of its outer wall. The voice acquisition device (3) has a data processing module inside and a weighted fusion algorithm inside. The data processing module has an access end, an output end, and eight sets of input and output terminals on its outer side. The signal input end and the output end of the data processing module are connected to the connecting line (7). The connecting line (7) passes through the lower radio station (2) and is connected to the controller (8). The top of the outer wall of the controller (8) is provided with a mechanical rod (9). The top of the outer wall of the mechanical rod (9) is provided with an operation screen (10). The controller (8) is connected to the TR robot (25) through the summing cable (12). The top of the outer wall of the voice collector (3) is provided with an upper radio station (5). The upper radio station (5) is a radio receiving structure that is wider at the top and narrower at the bottom.

2. The TR robot for easy interaction according to claim 1, characterized in that: The acquisition hole (4) is equipped with a directional microphone. The receiving end of the directional microphone is coaxially arranged with the acquisition hole (4) and the distance between them is 5mm. The output end of the directional microphone is connected to the data processing module through a data cable. The output ends of the eight directional microphones are respectively connected to the eight independent signal input ends of the data processing module. The data processing module is connected to the controller (8) through the connecting cable (7). The controller (8) is connected to the operation screen (10) through the data cable inside the mechanical rod (9).

3. The TR robot for easy interaction according to claim 1, characterized in that: The workbench (1) has a circular groove (13) at the top of its outer wall and four support rods (11) at the bottom of its outer wall. The inner wall of the circular groove (13) has a soundproof base (14) at the bottom. The inner wall of the circular groove (13) and the outer wall of the soundproof base (14) are fitted together. The outer wall of the soundproof base (14) has a soundproof ring (15) at the top. The outer wall of the soundproof ring (15) and the inner wall of the circular groove (13) are fitted together. The bottom of the outer wall of the soundproof ring (15) and the top of the outer wall of the soundproof base (14) are in contact. The inner wall of the soundproof ring (15) and the outer wall of the rubber seat (6) are fitted together. The bottom of the outer wall of the rubber seat (6) and the top of the outer wall of the soundproof base (14) are in contact.

4. The TR robot for easy interaction according to claim 3, characterized in that: The sound insulation base (14) and the sound insulation ring (15) are made of closed-cell rubber and each has an independent sealed air bubble inside. The closed-cell structure of the sound insulation base (14) and the sound insulation ring (15) blocks the mechanical vibration noise of the workbench (1) itself. The densely packed independent sealed air bubbles inside absorb the remaining mechanical sound wave energy of the workbench (1) through the vibration of the bubble wall.

5. The TR robot for easy interaction according to claim 1, characterized in that: Noise reduction mesh (21) is provided on the outer side of the upper radio station (5) and the lower radio station (2). Four second screw holes (22) are provided on the top of the outer wall of the noise reduction mesh (21). The second screw holes (22) and the first screw holes (19) are concentrically aligned. The first screw hole (19) is provided on the top of the outer wall of the fixing plate (18). The first screw hole (19) passes through the top of the outer wall of the fixing plate (18) and comes out from the bottom of the outer wall of the fixing plate (18). The bolt (20) passes through the first screw hole (19) and the second screw hole (22) to fix the fixing plate (18) to the top of the outer wall of the noise reduction mesh (21). A connecting rod (17) is provided on the bottom of the outer wall of the fixing plate (18).

6. The TR robot for easy interaction according to claim 5, characterized in that: The noise reduction mesh (21) is cylindrical and hollow inside. Noise reduction holes (23) are provided on the outer side of the noise reduction mesh (21). The inner wall of the noise reduction hole (23) is made of fiber composite material. The fiber composite material is composed of an outer layer and an inner layer. The outer layer is a dense fiber layer and the inner layer is a loose fiber layer. The noise reduction hole (23) is a regular hexagonal structure. A total of twenty groups of noise reduction holes (23) are provided on the outer side of the noise reduction mesh (21). Each group has six noise reduction holes (23). Each group of noise reduction holes (23) is evenly arranged from top to bottom on the outer side of the noise reduction mesh (21).

7. The TR robot for easy interaction according to claim 5, characterized in that: The outer side of the connecting rod (17) and the inner side of the connecting groove (16) are fitted together. The connecting groove (16) is set at the top of the outer wall of the upper radio station (5). There are four connecting grooves (16) at the top of the outer wall of the upper radio station (5). The four connecting grooves (16) are fitted together with the four connecting rods (17) respectively.

8. The TR robot for easy interaction according to claim 5, characterized in that: The bottom of the outer wall of the noise reduction mesh (21) is provided with a wire inlet groove (24). The wire inlet groove (24) is a rectangular structure. The connecting wire (7) passes through the wire inlet groove (24) into the lower radio station (2). The bottom of the outer wall of the noise reduction mesh (21) is in contact with the top of the outer wall of the sound insulation ring (15). The side of the outer wall of the noise reduction mesh (21) and the side of the inner wall of the circular groove (13) are interlocked.

9. An interaction method for an interactive TR robot, applicable to the interactive TR robot described in any one of claims 1-8, characterized in that: The interaction method includes the following steps: After the S1 and TR robot (25) are started, the sound insulation base (14) made of closed-cell rubber and the sound insulation ring (15) in the circular groove (13) at the top of the workbench (1) form a double sound insulation barrier through the interlocking structure. The sound insulation base (14) uses its closed-cell structure to block the transmission of the mechanical vibration of the workbench (1) to the voice collector (3). The densely packed independent closed bubbles inside absorb the remaining mechanical sound wave energy through the vibration of the bubble wall. The noise reduction net (21) set on the outside of the upper receiver (5) and the lower receiver (2) is interlocked and fixed with the connecting groove (16) of the upper receiver (5) through the connecting rod (17). The cylindrical hollow structure of the noise reduction net (21) and the twenty sets of regular hexagonal noise reduction holes (23) on the outer side of the outer wall form an acoustic filter layer. The dense fiber layer on the surface reflects human voice, and the loose fiber layer on the inner layer absorbs low-frequency noise. S2. When the operator issues a voice command, the sound wave first enters through the noise reduction hole (23) of the noise reduction net (21), and undergoes preliminary noise reduction filtering through the double-layer structure of the fiber composite material. Then, the acoustic convergence channel formed by the lower receiver (2) and the upper receiver (5) reflects and converges the sound wave, causing the sound wave to concentrate on the eight sets of upward tilting acquisition holes (4) on the side of the voice acquisition device (3). The directional microphone in the acquisition hole (4) receives the converged voice signal and converts the sound wave vibration into an analog electrical signal.

10. The interaction method for an interactive TR robot according to claim 9, characterized in that: The interaction method further includes the following steps: S3. The eight sets of analog electrical signals output by the directional microphone are respectively connected to the independent signal input terminal of the data processing module. The data processing module uses a weighted fusion algorithm to assign weights according to the signal-to-noise ratio (SNR) of each signal, and generates a fused signal by weighted superposition of the signals, thereby weakening the interference of low-quality signals. At the same time, the soundproof base (14) and the soundproof ring (15) continuously block the noise generated by the mechanical vibration of the workbench (1), so as to avoid it from interfering with the signal in the data processing process and ensuring the stability of the fused signal. S4. The data processing module converts the fused signal into a digital signal via analog-to-digital conversion and transmits it to the controller (8) via the connection line (7). The controller (8) compares and identifies the instruction with the preset semantic library and displays it on the operation screen (10). After the operator confirms the instruction, the controller (8) transmits the instruction to the TR robot (25) for execution via the summary line (12).

Images