An ultrasonic robot with autonomous scanning function and teaching learning method

Abstract

The present invention discloses an ultrasonic robot with autonomous scanning capabilities and a teaching learning method. The ultrasonic robot includes a single-chip microcontroller, a video capture card, a controller with a built-in teaching model, and an ultrasonic integrated device consisting of a flexible tactile sensor and an ultrasonic probe. During the teaching process, a posture monitoring device captures the ultrasonic probe's posture, a flexible tactile sensor collects contact data from the ultrasonic probe, and a video capture card captures ultrasonic images. A teaching dataset is constructed using the probe's posture, ultrasonic images, and contact data to train the teaching model. The ultrasonic integrated device is then mounted on a multi-degree-of-freedom mobile platform, which reads the ultrasonic images and contact data in real time. The teaching model is then used to predict and adjust the ultrasonic probe's posture to enable autonomous scanning by the ultrasonic robot. This invention simplifies the complex sensing and decision-making systems of the autonomous ultrasonic robot, making it easy to teach, deploy, and use, and has great potential for clinical application.

Description

Technical Field

[0001] The present invention belongs to the technical field of robots and relates to a medical ultrasonic robot, in particular to an ultrasonic robot with an autonomous scanning function and a teaching and learning method. Background Art

[0002] Ultrasound examinations are widely used in medicine due to their low cost, radiation-free nature, and ease of use. To promote the development of ultrasound medicine and address the shortage of ultrasound physicians, many researchers have recently proposed autonomous robotic ultrasound examinations. These robotic technologies, which utilize automated ultrasound scans of specific or multiple body parts, could significantly improve the efficiency of ultrasound examinations. However, current autonomous ultrasound robots are still far from clinically viable due to two key issues: First, autonomous ultrasound robots have limited ability to learn the scanning techniques of physicians and are unable to effectively simulate their scanning processes. Second, most autonomous ultrasound robot systems are complex, with highly coupled components, making deployment and maintenance difficult and unfavorable for clinical application. Summary of the Invention

[0003] In view of the above-mentioned deficiencies in the existing research on autonomous scanning of ultrasonic robots, the purpose of the present invention is to provide an ultrasonic robot with autonomous scanning capabilities and a teaching and learning method, which also has scanning teaching and scanning execution functions. The ultrasonic robot includes a multi-degree-of-freedom mobile platform and a detachable ultrasonic integrated device mounted thereon. The ultrasonic integrated device integrates a flexible tactile sensor and an ultrasonic probe, which can be held by a doctor to perform ultrasonic scanning teaching on the patient's designated area to be examined. After the teaching and learning is completed, the device is installed on a robot with six or more degrees of freedom at the end. The robot can then reproduce the doctor's ultrasonic scanning process on the patient and perform ultrasonic scanning on the patient. The ultrasonic robot and teaching and learning method described in the present invention are easy to teach, simple to deploy, and easy to use, and have great potential for clinical application.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] An ultrasonic robot with autonomous scanning function, characterized by comprising:

[0006] An ultrasonic probe is installed at the execution end of the multi-degree-of-freedom mobile platform and is used for performing ultrasonic scanning while being carried by the multi-degree-of-freedom mobile platform;

[0007] A flexible tactile sensor, fixed together with the ultrasound probe to form an ultrasound integrated device, is used to monitor the contact data between the ultrasound probe and the scanned object;

[0008] The controller has a built-in teaching model, which is used to record teaching information and train the teaching model during the manual teaching process. It also uses the trained teaching model to generate control instructions during scanning, and controls the multi-degree-of-freedom mobile platform carrying the ultrasonic integrated device for autonomous scanning according to the control instructions.

[0009] The teaching information includes ultrasonic images, contact data and posture information of the ultrasonic probe. The probe posture, ultrasonic images and contact data are used to construct a teaching data set for training the teaching model.

[0010] Preferably, the flexible tactile sensor and the ultrasonic probe are connected together via a fixing frame, the fixing frame comprising a first fixing plate and a second fixing plate, one end of the first fixing plate being configured as a connecting end connected to the multi-degree-of-freedom mobile platform, and the other end of the first fixing plate being detachably connected to the second fixing plate to form a clamping end for the ultrasonic probe;

[0011] There are two flexible tactile sensors, which are fixed on the first fixing plate and the second fixing plate on both sides of the ultrasonic probe respectively, and the detection surface of the flexible tactile sensor is configured to have the same shape and height as the scanning surface of the ultrasonic probe.

[0012] Preferably, the flexible tactile sensor comprises a sensor support plate, a flexible circuit board, an isolation net, a conductive sponge and a soft layer which are stacked in sequence;

[0013] The sensor support plate is used to provide overall support;

[0014] The flexible circuit board is provided with a plurality of electrode points distributed in an array and a voltage acquisition circuit for measuring the resistance signal between adjacent electrode points;

[0015] The isolation net is made of insulating material and has a large number of mesh holes, which are used to isolate the electrode points of the conductive sponge and the flexible circuit board under normal circumstances. When the conductive sponge is squeezed by an external force, the adjacent electrode points are connected through the mesh holes. The resistance between the adjacent electrode points is detected by the voltage acquisition circuit to obtain contact data, which includes the contact position and the contact force.

[0016] The soft layer is used for contacting with the scanned object and providing protection for the conductive sponge.

[0017] Preferably, the sensor support plate, flexible circuit board, isolation net, conductive sponge and soft layer are stacked in an arc-shaped package. After stacking, the outer surface of the soft layer forms a scanning surface with the same shape as the surface of the ultrasound probe.

[0018] Preferably, the electrode points include staggered excitation electrode points and sensing electrode points, and the voltage acquisition circuit includes a power supply, a voltage detector, and a plurality of reference resistors. All the excitation electrodes are connected in parallel to one electrode of the power supply, and each sensing electrode point is connected to another electrode of the power supply via a reference resistor.

[0019] The voltage detector detects the resistance between the excitation electrode point and the sensing electrode point by detecting the voltage score of each reference resistor.

[0020] Preferably, the voltage detector is a voltmeter.

[0021] The working principle of the flexible tactile sensor is as follows: when the tactile sensor is not in contact with an object, the conductive sponge and the flexible circuit board are separated by an isolation net. The excitation electrode and the sensing electrode on the flexible circuit board are disconnected, and the equivalent resistance is infinite. Therefore, the voltage measured by the reference resistance is approximately zero. When the soft layer is pressed, a depression is generated at the pressed part, and the conductive sponge inside is squeezed and contacts the flexible circuit board through the isolation net. The excitation electrode and the sensing electrode on the circuit board are connected. The greater the pressure, the more complete the contact between the conductive sponge and the electrode points and the lower the sponge's resistance, the smaller the equivalent resistance between the excitation electrode and the sensing electrode points, and thus the greater the voltage measured by the reference resistance. By measuring the voltage divider value of the reference resistance, the contact position and contact force of the tactile sensor can be calculated.

[0022] Preferably, the ultrasonic robot with autonomous scanning function further includes a video capture card, the controller includes a single-chip microcomputer and a computer with a built-in teaching model, and the video capture card is connected to the ultrasonic probe and the computer respectively, and is used to transmit the image information collected by the ultrasonic probe to the computer;

[0023] The single chip microcomputer is respectively connected to the flexible tactile sensor, the computer, and the multi-degree-of-freedom mobile platform, and is used to transmit the contact data detected by the flexible tactile sensor to the computer, and receive instructions from the computer to control the multi-degree-of-freedom mobile platform to carry the ultrasonic probe for scanning.

[0024] The present invention also protects a teaching method for an ultrasonic robot, which uses a posture monitoring device to capture the posture of an ultrasonic probe during the teaching process, and includes the following steps:

[0025] Install the posture monitoring device on the ultrasonic integrated device and set up the teaching scene;

[0026] The instructor uses a handheld ultrasonic integrated device to scan the test object for ultrasonic testing. During the scanning process, the flexible tactile sensor monitors the contact data of the ultrasonic probe, and the posture monitoring device detects the posture information of the ultrasonic probe.

[0027] Use the collected contact data and posture information to build a teaching model;

[0028] The posture monitoring device was removed, and the ultrasonic integrated device was installed on the execution end of the multi-degree-of-freedom mobile platform. The teaching model was used to drive the multi-degree-of-freedom mobile platform to carry the ultrasonic integrated device for autonomous scanning and ultrasonic testing.

[0029] Preferably, the posture monitoring device is a vicon motion capture system, which includes a positioning device and several visual cameras for visual capture. The positioning device is installed on the ultrasonic integrated device, and the shooting area of ​​the visual camera covers the active area of ​​the ultrasonic integrated device during the teaching process.

[0030] Preferably, the positioning device includes a cross positioning frame and a plurality of marking points installed on the cross positioning frame. During teaching, the cross positioning frame is installed on the top of the ultrasonic integrated device.

[0031] Preferably, the teaching model is a neural network model, including an autoencoder and a probe posture adjustment prediction neural network. The autoencoder is used to reduce the dimension of the image collected by the ultrasound probe, and the probe posture adjustment prediction neural network model is used to predict the change in the ultrasound probe posture at the next moment based on the ultrasound image data after dimension reduction, that is, the output of the neural network model is the control instruction.

[0032] Further preferably, the probe posture adjustment prediction neural network is a deep neural network.

[0033] The teaching process is as follows: a professional ultrasound doctor holds the ultrasound integrated device and performs an ultrasound examination on a kidney phantom, taking the examination of the kidney area as an example. The ultrasound image is recorded by a video acquisition card, the contact data of the flexible tactile sensor is recorded by a single-chip microcomputer, and the position and posture of the ultrasound integrated device is recorded by a Vicon motion capture system.

[0034] The learning process is as follows: the ultrasound images, contact data and posture data collected during the teaching process are made into a data set, which is input into a neural network for posture prediction to train and fit the relationship between the ultrasound images, contact data and the posture changes of the ultrasonic integrated device. The trained neural network model can be used for robot motion control during autonomous scanning by the ultrasonic robot.

[0035] The beneficial effects of the present invention are:

[0036] 1. An ultrasonic robot with autonomous scanning function is proposed. It integrates teaching learning and autonomous scanning functions, simplifies the complex sensing system and decision-making system of the autonomous ultrasonic robot, is easy to use, does not rely on a certain mobile robot platform, and has good portability.

[0037] 2. A flexible tactile sensor is designed for ultrasonic detection application scenarios. It adapts to the shape of the ultrasonic probe and can be used to collect data on contact with the human body during ultrasonic probe inspection.

[0038] 3. Based on the ultrasonic robot described in the present invention, a complete ultrasound teaching process and ultrasound examination skill learning method are proposed, which solves the problems of difficulty in doctor teaching and data collection during ultrasound scanning, and is applicable to most areas and organs of the human body. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Schematic diagram of the entire ultrasonic robot with autonomous scanning function in Example 1 of the present invention.

[0040] Figure 2 It is a schematic structural diagram of the ultrasonic integrated device of the ultrasonic robot in Example 1 of the present invention.

[0041] Figure 3 Schematic diagram of the flexible tactile sensor structure of the ultrasonic robot in Example 1 of the present invention.

[0042] Figure 4 Schematic diagram of the measurement principle of the flexible tactile sensor of the ultrasonic robot in Example 1 of the present invention.

[0043] Figure 5 This is a hardware wiring diagram of the ultrasonic integrated device of the ultrasonic robot of the present invention.

[0044] Figure 6 This is a teaching scene diagram of the ultrasonic robot teaching learning method of the present invention.

[0045] Figure 7 Schematic diagram of the teaching learning method and autonomous scanning process of the ultrasonic robot of the present invention.

[0046] Among them, 1-flexible tactile sensor, 2-fixing plate, 3-ultrasound probe, 4-posture monitoring device, 5-ultrasound integrated device, 6-silicone kidney phantom, 7-six-axis robotic arm, 11-sensor support plate, 111-mounting hole, 12-flexible circuit board, 13-isolation net, 14-conductive sponge, 15-soft layer, 121-excitation electrode point, 122-sensing electrode point, 123-reference resistor, 124-power supply, 125-voltmeter, 21-second fixing plate, 22-first fixing plate, 221-rectangular plate, 222-connecting hole, 41-cross positioning frame, 42-marker mark point, 43-visual camera, 16-controller, 161-single-chip microcomputer, 162-computer, 163-video acquisition card. DETAILED DESCRIPTION

[0047] In order to make the technical problems and technical solutions to be solved by the present invention clearer, the dual-arm puncture robot based on ultrasound image navigation provided by the present invention is described in detail below with reference to the accompanying drawings and embodiments, which are intended to explain rather than limit the present invention.

[0048] Example 1. Figure 1 As shown, an ultrasound robot with autonomous scanning function includes:

[0049] The ultrasonic probe 3 is mounted on the execution end of a multi-degree-of-freedom mobile platform (a six-axis robotic arm 7 is used in this embodiment) and is used to perform time-lapse scanning while being carried by the multi-degree-of-freedom mobile platform;

[0050] The flexible tactile sensor 1 is fixed together with the ultrasonic probe 3 to form an ultrasonic integrated device 5 for monitoring the contact data between the ultrasonic probe 3 and the scanned object;

[0051] The controller 16 has a built-in teaching model, which is used to record teaching information during the manual teaching process, use the teaching information to train the teaching model, and use the trained teaching model to generate control instructions during scanning. According to the control instructions, the multi-degree-of-freedom mobile platform carrying the ultrasonic integrated device 5 performs autonomous scanning; the teaching information includes ultrasonic images, contact data and posture information of the ultrasonic probe 3.

[0052] During the teaching process, firstly, the posture monitoring device 4 is installed on the ultrasonic integrated device 5 and the teaching scene is set up;

[0053] The instructor uses a handheld ultrasonic integrated device 5 to scan the test object for ultrasonic testing. During the scanning process, the flexible tactile sensor 1 monitors the contact data of the ultrasonic probe 3, and the posture monitoring device 4 detects the posture information of the ultrasonic probe 3.

[0054] Use the collected contact data and posture information to build a teaching model;

[0055] The posture monitoring device 4 is removed, and the ultrasonic integrated device 5 is installed on the execution end of the multi-degree-of-freedom mobile platform. The teaching model is used to drive the multi-degree-of-freedom mobile platform to carry the ultrasonic integrated device 5 for autonomous scanning and ultrasonic testing.

[0056] See also Figure 2The flexible tactile sensor 1 is a piezoresistive tactile sensor, consisting of a sensor support plate 11, a flexible circuit board 12, an isolation net 13, a conductive sponge 14, and a soft layer 15. The sensor support plate 11 has three mounting holes 111, which can be connected to the second fixing plate 21 or the first fixing plate 22 by screws. This allows the two flexible tactile sensors 1 to be fixed to either side of the ultrasound probe 3, respectively, to detect contact data between the probe and human skin without interfering with the imaging of the ultrasound probe 3. The second fixing plate 21 and the first fixing plate 22 fix the ultrasound probe 3 by clamping on both sides. The first fixing plate 22 is higher than the second fixing plate 21, and a rectangular plate 221 perpendicular to the first fixing plate 22 is provided on top of the first fixing plate 22 to form a connection end. The rectangular plate 221 is provided with several connection holes 222 for fixing to the actuator end of the multi-degree-of-freedom mobile platform via bolt connection.

[0057] See also Figures 2 to 4 The soft layer 15 at the bottom of the flexible tactile sensor 1 is made of flexible TPU, with a smooth and soft surface. The soft layer 15 has the same shape as the bottom of the ultrasound probe 3, designed to fit snugly onto the working surface of the bottom of the ultrasound probe 3. The soft layer 15 has edges around it and is directly fixed to the sensor support plate 11. An installation space is formed between the soft layer 15 and the sensor support plate 11, within which a conductive sponge 14, an isolation mesh 13, and a flexible circuit board 12 are sequentially arranged.

[0058] See also Figure 3 and Figure 4 The conductive sponge 14 is a mesh sponge structure, and its material contains conductive material. When pressed, it will be compressed and deformed, which increases the density of the internal conductive material and reduces the resistance. This is an inherent characteristic of the commercially available conductive sponge 14 itself, and there is no need to look for a conductive sponge 14 made of special materials.

[0059] See also Figure 4There are two types of electrode points distributed on the flexible circuit board 12, namely, excitation electrode points 121 and sensing electrode points 122. The voltage between adjacent excitation electrode points 121 and sensing electrode points 122 is detected by a voltage acquisition circuit; exemplarily, the voltage acquisition circuit includes a power supply 124, a voltage detector and several reference resistors 123. All excitation electrode points 121 are connected in parallel through a circuit to the positive pole of the power supply 124, each sensing electrode point 122 is connected to a 10 kilo-ohm reference resistor 123, and all reference resistors 123 are connected in parallel to the negative pole of the power supply 124. The excitation electrode points 121 and the sensing electrode points 122 are in a disconnected state on the flexible circuit board 12. The voltage detector (voltmeter 125) can measure the voltage divided value of the reference resistor 123 and convert it into the voltage between the excitation electrode point 121 and the sensing electrode point 122 through calculation. The voltage detector is set on the flexible circuit board 12 and is connected to the external microcontroller 161 for collecting voltage signals through a cable.

[0060] See also Figure 3 and Figure 4 The isolation net 13 is a single-layer grid structure made of insulating material, which can be made of flexible plastic or insulating rubber. It is used to separate the conductive sponge 14 and the flexible circuit board 12. However, when the conductive sponge 14 is squeezed, the sponge can contact the flexible circuit board 12 through the grid holes, so that the excitation electrode point 121 and the sensing electrode point 122 on it are conductive.

[0061] See also Figure 3 The sensor support plate 11 is a fixed base for the tactile sensor. It has protrusions and holes for fixing the flexible circuit board 12 and the soft layer 15, as well as mounting holes 111 connected to the fixed plate on the ultrasonic integrated device 5. It is made of resin using a 3D printing process.

[0062] See also Figure 5 The ultrasonic probe 3 is connected to the computer 162 through the video capture card 163 to transmit the ultrasonic image to the computer 162, and the flexible tactile sensor 1 is connected to the computer 162 through the single chip microcomputer 161 to transmit the collected contact data to the computer 162.

[0063] See also Figure 6The posture monitoring device 4 is a Vicon motion capture system commonly used in the prior art, which includes five visual cameras 43 and a positioning device. For example, the positioning device includes a cross positioning frame 41 and four markers mounted on the cross positioning frame 41. The Vicon motion capture system is a positioning system that can locate the position of marker points 42 using the visual cameras 43. The cross positioning frame 41 is fixed to the ultrasonic integrated device 5, and four marker points 42 are attached to establish a rigid coordinate system. This allows the position and posture of the ultrasonic integrated device 5 to be calculated based on the relative coordinates of the four marker points 42.

[0064] See also Figure 7 The present invention also provides a teaching and learning process of the ultrasonic robot, comprising the following steps:

[0065] Step 1: Fix the cross positioning frame 41 on the ultrasound integrated device 5 and attach four marker points 42; place the silicone kidney phantom 6 to be tested within the working area of ​​the visual camera 43 and set up the teaching scene;

[0066] Step 2: A professional ultrasound doctor holds an integrated ultrasound device 5 and performs an ultrasound examination on the silicone kidney phantom 6;

[0067] Step 3: The ultrasonic image is transmitted to the computer 162 via the video capture card 163, the contact data of the flexible tactile sensor 1 is collected via the single-chip microcomputer 161, and the position and posture of the ultrasonic integrated device 5 are recorded via the Vicon motion capture system. All data are transmitted to the computer 162 for storage;

[0068] Step 4: Repeat the scanning process several times to cover all important scanning locations and scanning sections as much as possible, and save and record all the data of the inspection process;

[0069] Step 5: The data recorded during the teaching process is made into a training data set, which is input into the probe posture adjustment prediction neural network for training, fitting the relationship between the ultrasound image, contact data and the posture change of the ultrasound integrated device 5, and constructing a teaching model;

[0070] Step 6: Remove the cross positioning frame 41 and the marker point 42 on the ultrasonic integrated device 5, and install the ultrasonic integrated device 5 on the end flange of the six-axis robot arm 7;

[0071] Step 7: Using the trained teaching model, the real-time ultrasound image and contact data are input into the network to predict the next moment's position change of the ultrasound integrated device 5, and convert it into the next moment's end position of the six-axis robot arm 7, driving the six-axis robot arm 7 to make corresponding adjustments;

[0072] Step 8: Control the six-axis robotic arm 7 to move to the next moment's position calculated in step 7, completing one cycle of motion prediction and execution. Then, repeat step 7 to control the six-axis robotic arm 7 to continuously scan, thereby realizing autonomous ultrasonic scanning. Ultrasonic images are recorded during the entire process.

[0073] Step 9: Monitor the change in the posture of the robotic arm. When the change in posture is lower than the set threshold for 10 consecutive seconds, terminate the ultrasonic scan.

[0074] It should be noted that the teaching model is based on a neural network model. The contact data and ultrasonic image of the ultrasonic probe 3 detected by the flexible tactile sensor 1 are the input of the neural network model, and the control instruction is the output of the neural network model. The neural network model includes an autoencoder and a probe posture adjustment prediction neural network. The autoencoder is used to reduce the dimension of the image collected by the ultrasonic probe 3. The probe posture adjustment prediction neural network model is used to predict the change in the posture of the ultrasonic probe 3 at the next moment based on the reduced-dimensional ultrasonic image data, that is, the control instruction, to control the movement of the six-axis manipulator 7 to change the posture of the ultrasonic probe 3. The control part of the six-axis robot can adopt existing technology. Specifically, the teaching model sends the control instruction information of the posture change amount. After receiving the information, the controller of the six-axis robot can solve the movement amount of each joint of the six-axis manipulator through the kinematics of the manipulator arm and then perform the corresponding action. The present invention does not limit the network structure of the probe posture adjustment prediction neural network, because most neural networks have the ability to fit this nonlinear relationship. It only needs to ensure that the root mean square error between the final predicted value of the neural network and the true value is within the allowable range.

[0075] Exemplarily, the probe posture adjustment prediction neural network is a deep neural network, which can provide better prediction results in the present invention.

[0076] It should be noted that the multi-degree-of-freedom mobile platform can adopt a six-axis robotic arm 7, but is not limited thereto. Any robotic mobile platform with six degrees of freedom at the end can be used.

[0077] The present invention first uses the ultrasonic integrated device 5 for teaching, and a professional ultrasound doctor holds it in hand to collect teaching data. After obtaining sufficient learning samples, it is used to train the probe posture adjustment prediction neural network to complete the learning of the doctor's scanning skills; then the ultrasonic integrated device 5 is installed at the end of the multi-degree-of-freedom mobile platform. The present invention uses a six-axis robotic arm 7, but is not limited to this. Any robot mobile platform with six degrees of freedom at the end can be used. Then, the trained neural network model is used to predict the next movement of the robot based on the real-time ultrasonic image and contact data during the scanning process, thereby controlling the robot to complete autonomous ultrasonic scanning.

[0078] The above embodiments are intended to illustrate the present invention only and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that various combinations, modifications, or equivalent substitutions of the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention and should be encompassed by the scope of the claims of the present invention.

Claims

1. An ultrasonic robot with autonomous scanning function, characterized in that: include: An ultrasonic probe is installed at the execution end of the multi-degree-of-freedom mobile platform and is used for performing ultrasonic scanning while being carried by the multi-degree-of-freedom mobile platform; A flexible tactile sensor, fixed together with the ultrasound probe to form an ultrasound integrated device, is used to monitor the contact data between the ultrasound probe and the scanned object; A controller having a built-in teaching model for recording teaching information and training the teaching model during manual teaching. The controller then uses the trained teaching model to generate control instructions during scanning, and controls the multi-degree-of-freedom mobile platform carrying the ultrasonic integrated device to perform autonomous scanning according to the control instructions. The teaching information includes ultrasonic images, contact data, and position information of the ultrasonic probe. The flexible tactile sensor comprises a sensor support plate, a flexible circuit board, an isolation net, a conductive sponge and a soft layer which are stacked in sequence; The sensor support plate is used to provide overall support; The flexible circuit board is provided with a plurality of electrode points distributed in an array and a voltage acquisition circuit for measuring the resistance signal between adjacent electrode points; The isolation net is made of insulating material and has a large number of mesh holes, which are used to isolate the electrode points of the conductive sponge and the flexible circuit board under normal circumstances. When the conductive sponge is squeezed by an external force, the adjacent electrode points are connected through the mesh holes. The resistance between the adjacent electrode points is detected by the voltage acquisition circuit to obtain contact data, which includes the contact position and the contact force. The soft layer is used for contacting with the scanned object and providing protection for the conductive sponge.

2. The ultrasonic robot with autonomous scanning function according to claim 1, characterized in that: The flexible tactile sensor and the ultrasonic probe are connected together through a fixing frame, and the fixing frame includes a first fixing plate and a second fixing plate. One end of the first fixing plate is configured as a connection end connected to the multi-degree-of-freedom mobile platform, and the other end is detachably connected to the second fixing plate to form a clamping end for the ultrasonic probe; There are two flexible tactile sensors, which are fixed on the first fixing plate and the second fixing plate on both sides of the ultrasonic probe respectively, and the detection surface of the flexible tactile sensor is configured to have the same shape and height as the scanning surface of the ultrasonic probe.

3. The ultrasonic robot with autonomous scanning function according to claim 1, characterized in that: The sensor support plate, flexible circuit board, isolation net, conductive sponge and soft layer are stacked in an arc-shaped package. After stacking, the outer surface of the soft layer forms a scanning surface with the same shape as the surface of the ultrasonic probe.

4. The ultrasonic robot with autonomous scanning function according to claim 1, characterized in that: The electrode points include staggered excitation electrode points and sensing electrode points. The voltage acquisition circuit includes a power supply, a voltage detector, and a plurality of reference resistors. All the excitation electrodes are connected in parallel to one electrode of the power supply, and each sensing electrode is connected to another electrode of the power supply via a reference resistor. The voltage detector detects the resistance between the excitation electrode point and the sensing electrode point by detecting the voltage score of each reference resistor.

5. The ultrasonic robot with autonomous scanning function according to any one of claims 1 to 4, characterized in that: The controller also includes a video capture card, the controller includes a single chip microcomputer and a computer with a built-in teaching model, and the video capture card is connected to the ultrasound probe and the computer respectively, and is used to transmit the image information collected by the ultrasound probe to the computer; The single chip microcomputer is respectively connected to the flexible tactile sensor, the computer, and the multi-degree-of-freedom mobile platform, and is used to transmit the contact data detected by the flexible tactile sensor to the computer, and receive instructions from the computer to control the multi-degree-of-freedom mobile platform to carry the ultrasonic probe for scanning.

6. A teaching method for an ultrasonic robot according to any one of claims 1 to 5, wherein a posture monitoring device is used to capture the posture of the ultrasonic probe during the teaching process, characterized in that: The following steps are involved: Install the posture monitoring device on the ultrasonic integrated device and set up the teaching scene; The instructor uses a handheld ultrasonic integrated device to scan the test object for ultrasonic testing. During the scanning process, the flexible tactile sensor monitors the contact data of the ultrasonic probe, and the posture monitoring device detects the posture information of the ultrasonic probe. Use the collected contact data and posture information to build a teaching model; The posture monitoring device was removed, and the ultrasonic integrated device was installed on the execution end of the multi-degree-of-freedom mobile platform. The teaching model was used to drive the multi-degree-of-freedom mobile platform to carry the ultrasonic integrated device for autonomous scanning and ultrasonic testing.

7. The teaching method of the ultrasonic robot according to claim 6, characterized in that: The posture monitoring device is a vicon motion capture system, which includes a positioning device and several visual cameras for visual capture. The positioning device is installed on the ultrasonic integrated device, and the shooting area of ​​the visual camera covers the active area of ​​the ultrasonic integrated device during the teaching process.

8. The teaching method of the ultrasonic robot according to claim 7, characterized in that: The positioning device includes a cross positioning frame and a plurality of marking points installed on the cross positioning frame. During teaching, the cross positioning frame is installed on the top of the ultrasonic integrated device.

9. The teaching method of the ultrasonic robot according to claim 6, characterized in that: The teaching model is a neural network model, including an autoencoder and a deep neural network. The autoencoder is used to reduce the dimension of the images collected by the ultrasound probe, and the deep neural network is used to predict the position of the ultrasound probe based on the ultrasound image data after dimension reduction.

Images