Robot foot device fused with force and position sensing and sensing method of robot foot device

Abstract

The invention discloses a robot foot device integrating force and position sensing and a sensing method of the robot foot device, and belongs to the technical field of force and position sensing. The device comprises a wear-resistant body, a clamping groove is formed in the surface of the wear-resistant body, a sensing assembly is clamped to the top of the wear-resistant body, a collecting plate is clamped to the top of the sensing assembly, a magnetic collecting module is arranged on the collecting plate, the sensing assembly comprises a base plate, an elastic body and a magnetic source, and the base plate is located between the wear-resistant body and the collecting plate; an elastic body is arranged on the surface of the base plate, and a magnetic source is fixedly connected to the surface of the elastic body. According to the robot foot device, the force position sensing technology is fused, the magnetic sensing algorithm, the flexible sensing algorithm and the neural network algorithm are integrated, the force position sensing algorithm is combined, and the foot contact position and the stress condition of the foot contact position can be accurately sensed. Therefore, the force position feedback control of the foot type robot can be realized, and the stability and the motion performance of the robot in a complex environment can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of force-position sensing, and specifically to a robot foot device integrating force-position sensing and its sensing method. Background Art

[0002] With the rapid development of technology, robots are increasingly widely used in many fields such as manufacturing, medical care, education and training. However, how to improve the sensing ability of robots, especially the force-position sensing ability of the feet, has always been a research hotspot in the field of robot technology. Force-position sensing technology is of great significance for improving the motion stability, positioning accuracy and environmental adaptability of robots.

[0003] In recent years, the development of magnetic sensing technology and flexible sensing technology has provided new solutions for force-position sensing of robot feet. Magnetic sensors utilize the principle of electromagnetic induction and can detect other parameters without contacting objects, with a long service life and high reliability. Flexible sensors, with their flexible structure, light weight, bendability and compatibility with various materials, enable robots to more naturally sense the surrounding environment.

[0004] In the application of force-position sensing of robot feet, magnetic sensors can be used to detect the position and posture of the feet, providing accurate positioning information; while flexible sensors can be used to sense the contact force and pressure distribution between the feet and the ground, providing real-time feedback for the motion control of the robot. By combining magnetic sensing and flexible sensing technologies, it is possible to achieve all-round and high-precision force-position sensing of the robot feet, thereby improving the motion performance and stability of the robot.

[0005] Aiming at the deficiencies of existing robot foot sensing technologies, the present invention proposes a force-position sensing robot foot device, which integrates magnetic sensing and flexible sensing technologies, aiming to improve the sensing ability and motion performance of robots.

[0006] Currently, there are some robot foot devices on the market that use sensing technologies for force-position sensing. Most of these devices use traditional rigid sensors, such as pressure sensors, force sensors, etc., to detect the force on the feet. Although these sensors can provide certain force-position information, they still have limitations in some aspects; Firstly, due to their fixed shape and size, rigid sensors are difficult to adapt to the changing shapes of robot feet on complex terrains, which may lead to inaccurate measurement or untimely response of the sensors in some cases. Secondly, traditional sensors are often heavy, increasing the mass of the robot feet, which may have an adverse impact on the motion performance and energy consumption of the robot. In addition, existing foot sensing systems usually have limited data processing capabilities and are difficult to achieve the fusion and optimization of data from multiple sensors, thus affecting the accuracy and real-time performance of sensing. Summary of the Invention

[0007] In view of the deficiencies of the prior art, the present invention provides a robot foot device integrating force and position sensing, which solves the problem that most of the prior art uses traditional rigid sensors, resulting in limitations in some aspects.

[0008] To achieve the above objectives, the present invention is realized through the following technical solutions: A robot foot device integrating force and position sensing, comprising a wear-resistant body, a sensing component, a collection board, and a connecting piece; A card slot is provided at the top of the wear-resistant body; The sensing component includes a substrate, an elastomer, and a magnetic source. The substrate is inlaid with the elastomer, and the magnetic source is fixedly connected to the top surface of the elastomer. The top of the elastomer is clamped in the top card slot of the wear-resistant body. Among them, the magnetic source is stationary relative to the wear-resistant body, and the elastomer can generate deformation; The collection board is clamped on the top of the substrate of the sensing component, so that the collection board is stationary relative to the substrate, and a magnetic collection module is provided on the collection board; The connecting piece is fixedly screwed to the collection board through a threaded hole, is stationary relative to the collection board, and deforms with the elastomer of the sensing component, so that the connecting piece and the collection board move relative to the wear-resistant body.

[0009] Preferably, the wear-resistant body is made of hard silicone by die casting, and wear-resistant patterns are provided on its surface. The wear-resistant body includes but is not limited to a semi-circular structure.

[0010] Preferably, the sensing component is integrally inlaid with a substrate, an elastomer, and a magnetic source. Among them, the material of the elastomer is deformable soft silicone, so that the magnetic source can move relative to the substrate as the elastomer deforms.

[0011] Preferably, the collection board is a PCB rigid circuit board or an FPC flexible circuit board, which integrates a control module, a magnetic collection module, and a power management module. Among them, the magnetic collection module can use a magnetoresistive sensor, a Hall effect sensor, a magnetic induction coil, etc. The control module selects but is not limited to common 32-bit single-chip microcontrollers, such as the STM32 series or the ESP32 series. The controller is equipped with program modules for data acquisition, data processing, and data output.

[0012] Preferably, the preparation method of the robot foot device integrating force and position sensing specifically includes the following steps: The preparation method of the wear-resistant body is: position and fasten the two parts of the mold, cover the mold cover, configure the wear-resistant body material, inject the hard silicone material through the pouring hole, and demold to obtain the wear-resistant body; The preparation method of the sensing component is as follows: Assemble the mold, seal the perimeter with a rubber band, prepare the sensing layer material, inject the silicone material into the mold cavity through the pouring hole, place it in a vacuum pump to exhaust air, repeat the injection of the material 2 - 3 times, heat and cure to form, cut off the excess part after demolding, and inlay and fix the magnetic source; For the acquisition board, select a rigid PCB circuit board or a flexible FPC circuit board for SMT processing in different scenarios; For the connecting piece, use aluminum CNC machining under working conditions requiring high strength, and use nylon 3D printing under normal working conditions. Fillet or chamfer treatment is required at the connection with the robot ankle joint.

[0013] Preferably, the assembly method of the wear-resistant body and the sensing component is as follows: Inlay and install the magnetic source in the groove of the elastomer of the sensing component, fill a small amount of silicone adhesive and remove the excess, let it stand and cure, position the elastomer layer with the groove on the flat side of the wear-resistant body, apply acidic adhesive to fit, complete the connection, bond the elastomer layer and the wear-resistant body and position them with the substrate and the positioning holes on the acquisition board, and use locknuts to fasten with the connecting piece to ensure stress isolation of the acquisition board.

[0014] A force and position sensing method for a robot foot device integrating force and position sensing, the specific features include the following steps: Step 1: Optimize the three-dimensional magnetic field data from the magnetic sensor array by using various techniques such as moving average filtering, statistical filtering, and low-pass filtering to effectively reduce the data volume and improve the data quality; Step 2: According to the task requirements, add corresponding labels to the data points, and synchronize the magnetic field data with the position information of the sensor, the data of the force sensor, etc. in time or space; Step 3: Extract useful features from the original three-dimensional magnetic field data, and perform standardization or normalization processing on the feature data to ensure that the range of the input features is consistent, which is convenient for better training of the neural network; Step 4: Divide the processed data set into a training set, a validation set, and a test set. The common ratio is 70% - 80% for training, 10% - 15% for validation, and the rest for testing; If the data volume is insufficient, more samples can be generated through data augmentation techniques such as adding noise, data rotation, and translation; Step 5: Load the processed data set into the neural network training framework, specify the input data and labels, set the batch size and learning rate according to the network structure, and perform iterative training; Step 6: After training, use the validation set or the test set to evaluate the prediction accuracy of the model, and the mean square error or other regression evaluation metrics can be calculated; Step 7: Based on the optimized model, perform real-time force and position sensing on the three-dimensional magnetic field data of the robot foot device.

[0015] The present invention provides a robot foot device integrating force and position sensing, with the following beneficial effects: (1) The robot foot device integrating force and position sensing can realize the sensing of the contact position and force of the foot by combining magnetic sensing and flexible sensing technologies. This not only helps to improve the stability and motion performance of the robot in complex environments, but also provides it with more accurate navigation and positioning capabilities.

[0016] (2) The robot foot device integrating force and position sensing can better adapt to the shape changes of the robot foot through the use of flexible sensors, ensuring accurate sensing data can be provided in various terrains and postures. This flexibility greatly improves the adaptability and operation performance of the robot in different scenarios.

[0017] (3) The robot foot device integrating force and position sensing realizes lightweight design: compared with traditional rigid sensors, the sensors adopted in this patent are lighter in weight, reducing the burden on the robot foot and helping to improve the motion efficiency and battery life of the robot.

[0018] (4) The robot foot device integrating force and position sensing realizes real-time data processing: the device of this patent has strong data processing capabilities, can fuse and optimize data from different sensors in real time, so as to provide more accurate and timely sensing information to support the rapid decision-making and response of the robot.

[0019] (5) The robot foot device integrating force and position sensing, through innovative combinations of sensing technologies and data processing methods, not only achieves high-performance sensing in this patent, but also takes into account cost-effectiveness. This enables the device to maintain a reasonable manufacturing cost while improving the performance of the robot, helping to promote the application of robots in a wider range of fields. Brief Description of the Drawings

[0020] Figure 1 is an exploded schematic view of the structure in the present invention; Figure 2 is a schematic view of the wear-resistant body structure in the present invention; Figure 3 is a physical diagram of the robot foot device in the present invention; Figure 4 is a calibration experimental scene diagram of the magnetic-force position conversion model in the present invention; Figure 5 is a three-dimensional pressure change diagram of the gait experiment in the present invention.

[0021] In the figures, 1 - wear-resistant body, 2 - sensing component, 21 - substrate, 22 - elastomer, 23 - magnetic source, 3 - acquisition board, 4 - connecting piece. Detailed Embodiments

[0022] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0023] Please refer to Figures 1-3 , an embodiment of the present invention provides a technical solution: a robot foot device integrating force and position sensing, including a wear-resistant body 1. A card slot is provided on the surface of the wear-resistant body 1. A sensing component 2 is clamped on the top of the wear-resistant body 1. A collection board 3 is clamped on the top of the sensing component 2. A connecting piece 4 is clamped on the top of the collection board 3. A sensor is fixedly connected to the outer surface of the sensing component 2. A magnetic collection module is provided on the collection board 3. The collection board 3 is a PCB rigid circuit board or an FPC flexible circuit board, which integrates a control module, a magnetic collection module, and a power management module. Among them, the magnetic collection module can use a magnetoresistive sensor, a Hall effect sensor, a magnetic induction coil, etc. The control module selects a common 32-bit single-chip microcomputer, such as the STM32 series or the ESP32 series. The controller is equipped with program modules for data collection, data processing, and data output. The sensing component 2 includes a substrate 21. The substrate 21 is located between the wear-resistant body 1 and the collection board 3. An elastic body 22 is provided on the surface of the substrate 21. A magnetic source 23 is fixedly connected to the surface of the elastic body 22. The material of the wear-resistant body 1 is made of silicone. The elastic body 22 is made of extremely soft silicone material. Wear-resistant patterns are etched or die-cast on the surface of the wear-resistant body 1. The wear-resistant body 1 includes, but is not limited to, a semi-circular structure. The sensing layer is the core part for the device to achieve the sensing function, and is composed of a magnetic source, an elastic body, and a substrate. The magnetic source can be a magnetic body with appropriate size and magnetic field strength, such as a micro magnet or a magnetic film. In this example, a micro-cylindrical magnet (neodymium iron boron strong magnet, grade N35, surface galvanized treatment) is used. Among them, the specific structure of the elastic layer can be changed according to the actual situation, and the stiffness of the elastic body is changed to change the range and sensitivity of force sensing. In this example, a solid silicone column is used, and the magnetic source is embedded at the top of the cylinder. The forming method is deposition forming. The silicone material uses ecoflex-0010 or 0030 produced by Smooth-On Company. The wear-resistant body 1 is the part of the foot in contact with the ground, and will be firmly connected to the sensing layer through an acidic adhesive to become an integrated body.

[0024] A robot foot device integrating force and position sensing, comprising a wear-resistant body 1, on the surface of which there are card slots; on the top of the wear-resistant body 1 there is a sensing component 2, on the top of the sensing component 2 there is a collection board 3, and on the top of the collection board 3 there is a connecting piece 4; a sensor is fixedly connected to the outer surface of the sensing component 2; a magnetic collection module is integrated on the collection board 3. The wear-resistant body 1 is made of a hard material, and the elastomer 22 is made of a softer material. Due to the deformation of the elastomer, the wear-resistant body can be displaced relative to the collection board. The surface of the wear-resistant body 1 is fixedly connected with wear-resistant patterns. The wear-resistant body 1 can be semi-circular or other structures. The sensing component 2 includes a substrate, an elastomer and a magnetic source. The hardness of the elastomer is much smaller than that of the wear-resistant body. The collection board 3 can be a PCB rigid circuit board or an FPC flexible circuit board, which integrates a control module, a magnetic collection module and a power management module. Among them, the magnetic collection module can use magnetoresistive sensors, Hall effect sensors, magnetic induction coils, etc. The control module can select single-chip microcomputers of the STM32 series or ESP32 series, and is equipped with data acquisition, processing and output program modules. The collection board 3 is equipped with a magnetic-force and position conversion model, the input of which is three-dimensional magnetic field data, and the output is three-dimensional force data and the position of the contact point.

[0025] A preparation method of a robot foot device, comprising the following steps: Preparation of the wear-resistant body 1: Position and fasten the mold, configure the mold rubber, stir and evacuate, pour while tilting the mold, heat and cure at 45 °C, demold. Preparation of the substrate in the sensing component 2: Assemble the mold, configure soft silica gel, stir and evacuate, pour and cure, demold and cut. Preparation of the connecting piece 4: Use aluminum CNC machining under high-strength working conditions, and use nylon 3D printing under normal working conditions. The connection part needs to be rounded or chamfered. The assembly method of the wear-resistant body 1 and the sensing component 2 is: Install a magnetic source with permanent magnetism at the bottom of the sensing component 2 and fix it with an adhesive. Position the magnetic source and the wear-resistant body 1 and bond them with an acidic adhesive. Connect the sensing layer and the collection board 3 through the mold and positioning holes, and fasten them with a locknut and the connecting piece 4.

[0026] A preparation method of a robot foot device integrating force and position sensing specifically includes the following steps; The preparation method of the wear-resistant body 1 is: Position and fasten the two parts of the mold, cover the mold cover, and at the same time configure the material of the contact layer 4: Weigh 80 grams of food-grade mold silica gel E640A agent and mix it with the B agent in a ratio of 1:1, stir for 2 minutes, put the mixed silica gel material into a vacuum pump to evacuate, let it stand for 2 - 3 minutes and then exhaust. When pouring, tilt the mold slightly and inject the silica gel material through the pouring hole. Heat at 45 °C for 1.2 hours and demold to obtain the wear-resistant body 1; The preparation method of the substrate in the sensing component 2 is as follows: Assemble the mold, seal the four sides with rubber rings, and configure the sensing layer material: Weigh 50 grams of Ecoflex 00-10 A agent and mix it with B agent in a ratio of 1:1, stir for 1 minute, put the mixed material into a vacuum pump to evacuate, let it stand for 1-2 minutes and then exhaust, inject the silicone material into the mold cavity through the pouring hole, put it into the vacuum pump to exhaust, repeat injecting the material 2-3 times, heat it at 60 °C for 45 minutes to cure and form, and cut off the excess part after demolding; The connecting piece 4 is machined by aluminum CNC under working conditions requiring high strength, and is 3D printed by nylon under normal working conditions. The connection with the robot ankle joint requires rounding or chamfering treatment.

[0027] The assembly method of the wear-resistant body 1 and the sensing component 2 is as follows: Install a cylindrical magnet with a diameter of 2 mm in the bottom groove of the sensing component 2, fill a small amount of silicone adhesive and remove the excess part, let it stand for 12 minutes to cure, position the magnet in the sensing layer with the groove on one side of the flat surface of the wear-resistant body 1, apply acidic adhesive to fit, complete the connection, bond the sensing layer and the wear-resistant body 1 and position them with the mold and the positioning holes on the acquisition board 3, and use locknuts to fasten with the connecting piece 4 to ensure that the acquisition board completes stress isolation.

[0028] During use, when the wear-resistant body 1 contacts the ground, deformation will occur. This deformation is transmitted to the sensor in the sensing component 2. The spatial position of the magnetic source 23 in the elastic body 22 in the sensing component 2 changes due to the deformation, which in turn causes the three-dimensional magnetic field intensity to change. The device collects these changes in magnetic field intensity through the magnetic acquisition module installed on the acquisition board 3. The magnetic-force position conversion model is integrated in the control module on the acquisition board 3. Through the calculation of the control module, the magnitude of the three-dimensional force is converted. This method enables the device to accurately sense and calculate the force generated by the ground contact, and obtain the position where the wear-resistant body 1 contacts the ground, thereby realizing high-precision force-position feedback and measurement.

[0029] To sum up As Figure 4 shown, it is a physical diagram of the robot foot device installed on a quadruped robot. During actual use, it can be installed at the end ankle joint of various robots.

[0030] As Figure 5 shown, it is a calibration experimental diagram of the magnetic-force position conversion model. The magnetic-force position conversion model is a neural network model obtained through a calibration method. The calibration method is as follows: (1) Experimental preparation: Place the device on a three-dimensional mobile test platform with three-axis position feedback to ensure stable and reliable installation; (2) Force sensor calibration: Install a high-precision six-axis force sensor on the mobile end of the platform to record the actually applied three-dimensional force. The force sensor needs to be precisely calibrated to ensure the accuracy of its measurement; (3) Multi-point measurement: Apply forces with known magnitudes and directions at different positions and angles to ensure that various situations within the working range of the device are covered. At each measurement point, record the corresponding three-dimensional magnetic field intensity in the sensing layer; (4) Data acquisition: Through the magnetic acquisition module on the acquisition board 3, acquire the magnetic field intensity data at each measurement point, and at the same time record the true three-dimensional force value measured by the force sensor and the position feedback of the three-dimensional mobile platform; (5) Data processing: Pair the acquired magnetic field intensity data with the corresponding true force values and the positions of the measurement points to form an input-output data set. The data set contains multi-dimensional information to ensure that the neural network model can learn the force-magnetic field relationship under different situations; (6) Model training: Use the acquired data set to train a neural network model. The input is the change in magnetic field intensity measured by the magnetic acquisition module, and the output is the corresponding three-dimensional force value and the force application position. Adopt the method of supervised learning. By adjusting the network parameters, minimize the error between the force position predicted by the model and the actual force position; (7) Model verification: Use a part of the data set for model verification to ensure that the neural network model can perform well on unseen data. Through methods such as cross-validation, evaluate the generalization ability and accuracy of the model; (8) Model optimization: According to the verification results, it may be necessary to adjust the structure of the neural network or re-acquire more data to improve the accuracy of the model. The optimized model should be able to stably convert the change in magnetic field intensity into accurate three-dimensional force positions in practical applications.

[0031] Through the above calibration method, the obtained neural network model can reliably perform magnetic-force position conversion, enabling the device to accurately sense and calculate the force generated by ground contact and the surface position of the contact point on the wear-resistant body 1 in practical applications, thereby realizing high-precision force position feedback and measurement.

[0032] To illustrate the good measurement effect of the robot spherical foot device, Figure 5 the three-dimensional pressure data obtained from real-time tests on a single-foot experimental platform are given. For the convenience of describing the motion characteristics, the direction of the three-dimensional force received by the foot and the ground is as Figure 5 shown in (d) below. The spherical foot test data are converted into a three-dimensional force through the model. Since it is difficult to determine the contact area of the spherical foot, the force is considered to act on the elastic body. For example, dividing the Z-direction by the total cross-sectional area of the elastic body gives the three-dimensional pressure situation, as shown in Figure 5 (a), (b), and (c) below.

[0033] The robot ball foot device provided by the present invention is particularly applicable to the environmental perception of legged robots.

[0034] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation.

[0035] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A robot foot device integrating force and position sensing, characterized in that: It includes a wear-resistant body (1), a sensing component (2), a collecting board (3), and a connecting piece (4); A card slot is provided at the top of the wear-resistant body (1); The sensing component (2) includes a substrate, an elastic body, and a magnetic source. The substrate is inlaid with the elastic body. The top surface of the elastic body is fixedly connected with the magnetic source, and the top of the elastic body is clamped in the top card slot of the wear-resistant body (1). Among them, the magnetic source is stationary relative to the wear-resistant body, and the elastic body can generate deformation; The collecting board (3) is clamped on the top of the substrate of the sensing component, so that the collecting board is stationary relative to the substrate, and a magnetic collecting module is arranged on the collecting board; The connecting piece (4) is fixedly screwed to the collecting board through a threaded hole, is stationary relative to the collecting board, and deforms with the elastic body of the sensing component, so that the connecting piece and the collecting board move relative to the wear-resistant body.

2. The robot foot device with integrated force and position sensing according to claim 1, characterized in that: The wear-resistant body (1) is made of hard silicone by die casting, and wear-resistant patterns are provided on its surface. The wear-resistant body (1) includes but is not limited to a semi-circular structure.

3. The robot foot device with integrated force and position sensing according to claim 1, characterized in that: The sensing component (2) is inlaid with the substrate, the elastic body, and the magnetic source as a whole. Among them, the material of the elastic body is deformable soft silicone, so that the magnetic source can move relative to the substrate as the elastic body deforms, and the magnetic source is at least but not limited to one.

4. The robot foot device with integrated force and position sensing according to claim 1, characterized in that: The collecting board (3) is a PCB rigid circuit board or an FPC flexible circuit board, which integrates a control module, a magnetic collecting module, and a power management module. Among them, the magnetic collecting module can use a magnetoresistive sensor, a Hall effect sensor, a magnetic induction coil, etc. The control module selects but is not limited to common 32-bit single-chip microcomputers, such as the STM32 series or the ESP32 series. The controller is equipped with program modules for data acquisition, data processing, and data output.

5. A robot foot device integrating force and position sensing according to any one of claims 1-4, characterized in that: Its preparation method specifically includes the following steps: The preparation method of the wear-resistant body (1) is: position and fasten the two parts of the mold, cover the mold cover, configure the material of the wear-resistant body (1), inject hard silicone material through the pouring hole, and demold to obtain the wear-resistant body (1); The preparation method of the sensing component (2) is: assemble the mold, seal the periphery with a rubber ring, configure the material of the sensing layer, inject the silicone material in the mold cavity through the pouring hole, put it into a vacuum pump to exhaust air, repeat injecting the material 2-3 times, heat and cure to form, cut off the excess part after demolding, and inlay and fix the magnetic source; The collecting board (3) selects a PCB rigid circuit board or an FPC flexible circuit board for SMT processing in different scenarios; The connecting piece (4) is processed by CNC with aluminum in working conditions that require high strength, and is 3D printed with nylon in ordinary working conditions. The connection part with the robot ankle joint needs to be rounded or chamfered.

6. The robot foot device with integrated force and position sensing according to claim 1, characterized in that: The assembly method of the wear-resistant body (1) and the sensing component (2) is: inlay and install the magnetic source in the groove of the elastic body of the sensing component (2), fill a small amount of silicone adhesive and remove the excess part, let it stand and cure, position the elastic layer and the groove on the flat side of the wear-resistant body (1), apply acidic adhesive to fit, complete the connection, and bond the elastic layer and the wear-resistant body (1) and then position them with the collecting board (3) through the positioning holes on the substrate and the collecting board (3), and use a locknut to fasten the connecting piece (4) to ensure that the stress isolation of the collecting board is completed.

7. The force and position sensing method of a robot foot device integrating force and position sensing according to any one of claims 1-6, characterized in that, The specific features include the following steps: Step 1: Optimize the three-dimensional magnetic field data from the robot foot device by using various techniques such as moving average filtering, statistical filtering, and low-pass filtering, effectively reducing the data volume and improving the data quality; Step 2: According to the task requirements, add corresponding labels to the data points, and synchronize the magnetic field data with the position information of the sensor, the data of the force sensor, etc. in time or space; Step 3: Extract useful features from the original three-dimensional magnetic field data, and perform standardization or normalization processing on the feature data to ensure that the range of the input features is consistent, facilitating better training of the neural network; Step 4: Divide the processed data set into a training set, a validation set, and a test set. A common ratio is 70%-80% for training, 10%-15% for validation, and the rest for testing; if the data volume is insufficient, more samples can be generated through data augmentation techniques such as adding noise, data rotation, and translation; Step 5: Load the processed data set into the neural network training framework, specify the input data and labels, set the batch size and learning rate according to the network structure, and perform iterative training to obtain a training model; Step 6: After training, use the validation set or the test set to evaluate the prediction accuracy of the model, and the mean square error or other regression evaluation metrics can be calculated; Step 7: Based on the optimized model, perform real-time force and position perception on the three-dimensional magnetic field data of the robot foot device.

8. The force and position sensing method of a robot foot device with integrated force and position sensing according to claim 7, characterized in that: The original three-dimensional magnetic field data is the data obtained by the magnetic acquisition module on the acquisition board measuring the magnetic field generated by the magnetic source on the sensing component.

9. The force and position sensing method of a robot foot device with integrated force and position sensing according to claim 7, characterized in that: The features can be, but are not limited to, the magnetic field intensity amplitude, the magnetic field direction, and the change rate of the magnetic field over time or space.

10. The force and position sensing method of a robot foot device with integrated force and position sensing according to claim 7, characterized in that: The model will be carried by the control module on the acquisition board, and the data conversion process will be performed by the control module.

Images