Glove-based multi-modal data acquisition system, multi-modal synchronous data acquisition method

Abstract

The embodiment of the application provides a glove-based multi-modal data acquisition system and a multi-modal synchronous data acquisition method, wherein the joint sliding electric contact node device comprises: an insulating base body; a conductive track; a contact assembly comprising a first contact part in sliding contact with a first type of track and a second contact part in sliding contact with a second type of track, the first contact part being configured to output a first electric signal for representing a rotation angle of the joint based on a contact position of the first contact part with the first type of track, and the second contact part being configured to output a second electric signal for representing a reference position event when the joint rotates through the reference position; and a master control unit electrically connected with the joint sliding electric contact node device, the master control unit generating a synchronous clock reference signal according to the second electric signal. The technical scheme of the embodiment of the application can realize joint angle detection and hardware level synchronous triggering simultaneously in a narrow space of the joint without increasing discrete devices.

Description

Technical Field

[0001] This application relates to the field of robotics technology, and in particular to a glove-based multimodal data acquisition system and a multimodal synchronous data acquisition method. Background Technology

[0002] With the rapid development of applications such as humanoid robot teleoperation, embodied intelligence dataset construction, and rehabilitation medical assessment, wearable data acquisition gloves, as key front-end devices for acquiring hand movement and tactile information, directly affect the performance of back-end algorithms through their acquisition accuracy and data quality. Among these, the temporal consistency of multimodal data, including tactile, joint posture, and visual data, is a crucial foundation for achieving high-fidelity teleoperation and high-quality model training.

[0003] Currently, the electrical solutions for wearable hand data acquisition mainly suffer from the following technical problems: Firstly, the angle detection and synchronous triggering functions at the joint rely on discrete sensors and conditioning circuits independent of the slip ring, resulting in the stacking of functional components and a large axial thickness at the joint, making it difficult to adapt to the narrow installation space of the finger joint.

[0004] Secondly, multimodal sensors sample according to their own independent clock domains, and cross-modal synchronization lacks a hardware-level synchronization reference signal. It relies on operating system scheduling or software post-processing for timestamp alignment, and the synchronization accuracy is difficult to guarantee due to interrupt response and bus contention.

[0005] Third, the joint harness is repeatedly bent during long-term use, which poses a risk of fatigue fracture and insufficient system reliability. Summary of the Invention

[0006] This application provides a glove-based multimodal data acquisition system and a multimodal synchronous data acquisition method, which can simultaneously achieve joint angle detection and hardware-level synchronous triggering within the narrow space of a joint without adding discrete components.

[0007] As one aspect of this application, this application provides a glove-based multimodal data acquisition system, including: A joint sliding electrical contact node device, comprising: Insulating substrate; Conductive tracks are disposed on the surface of the insulating substrate, the conductive tracks comprising first-class tracks and second-class tracks electrically isolated from each other; A contact assembly, comprising a first contact portion that slides in contact with a first type of track and a second contact portion that slides in contact with a second type of track, wherein the first contact portion is configured to output a first electrical signal characterizing a rotation angle of a joint based on the contact position between the first contact portion and the first type of track, and the second contact portion is configured to output a second electrical signal characterizing a reference position event when the joint rotates through a reference position; and The main control unit is electrically connected to the joint sliding electrical contact node device. The main control unit is configured to receive the first electrical signal and the second electrical signal, and generate a synchronous clock reference signal based on the second electrical signal.

[0008] As another aspect of the embodiments of this application, the embodiments of this application provide a multimodal synchronous data acquisition method, including: Acquire the zero-position electrical signal generated by the joint sliding electrical contact node device set at the finger joint when the joint rotates through the reference position; A synchronous clock reference signal is generated based on the zero-position electrical signal; The data acquisition device is controlled to perform synchronous sampling according to the synchronous clock reference signal; The data acquisition device includes multiple different types of acquisition units, which include two or more of the following: a tactile sensing unit, an inertial measurement unit, and a vision unit.

[0009] This application integrates the angle detection function (first type of track and first contact part) and the zero-position synchronization trigger function (second type of track and second contact part) into the same joint sliding electrical contact node device. The two functions share the same insulating substrate and contact components, eliminating the need for separate encoders or synchronization switching devices outside the slip ring. This significantly reduces the number of devices and axial thickness at the joint, making it suitable for the confined installation space of finger joints. Simultaneously, the second electrical signal output by the second type of track and second contact part is a hardware-generated transition signal, which can serve as a hardware-level synchronization reference, avoiding the timing uncertainties introduced by operating system scheduling and interrupt responses in software timestamping schemes.

[0010] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of this application will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0011] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments disclosed in this application and should not be construed as limiting the scope of this application.

[0012] Figure 1 A schematic block diagram of a multimodal data acquisition system according to an embodiment of this application is shown.

[0013] Figure 2 A schematic block diagram of a joint sliding electrical contact node device according to an embodiment of this application is shown.

[0014] Figure 3 A structural diagram of an insulating substrate according to an embodiment of this application is shown.

[0015] Figure 4 A spherical unfolded view of the surface of an insulating substrate according to an embodiment of this application is shown.

[0016] Figure 5 A schematic block diagram of a joint sliding electrical contact node device according to another embodiment of this application is shown.

[0017] Figure 6 An electrical schematic diagram of a multimodal data acquisition system according to an embodiment of this application is shown.

[0018] Figure 7 An electrical architecture block diagram of a multimodal data acquisition system according to an embodiment of this application is shown.

[0019] Figure 8 A flowchart illustrating a multimodal synchronous data acquisition method according to an embodiment of this application is shown.

[0020] Figure 9 A timing diagram of the zero-bit hardware synchronization link according to an embodiment of this application is shown.

[0021] Figure 10 A three-level priority bus scheduling time slot allocation diagram according to an embodiment of this application is shown.

[0022] Figure 11 A detailed flowchart of a multimodal synchronous data acquisition method according to an embodiment of this application is shown.

[0023] Explanation of reference numerals in the attached figures: 1000, Multimodal data acquisition system; 100, Joint sliding electrical contact node device; 110, Insulating substrate; 120, Conductive track; 121, Type I track; 1211, Conductive segment; 1212, Insulating segment; 122, Type II track; 123, Type III track; 124, Type IV track; 130, Contact assembly; 131, First contact part; 132, Second contact part; 140, Local processing circuit; 200, Main control unit. Detailed Implementation

[0024] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of this application. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0025] Please refer to the following: Figure 1 and Figure 2 The multimodal data acquisition system 1000 based on gloves provided in this application embodiment includes a joint sliding electrical contact node device 100 and a main control unit 200.

[0026] The joint sliding electrical contact node device 100 includes an insulating substrate 110, a conductive track 120, and a contact assembly 130. The conductive track 120 is disposed on the surface of the insulating substrate 110 and includes a first type track 121 and a second type track 122 electrically isolated from each other. The contact assembly 130 includes a first contact portion 131 that slides in contact with the first type track 121 and a second contact portion 132 that slides in contact with the second type track 122. The first contact portion 131 is configured to output a first electrical signal characterizing the rotation angle of the joint based on the contact position between the first contact portion 131 and the first type track 121. The second contact portion 132 is configured to output a second electrical signal characterizing a reference position event when the joint rotates through a reference position. A main control unit 200 is electrically connected to the joint sliding electrical contact node device 100 and is configured to receive the first and second electrical signals and generate a synchronous clock reference signal based on the second electrical signal.

[0027] Specifically, the insulating substrate 110 serves as the supporting foundation for the conductive rails 120. Its material can be ceramic, engineering plastics (such as polyimide, polyetheretherketone), or other rigid materials with good insulating properties to ensure the electrical isolation stability between the conductive rails 120. The conductive rails 120 can be formed on the surface of the insulating substrate 110 using low-resistivity metals such as copper, gold, or silver through processes such as electroplating, sputtering, and etching. Electrical isolation between the conductive rails 120 can be achieved by setting physical gaps or insulating filler layers between the rails.

[0028] The first type of track 121 and the first contact portion 131 constitute an angle detection unit. The first type of track 121 has electrical characteristics that vary with the angle (such as segmented conductivity and insulation structure). When the insulating substrate 110 rotates with the joint, the relative contact position between the first contact portion 131 and the first type of track 121 changes, causing the electrical signal (such as voltage, current, or on / off state) output at the first contact portion 131 to change accordingly. This changing electrical signal is the first electrical signal. The first electrical signal characterizes the current rotation angle of the joint.

[0029] The second type of track 122 and the second contact portion 132 constitute a zero-position detection unit. The second type of track 122 has conductive characteristics only in a local area on the surface of the insulating substrate 110 corresponding to the reference position. When the joint rotates to the reference position, the second contact portion 132 and the second type of track 122 form electrical contact, the circuit is turned on, and a transient electrical signal is generated, which is the second electrical signal. The appearance of the second electrical signal indicates that the joint has just passed the reference position, and therefore can be used as a timestamp of the reference position event. When the joint leaves the reference position, the second contact portion 132 and the second type of track 122 disconnect, and the circuit is cut off.

[0030] The main control unit 200 can be a microcontroller, digital signal processor, field-programmable gate array, or other electronic device with signal processing capabilities. After receiving the first electrical signal, the main control unit 200 can extract the joint angle information through analog-to-digital conversion, level judgment, or decoding algorithms. After receiving the second electrical signal, since the second electrical signal is a hardware-generated transient transition signal (rising edge or falling edge), the timing of its occurrence is determined by the physical contact process. The main control unit 200 can directly use this transition edge to generate a synchronization clock reference signal to trigger the synchronization operation of other data acquisition devices, thereby realizing a hardware-level synchronization reference.

[0031] It should be noted that the "reference position" mentioned above refers to the preset angle position that the joint needs to be accurately identified during the movement. It is usually the zero angle position of the joint (i.e., the position when the finger is straight), but it can also be set to any other angle position that needs to be triggered synchronously or marked as needed.

[0032] It should also be noted that the aforementioned "sliding contact" refers to a connection method in which the contact part maintains physical contact with the track surface and can move relative to it. During the joint rotation, the contact part and the track are in sliding friction contact, and electrical conductivity is always maintained between them, thereby continuously transmitting electrical signals during the motion state.

[0033] It should also be noted that the joint sliding electrical contact node device 100 can be disposed at any finger joint of the glove, including but not limited to the metacarpophalangeal joint, proximal interphalangeal joint, and distal interphalangeal joint. When the joint sliding electrical contact node device 100 is disposed at multiple joints, each device can independently detect the angle and reference position events of its respective joint. The aforementioned glove can be any type of wearable hand device, including industrial teleoperation data gloves, rehabilitation training gloves, virtual reality interactive gloves, etc.

[0034] The aforementioned "multimodal data acquisition" refers to the process of simultaneously acquiring data from two or more different types of sensors. In this application, it can specifically refer to the synchronous data acquisition between tactile sensing units (such as piezoresistive tactile sensor arrays), inertial measurement units (such as six-axis IMU sensors), and vision units (such as miniature cameras).

[0035] This embodiment integrates the angle detection function (first type of track 121 and first contact part 131) and the zero-position synchronization trigger function (second type of track 122 and second contact part 132) into the same joint sliding electrical contact node device 100. The two functions share the same insulating substrate 110 and contact component 130, eliminating the need for separate encoders or synchronization switching devices outside the slip ring. This significantly reduces the number of devices and axial thickness at the joint, making it suitable for the narrow installation space of finger joints. Simultaneously, the second electrical signal output by the second type of track 122 and second contact part 132 is a hardware-generated transition signal, which can serve as a hardware-level synchronization reference, avoiding the timing uncertainties introduced by operating system scheduling and interrupt responses in software timestamping schemes.

[0036] like Figure 3 As shown, in some embodiments, the insulating substrate 110 includes at least a spherical portion, and the conductive tracks 120 are circumferentially disposed on the surface of the insulating substrate 110. Specifically, the insulating substrate 110 adopts a spherical structure. Figure 3The sphere (shown schematically, but not representing the actual structure of the insulating substrate 110) preferably has a diameter ranging from 4 mm to 6 mm to accommodate the geometry of a human finger joint. The surface of the spherical insulating substrate 110 is a continuous curved surface of revolution, and the conductive track 120 can be arranged along the latitudinal direction (i.e., circumferential direction) of the sphere to form a ring or arc track. When the spherical insulating substrate 110 rotates around its center as the joint rotates, the fixedly arranged contact assembly 130 slides relative to it along the meridian direction on the sphere. The latitude of the contact point changes with the rotation angle. For example, when the joint is extended, the contact brush corresponds to the latitude near the equator of the sphere, and when the joint is flexed, the contact brush gradually shifts towards the polar direction of the sphere. Since different functional tracks (angle encoding track, zero position marking track, power supply track, signal track, etc.) are respectively arranged on different latitude circles, the contact brush sweeps over or stops on the corresponding functional track area in sequence as the latitude changes, thereby forming electrical contact with different types of tracks in the same rotation stroke, realizing multi-functional spatial partitioning coordination such as angle detection, zero position triggering, power supply conduction and signal transmission.

[0037] The insulating substrate 110 includes at least a spherical portion, specifically resembling a spherical crown or a pot shape. This portion can be one-quarter or one-eighth of the sphere. The insulating substrate 110 can be disposed on the rotating side of the finger joint, without occupying additional joint space during finger flexion and extension, thus occupying a small volume. Specifically, two quasi-spherical structures are provided at each finger joint, arranged concentrically and in the same direction. One is an inner quasi-spherical structure with conductive tracks on its surface, and the other is an outer quasi-spherical structure with conductive tracks on its surface. The conductive tracks on the inner and outer sides form a sliding electrical contact through an elastic contact brush. The contact brush is fixed to one of the quasi-spherical structures and maintains elastic sliding contact with the corresponding conductive track on the surface of the other quasi-spherical structure during joint rotation, thereby realizing continuous transmission of electrical signals. Taking the distal interphalangeal joint as an example, the distal finger is rigidly connected to the inner spherical slip ring structure, and the middle finger is rigidly connected to the outer spherical slip ring structure. When the finger is bent, the distal finger drives the inner spherical slip ring structure to rotate around the center, while the middle finger keeps the outer spherical slip ring structure fixed. Both rotate around the same center, and the elastic contact brush maintains sliding contact with the conductive tracks on the surfaces of the two spherical slip ring structures, thereby realizing the continuous transmission of electrical signals. The structure at the proximal interphalangeal joint of the middle finger is similar. The middle finger is rigidly connected to the inner spherical slip ring structure, and the proximal finger is rigidly connected to the outer spherical slip ring structure, and so on to each finger joint.

[0038] It should be noted that although a sphere or a portion thereof is the preferred shape, those skilled in the art will understand that the shape of the insulating substrate 110 is not limited to a sphere, and other arcuate shapes of revolution, such as ellipsoids, cylinders, frustums, or disks, can also be used, as long as their surfaces can provide a continuous curved surface for sliding contact of the contact assembly 130, and multiple conductive tracks can be arranged in sections on the surface. For cylindrical or disk-shaped insulating substrates, conductive tracks can be arranged along the circumferential direction or the end face direction, which can also realize the output of angle encoding and zero-position trigger electrical signals. In addition, "spherical" in this application should be understood to include arcuate conductive substrate structures in a broad sense, covering shapes of revolution such as spheres, ellipsoids, cylinders, and disks with arcuate conductive surfaces.

[0039] In some embodiments, such as Figure 3 As shown, the first type of track 121 is arranged circumferentially along the insulating substrate 110 and has multiple segments. The multiple segments include conductive segments 1211 and insulating segments 1212, which are alternately distributed circumferentially along the insulating substrate 110. The first contact portion 131 rotates with the insulating substrate 110 and sequentially sweeps across each segment. The first electrical signal is the on / off signal sequence generated when the first contact portion 131 contacts each segment.

[0040] Specifically, the first type of track 121 extends circumferentially (i.e., in the direction of joint rotation) on the surface of the insulating substrate 110, forming a ring or arc-shaped trajectory. This trajectory is divided into multiple continuous segments circumferentially, each segment being either a conductive segment 1211 (made of conductive material) or an insulating segment 1212 (made of insulating material or a blank gap), with the conductive and insulating segments 1211 alternating circumferentially. When the insulating substrate 110 rotates with the joint, the fixed first contact portion 131 sweeps across each segment circumferentially. When the first contact portion 131 passes through the conductive segment 1211, it makes contact with the conductive segment 1211, the circuit is connected, and a high-level signal (or a low-level signal, depending on the circuit design) is output; when the first contact portion 131 passes through the insulating segment 1212, it does not conduct, the circuit is broken, and a signal of the opposite level is output. Thus, as the joint rotation angle changes continuously, the first contact 131 outputs an on / off signal sequence consisting of alternating high and low levels. The encoding pattern of this on / off signal sequence changes with the angle, thereby representing the current rotation angle.

[0041] On / off signal sequences, as the physical layer representation of angle encoding, have advantages such as clear signal levels, strong noise immunity, and direct input to digital logic circuits for decoding. Compared with analog output angle sensors, on / off signals can be directly used for angle discrimination without analog-to-digital conversion, have fast response speed, and no quantization noise.

[0042] In some embodiments, the conductive segment 1211 may be made of a conductive material (such as copper, gold, or silver) and be coplanar with or slightly above the surface of the insulating substrate 110 to ensure reliable contact with the first contact portion 131. The insulating segment 1212 may be an exposed portion of the surface of the insulating substrate 110 (when the insulating substrate 110 itself is an insulating material) or may be formed by covering the conductive layer with an insulating coating.

[0043] It should be noted that the circumferential distribution pattern of the conductive segments 1211 and the insulating segments 1212 determines the angular encoding method. The more segments there are, the higher the angular resolution. The lengths of the segments can be equal or unequal, depending on the encoding strategy employed.

[0044] In some embodiments, the conductive segments 1211 and insulating segments 1212 are distributed circumferentially along the insulating substrate 110 according to Gray code encoding, and the on / off signal sequence is a Gray code encoded signal. Gray code, also known as cyclic code or reflected binary code, is a binary encoding method characterized by only one bit changing between two adjacent encoded values. In some embodiments, multiple conductive segments 1211 and insulating segments 1212 of the first type of track 121 are arranged circumferentially according to a Gray code encoding table. The combination of level states read at each contact point corresponding to different angular positions constitutes the Gray code encoded value.

[0045] When the joint rotates continuously from one angular position to an adjacent angular position, only one bit of the on / off signal sequence output by the first contact 131 changes, avoiding code ambiguity or race conditions that may be caused by multiple bits of the signal flipping simultaneously. In other words, at the boundary between any two adjacent angular intervals, only one segment undergoes a switching between conductive and insulating states, thereby eliminating the coding uncertainty that may be introduced by non-single-step changes.

[0046] When the Gray code has n bits, it corresponds to 2^n. n One encoded value, angular resolution = 360° / 2 n (For a full circle) or 90° / 2 n (For the flexion-extension range of 0° to 90°). In this application, the Gray code angle encoding track has 12 bits, corresponding to an angle resolution of approximately 0.022°, and an absolute angle measurement error ≤0.2°, meeting the requirements for high-precision joint angle detection. However, those skilled in the art will understand that the number of bits can be adjusted according to actual accuracy requirements, such as 8 bits, 10 bits, 12 bits or more, with different bit numbers corresponding to different angle resolutions and the number of track segments.

[0047] In some embodiments, the Gray code encoded signal is composed of the on / off states generated by the first contact portion 131 contacting each conductive segment 1211 (or insulating segment 1212) of the first type of track 121. This encoded signal can be directly input into the decoding circuit (such as the decoding circuit in the local processing circuit described later) to convert Gray code to binary code and obtain joint angle data.

[0048] In some embodiments, the second type of track 122 is disposed on the surface of the insulating substrate 110 in a region corresponding to a reference position, which is a preset angle position of the joint. Specifically, unlike the first type of track 121, the second type of track 122 is not continuously or segmented along the entire circumference, but is only partially disposed in a specific region on the surface of the insulating substrate 110. This specific region spatially corresponds to a preset angle position on the joint's movement trajectory. When the joint rotates to this preset angle position, the region of the second type of track 122 on the insulating substrate 110 rotates precisely to the position where the second contact portion 132 is located, so that the second contact portion 132 and the second type of track 122 form electrical contact.

[0049] The preferred reference position is the joint zero-angle position, which is the angle position of the joint when the finger is naturally extended. During data acquisition, the zero-angle is the starting reference point for joint movement, and the flexion and extension angles of the joint are measured with the zero position as the reference. Using the zero-angle as the hardware synchronization anchor point, synchronization can be automatically triggered every time the joint passes through the zero position, without the need for additional initialization or external trigger signals.

[0050] However, it should be noted that the reference position can also be set to any preset angle position that needs to be precisely marked, other than the zero angle, such as the position where the finger is bent at 30°, 45°, or 90°, depending on the needs of the application scenario. In rehabilitation assessment scenarios, a second type of track can be set at multiple preset angle positions to mark different key angles.

[0051] In some embodiments, the second type of track 122 is a bump conductor formed on the surface of the insulating substrate 110. The second contact portion 132 makes contact with the bump conductor when the joint rotates through a preset angle position of the joint to generate a second electrical signal with a transition edge based on the contact conduction.

[0052] Specifically, the bump conductor is a raised conductive structure formed in a localized area on the surface of the insulating substrate 110. Its material can be a good conductor such as copper, gold, or silver, and it is formed on the surface of the insulating substrate 110 through electroplating, welding, or embedding. The bump conductor has a slight height protrusion relative to the surface of the insulating substrate 110 to ensure reliable physical contact and electrical conduction between the second contact portion 132 and the bump conductor at the contact location.

[0053] The angular range occupied by the convex conductor in the circumferential direction corresponds to a very small angular interval, and the width of this angular interval determines the timing accuracy of the second electrical signal transition edge. In some embodiments, the conduction width of the convex conductor corresponds to an angle of 0.3°, that is, the second contact portion 132 and the convex conductor are in a conducting state only when the joint rotates through this 0.3° range, generating a steep transition edge at the instant of switching between conduction and cutoff.

[0054] The specific form of the transition edge can be a rising edge (transition from low to high) or a falling edge (transition from high to low), depending on the default level configuration of the circuit. In some embodiments, when the second contact 132 is not in contact with the bump conductor, the second contact 132 and the second type of track 122 are in an open circuit state, and the output at the second contact 132 is low (e.g., grounded through a pull-down resistor). When the joint rotates through a preset position, the second contact 132 makes contact with the bump conductor and conducts, and the level at the second contact 132 transitions from low to high, generating a rising edge transition signal. When the joint leaves the preset position, the second contact 132 disconnects from the bump conductor, and the level transitions from high to low, generating a falling edge transition signal. The rise time (or fall time) of the transition edge is determined by the physical process at the moment of contact conduction and is related to the contact material, contact pressure, and load circuit. In some embodiments, the rise time of the second electrical signal is ≤100ns, ensuring that the transition edge has sufficient time accuracy and can serve as a nanosecond-level hardware synchronization reference.

[0055] It should be noted that the convex conductor is only one implementation of the second type of track 122. In alternative embodiments, the second type of track 122 can also be a conductive sheet flush with the surface of the insulating substrate 110, a conductive post embedded in the surface of the insulating substrate 110, or a conductive ring segment, as long as it can form a reliable transient electrical contact with the second contact portion 132 at a preset angular position.

[0056] like Figure 3 As shown, in some embodiments, the conductive track 120 further includes a third type of track 123, which includes a positive power supply track and a common ground track, both of which are continuous annular conductive tracks used for transmitting electrical energy. Both the positive power supply track and the common ground track are annular tracks continuously arranged circumferentially along the insulating substrate 110, and are continuously conductive throughout the entire circumferential angle range. The positive power supply track is used to transmit the positive DC power supply (e.g., 3.3V), and the common ground track is used to provide a power circuit. Because they are continuous annular, regardless of the joint rotation angle, the positive power supply and the common ground can be continuously and stably powered through their corresponding contacts, achieving uninterrupted power transmission.

[0057] In some embodiments, both the positive power rail and the common ground rail can be made of low-resistivity conductive materials (such as copper or gold), with a contact resistance ≤20mΩ, to ensure power supply efficiency. For applications requiring power to the fingertip sensor, power can be transmitted step-by-step to the fingertip via sliding joints at each joint, achieving remote power supply.

[0058] In some embodiments, the conductive track 120 further includes a fourth type of track 124, which is a continuous annular conductive track used to transmit tactile analog signals. The fourth type of track 124 is also an annular track continuously arranged circumferentially along the insulating substrate 110, and is continuously conductive across the entire circumferential angle range. The number of fourth type tracks 124 can be set according to the channel requirements of the tactile sensor, for example, 2 to 4 tracks. The analog signal generated by the tactile sensor (such as a flexible piezoresistive array) on the fingertip side is transmitted through the fourth type of track 124 and the corresponding contact portion to the analog-to-digital converter (ADC) input terminal of the main control unit 200, eliminating the need for signal conditioning chips or ADCs at the joint nodes, thus achieving pure hardware transmission without intermediate chips.

[0059] The tactile analog signal is not processed by any active device, avoiding noise, offset and nonlinear distortion introduced during signal conditioning, and ensuring high-fidelity transmission of the tactile signal.

[0060] In some embodiments, the impedance matching of the fourth type track 124 is ≤50Ω, the signal bandwidth is ≥10kHz, and the crosstalk rejection ratio is ≥60dB, ensuring that the tactile analog signal does not undergo significant attenuation or distortion during transmission.

[0061] It should be noted that the first type of orbit 121, the second type of orbit 122, the third type of orbit 123, and the fourth type of orbit 124 are located on the surface of the insulating substrate 110 along the latitudinal direction (i.e., Figure 2 The tracks are arranged in layers (vertically as shown), with a certain distance between each track in the latitudinal direction, and are electrically isolated by insulating grooves or insulating gaps. The specific latitudinal position and width of each track can be designed according to the actual spatial layout and functional requirements.

[0062] In some embodiments, insulating grooves are provided between each conductive track to electrically isolate adjacent conductive tracks. The insulating groove is a physical isolation structure provided on the surface of the insulating substrate 110 between adjacent conductive tracks. The insulating groove can be in the form of a recess (i.e., a groove formed by etching or cutting on the surface of the insulating substrate 110), a raised insulating ridge (i.e., a raised insulating isolation strip provided between adjacent tracks), or an exposed insulating gap (i.e., a blank area of ​​a certain width left between adjacent tracks so that they are not directly connected).

[0063] The width of the insulating groove should be sufficient to ensure that no accidental electrical continuity occurs between adjacent conductive tracks under conditions of manufacturing tolerances, wear and tear during use, and environmental changes (such as temperature and humidity). In some embodiments, the width of the insulating groove is ≥0.2mm and the insulation resistance is >100MΩ, ensuring high-reliability electrical isolation between the conductive tracks.

[0064] When multiple types of conductive tracks are densely arranged on a limited area of ​​insulating substrate surface, the insulating groove ensures that there will be no leakage or short circuit between the tracks due to conductive material residue, particulate contamination or surface creepage, thereby ensuring the independence of each electrical function and signal integrity.

[0065] In a specific embodiment, such as Figure 4 As shown, this figure is a spherical unfolded view of the surface of the spherical insulating substrate 110, which fully shows the orbital arrangement relationship of the sphere in the 360° circumferential direction and the latitudinal direction. Figure 4 The horizontal arrangement of brushes 1, 2, 3, 4, 5, ..., 16, 17, 18, 19, and 20 at the top refers to the different angular positions of the same contact brush assembly, used to represent the sweeping trajectory of the contact brush at different circumferential positions on the spherical surface.

[0066] Specifically, Figure 4 The positive power supply track and the common ground track, corresponding to the positive power supply and ground respectively, belong to the third type of track 123, which are continuous circular conductive tracks. Signal track 1 and signal track 2 belong to the fourth type of track 124, which are tactile analog signal pass-through tracks, also complete and continuous circular, realizing chip-free pass-through of tactile analog signals. The 12-bit Gray code encoding area belongs to the first type of track 121, with conductive segments 1211 arranged alternately along the circumference 360° (i.e., Figure 4 The first type of track (black grid) and insulation segment 1212 (i.e. Figure 4 The white grid of the first type of track forms a Gray code on / off signal sequence, outputting the first electrical signal representing the joint rotation angle. The single-point protrusion (zero position) is for the second type of track 122, with a conductive protrusion only set at a single angular position in the circumferential direction. This position corresponds to the joint reference zero position. When the spherical insulating substrate rotates to the reference position, this set of elastic metal brushes instantaneously contacts and conducts with the single-point protrusion, generating a second electrical signal with a jumping edge, outputting a hardware zero-position synchronization event. Insulating grooves (i.e., ...) are set between the tracks. Figure 4 Insulating slots 1, 2, ..., 14, and 15 isolate adjacent tracks with different functions to prevent electrical crosstalk between the power rail, analog signal rail, Gray code encoding rail, and zero-position convex point rail.

[0067] In some embodiments, please refer to Figure 5The joint sliding electrical contact node device 100 also includes a local processing circuit 140, which includes a decoding circuit and / or a detection circuit. The decoding circuit is electrically connected to the contact component 130 and is used to decode the first electrical signal to obtain joint angle data. The detection circuit is electrically connected to the contact component 130 and is used to perform debouncing and edge extraction processing on the second electrical signal to output a detection signal with a transition edge.

[0068] The local processing circuit 140 is integrated on the back or near the mounting base of the joint sliding electrical contact node device 100, and its size is preferably ≤5mm×5mm to accommodate the narrow space of the finger joint. The local processing circuit 140 can be implemented using pure hardware logic (such as discrete components or programmable logic devices such as digital logic gates and flip-flops), without the need for software intervention, thereby ensuring the real-time performance and determinism of signal processing, and eliminating software operation delay.

[0069] The function of the decoding circuit is to decode the on / off signal sequence (Gray code encoded signal) output by the first contact 131 into angle values. The decoding circuit can use Gray code to binary code conversion logic, such as a Gray code decoder composed of an XOR gate array, to convert the Gray code encoded value into the corresponding binary angle value. The decoded joint angle data can be reported to the main control unit 200 via a cascaded bus.

[0070] The function of the detection circuit is to preprocess the second electrical signal output from the second contact 132 to output a detection signal with a clear rising edge. Specifically, the detection circuit includes a debouncing circuit and an edge extraction circuit. The debouncing circuit eliminates mechanical jitter or electrical oscillations (i.e., contact bounce) that may occur when the second electrical signal is turned on, ensuring a clean, single-rise signal output. The debouncing circuit can be implemented using a Schmitt trigger, an RC filter circuit, or digital debouncing logic (such as delayed sampling). The edge extraction circuit extracts the rising edge from the debouncing signal and outputs a detection signal with a clear rising or falling edge, whose rising edge is synchronized with the rising edge of the second electrical signal.

[0071] In some embodiments, the detection signal output by the detection circuit is sent to the synchronization trigger module of the main control unit 200 to generate a synchronization clock reference signal. Since the detection circuit is implemented using pure hardware logic, the output delay of the detection signal is extremely small (nanosecond level), ensuring the real-time performance of the synchronization trigger.

[0072] It should be noted that the local processing circuit 140 may include both a decoding circuit and a detection circuit, or it may include only one of them. If the node is only responsible for angle reporting and synchronous triggering is the responsibility of other nodes, then only a decoding circuit can be set up; otherwise, it is also possible.

[0073] In some embodiments, the local processing circuit 140 stores node hardware address codes for responding to polling in cascaded communication. The node hardware address code is a unique identifier (e.g., a 4-bit binary code) pre-assigned to each joint sliding electrical contact node device 100, used to distinguish nodes at different joints on the same finger (distal joint, middle joint, proximal joint) or nodes on different fingers. When the master control unit 200 polls each node via the cascade bus, each node identifies whether it is the target node addressed by the master control unit 200 based on its stored hardware address code. Only nodes with matching addresses will respond to bus commands and upload data.

[0074] The node hardware address code can be stored in a non-volatile memory (such as EEPROM or flash memory) in the local processing circuit 140 or set via a hardware jumper / DIP switch. In this embodiment, a 4-bit address code is used, which can support 16 nodes with different addresses, sufficient to cover the addressing needs of all finger joints of a single glove (5 fingers × 3 joints = 15 nodes).

[0075] In some embodiments, the contact component 130 includes multiple sets of elastic metal brushes, each set of elastic metal brushes making contact with a corresponding conductive track.

[0076] The flexible metal brush is a slender brush-like structure made of a flexible conductive material (such as beryllium copper, phosphor bronze, or stainless steel), and its surface can be plated with gold or silver to improve conductivity and wear resistance. Each set of flexible metal brushes corresponds to one conductive track, maintaining continuous elastic contact with the corresponding track surface. The number of flexible metal brushes corresponds to the total number of conductive tracks. For example, when the conductive tracks include a first type track 121, a second type track 122, a third type track 123 (two tracks: positive power and common ground), and a fourth type track 124 (one or more tracks), the number of flexible metal brush sets in the contact assembly 130 is 6 to 8 sets.

[0077] The contact pressure of the elastic metal brush (i.e., the positive force applied by the elastic metal brush to the track surface) is a key parameter affecting contact reliability. Insufficient contact pressure may lead to increased contact resistance or intermittent contact, while excessive contact pressure may accelerate wear on the track and contact brush. In some embodiments, the contact pressure of each group of elastic metal brushes is 10mN to 30mN. This pressure range ensures reliable contact while keeping wear within acceptable limits, thereby achieving a bending life of ≥1 million cycles. The contact resistance of the elastic metal brush (i.e., the contact interface resistance between the elastic metal brush and the track surface) should remain stable throughout the full angular rotation range. In some embodiments, the contact resistance fluctuation is ≤10mΩ, ensuring the stability of signal transmission.

[0078] It should be noted that although this embodiment uses an elastic metal brush as the contact part, those skilled in the art will understand that the contact part can also adopt other forms of sliding contact structure, such as elastic stylus, conductive rubber contact, carbon brush, etc., as long as it can maintain continuous electrical contact with the track surface and output electrical signals during the joint rotation.

[0079] In some embodiments, there are multiple joint sliding electrical contact node devices 100, which are respectively disposed at different joints of the same finger, and each joint sliding electrical contact node device 100 is cascaded in series through a set of bus harnesses.

[0080] For example, such as Figure 6 As shown, a joint sliding electrical contact node device 100 is set at the distal, middle, and proximal joints of a finger, namely the proximal node (ADDR=00), middle node (ADDR=01), and distal node (ADDR=10). The three nodes are serially cascaded through a set of 6-core bus harnesses, with the address increasing from the proximal end to the distal end of the finger (00→01→10), facilitating sequential polling by the main control unit 200. The starting point of the bus harness is the back-of-hand main control interface, which is located on the back of the hand and electrically connected to the main control unit 200. This interface is responsible for transmitting the power, communication, and analog signals of the main control unit 200 to each joint node through the bus harness. The 6-core harness is led out from the back-of-hand main control interface along the finger direction, passing sequentially through the proximal node (ADDR=00), middle node (ADDR=01), and distal node (ADDR=10).

[0081] The core wires of the bus harness specifically include: core wire 1 is the power supply line (VCC, 3.3V DC power supply), core wire 2 is the ground line (GND, common ground loop), core wires 3 and 4 are analog signal lines (Analog 1 and Analog 2, used to transmit analog signals from the fingertip tactile array), and core wires 5 and 6 are digital communication lines (communication + and communication -, used for half-duplex serial cascaded communication). Analog 1 and Analog 2 form a pair of analog signal transmission lines, used to transmit the tactile analog signals generated by the fingertip tactile sensor from the distal finger node to the proximal finger node step by step, and finally send them to the ADC unit of the main control unit 200; communication + and communication - form a pair of digital communication lines, used for half-duplex serial cascaded communication between the main control unit 200 and each node. This 6-core wire harness allows for simultaneous power transmission (core wires 1 and 2), analog signal pass-through (analog 1 and analog 2 of core wires 3 and 4), and digital communication (communication + and communication - of core wires 5 and 6) on the same harness, eliminating the need for separate wiring for each function. This 6-core wire harness corresponds to the physical structure of the aforementioned third-type rail 123 (power positive / common ground) and fourth-type rail 124 (analog signal pass-through). Core wires 1 and 2 correspond to the power positive rail and common ground rail, respectively; core wires 3 and 4 correspond to the analog signal pass-through rails (analog 1 and analog 2); and core wires 5 and 6 correspond to the digital communication lines (communication + and communication -). These digital communication lines are established through the communication interface in the local processing circuit 140 of each node (i.e.,...). Figure 6 The cascaded bus interface in the local processing circuit 140 is responsible for level conversion and signal transmission / reception. The cascaded bus interface is the interface module in the local processing circuit 140 responsible for bus communication with the main control unit 200. It contains a bus transceiver, level conversion circuit and protocol parsing logic, which is used to convert differential signals on the bus into digital level signals that can be processed inside the node, and to encode the data to be sent by the node into bus signals for external transmission.

[0082] Each node is equipped with a local processing circuit 140, which includes angle decoding and zero-position detection. The angle decoding function is implemented by a decoding circuit within the local processing circuit 140, which is electrically connected to the contact component 130. This decoding circuit decodes the first electrical signal (Gray code encoded signal) generated by the contact between the first type of track 121 and the first contact part 131 from Gray code to binary code, obtaining the current joint rotation angle data. The zero-position detection function is implemented by a detection circuit within the local processing circuit 140, which is electrically connected to the contact component 130. This detection circuit performs debouncing and edge extraction processing on the second electrical signal generated by the contact between the second type of track 122 and the second contact part 132 at the reference position, obtaining a detection signal with a clear transition edge. This detection signal is used to characterize whether the joint has passed the reference position. When responding to polling by the main control unit 200, the node replies to the main control unit 200 via a cascaded bus interface with the decoded angle data and the zero-position status data output by the zero-position detection circuit. A fingertip tactile sensor is located at the fingertip of the finger and is a flexible pressure-controlled array (i.e.,...). Figure 6 The flexible piezoresistive array (in the design) consists of multiple piezoresistive sensing units. Each sensing unit changes its resistance when subjected to pressure, generating an analog voltage signal that varies with pressure through resistive voltage division or constant current source excitation. This analog voltage signal is directly transmitted through core wire 3 (analog 1) and core wire 4 (analog 2), bypassing local processing circuits at each node, and directly reaching the input of the ADC unit of the main control unit 200 in a chip-free transmission manner. Analog 1 and Analog 2 can transmit analog signals from different rows or channels of the fingertip pressure control array, for example, Analog 1 transmits the signal from the first half of the fingertip sensor array, and Analog 2 transmits the signal from the second half, or Analog 1 and Analog 2 can be transmitted differentially to improve anti-interference capability. Since the signal does not undergo any active devices or logic processing, noise, offset, and nonlinear distortion introduced by intermediate links are avoided, ensuring high-fidelity transmission of the tactile signal.

[0083] The cascaded communication employs a master-slave polling mechanism, with the master control unit 200 acting as the master and each joint sliding electrical contact node device 100 acting as a slave. Each node stores a unique hardware address code; specifically, the address code for the proximal finger node is ADDR=00, the middle finger node is ADDR=01, and the distal finger node is ADDR=10. The master control unit 200 sends a polling frame containing the target address via digital communication lines (communication + of core wire 5 and communication - of core wire 6). The polling frame is transmitted cascaded through the cascaded link to each node. Upon receiving the polling frame, each node compares the target address in the frame with its own stored hardware address code. If they match, the node replies with its own data (angle data and zero-position status) to the master control unit 200 via the cascaded bus interface within a predetermined time slot. If they do not match, the node ignores the frame and continues to wait for the next frame.

[0084] The bus harness extends along the fingers in a cascaded manner. Specifically, the back-of-hand main control interface is connected to the proximal finger node (ADDR=00) via a set of 6-core harnesses. The proximal finger node is connected to the middle finger node (ADDR=01) via another set of 6-core harnesses. The middle finger node is then connected to the distal finger node (ADDR=10) via a set of 6-core harnesses. The 6-core harnesses at each node physically pass through or bypass the node's mounting bracket. Core wire 1 (VCC) and core wire 2 (GND) need to be connected to the node's local processing circuit at each node to provide operating power. Core wire 5 (communication+) and core wire 6 (communication-) need to be connected to the cascaded bus interface in the local processing circuit at each node to realize the transmission and reception control of cascaded communication. Core wire 3 (analog 1) and core wire 4 (analog 2) are directly bypassed at the proximal and middle finger nodes (i.e., directly bridging at the physical layer without passing through the local processing circuit), and only connected to the fingertip flexible pressure control array at the distal finger node to realize the chip-free transmission of analog signals from the fingertip to the main control.

[0085] The single-finger three-node design uses a set of 6-core wire harnesses cascaded in series. Compared to the traditional solution where each joint has its own independent wiring (12 to 16 wire harnesses per finger), the number of wire harnesses is reduced by more than 65%, significantly reducing the complexity and density of internal wiring. At the same time, the electrical connection at each joint is achieved through sliding contact, and the bus harness itself does not need to bend with the joint, eliminating the risk of wire harness fatigue breakage.

[0086] It should be noted that although this embodiment uses three joints of a single finger as an example, those skilled in the art will understand that this cascaded topology is also applicable to more fingers. The bus of each finger can be connected to different interfaces of the main control unit 200, or more nodes can be connected on the same bus through address expansion. The hardware address encoding of each node can be assigned according to the position of the node on the finger and the finger to which it belongs, and is not limited to the specific values ​​in the above example.

[0087] In some embodiments, the main control unit 200 includes a synchronization trigger module, which includes a cascaded first register and a second register. The data input terminal of the first register is used to receive a detection signal, the clock input terminal of the first register is used to connect to the system clock, the data output terminal of the first register is connected to the data input terminal of the second register, the clock input terminal of the second register is used to connect to the system clock, and the data output terminal of the second register is used to output a synchronization clock reference signal.

[0088] The synchronization trigger module is a circuit module used to process the zero-position detection signal and generate a hardware synchronization reference. As mentioned earlier, the detection circuit in the local processing circuit 140 debouncing and edge-extracting the second electrical signal outputs a detection signal with a rising edge. This detection signal is an asynchronous signal and belongs to a different clock domain from the system clock inside the main control unit 200 (i.e., there is no fixed phase relationship between them). When the asynchronous signal is directly input to the synchronization circuit, it may not meet the setup and hold time constraints of the trigger, causing the trigger to enter a metastable state (i.e., the output level is uncertain, oscillating, or has an excessively long delay), thereby affecting the timing determinism of the synchronization clock reference signal.

[0089] To address this issue, this application employs a two-stage register cascading (i.e., a cascaded first and second register) to eliminate the risk of metastability across clock domains. The first register receives the detection signal and is sampled by the system clock. If the transition edge of the detection signal occurs within the setup time window of the first register, the output of the first register may exhibit metastability. The second register resamples the output of the first register. Since the metastability usually stabilizes within one clock cycle, the output of the second register essentially eliminates the metastability, outputting a clean transition signal synchronized with the system clock.

[0090] Furthermore, in some embodiments, both the first register and the second register are D flip-flops (data flip-flops). The data input terminal (D terminal) of the first register receives the detection signal, the clock input terminal (CLK terminal) is connected to the system clock, and the data output terminal (Q terminal) is connected to the data input terminal (D terminal) of the second register. The clock input terminal (CLK terminal) of the second register is connected to the same system clock, and the data output terminal (Q terminal) outputs the signal after synchronization processing, i.e., the synchronization clock reference signal.

[0091] After two-stage register synchronization, the jitter of the synchronization clock reference signal is ≤50ns, ensuring the timing accuracy of synchronous triggering. Meanwhile, the synchronization triggering module is implemented using pure hardware logic, with no software scheduling delay, and the total delay of the triggering link is ≤1us.

[0092] It should be noted that although the embodiments of this application use two-stage register pacing to achieve metastable synchronization, those skilled in the art will understand that single-stage register synchronization can also be used in scenarios with lower system clock frequencies or lower metastability risks; and in scenarios requiring higher reliability, three-stage or more register synchronization can be used. Two-stage registers are a preferred solution that balances reliability and latency.

[0093] Furthermore, the synchronous trigger module is not limited to a two-level register implementation. In an alternative embodiment, the synchronous trigger module can also be constructed from combinational logic circuitry to transmit the detection signal from the signal input terminal to the signal output terminal and output it as a synchronous clock reference signal. This combinational logic circuitry can be constructed from digital logic gates such as buffers, inverters, and AND gates, and its main function is to provide signal driving and shaping, without involving clock synchronization processing.

[0094] In some embodiments, the main control unit further includes a data acquisition device comprising multiple acquisition units of different types; the synchronization trigger module outputs multiple synchronous trigger signals of the same source based on a synchronization clock reference signal, which are used to trigger different types of acquisition units to perform synchronous sampling. "Multiple sources of the same source" means that all multiple synchronous trigger signals are generated from the same synchronization clock reference signal, i.e., they share the same trigger source (the second electrical signal generated by the contact and conduction between the second type of track and the second contact part). Since each trigger signal originates from the same source, the time deviation between them is extremely small, determined by the signal allocation delay (gate delay) within the synchronization trigger module, typically less than nanoseconds.

[0095] The synchronous trigger module can use a fan-out buffer to copy the synchronous clock reference signal into multiple outputs, or output multiple synchronous trigger signals through a clock distribution network. Each synchronous trigger signal has the same timing relationship (i.e., their transition edges are aligned at the same time), only the output ports are different, and they are connected to different types of acquisition units.

[0096] Data acquisition equipment refers to other sensor devices in the system besides the joint sliding electrical contact node device 100, including but not limited to tactile sensing units, inertial measurement units, and vision units. These sensor devices typically have their own independent sampling clocks and sampling periods, operating freely at their respective frequencies in the absence of external synchronization signals. When a synchronization trigger signal is received, these sensor devices respond to the rising edge of the synchronization trigger signal, immediately performing sampling operations and latching the sampled data at the current moment, thereby achieving synchronous sampling by multiple sensors at the same time.

[0097] It should be noted that "different types of acquisition units" refers to sensor units that perform measurements of different physical quantities. For example, a tactile sensing unit measures contact force / pressure distribution, an inertial measurement unit measures acceleration and angular velocity, and a vision unit measures image information. Different types of acquisition units do not have a fixed sampling frequency synchronization relationship; therefore, an external synchronization trigger signal is needed to achieve time alignment between them.

[0098] In some embodiments, the multi-source synchronous trigger signal includes at least two of the following: a first synchronous trigger signal for triggering the analog-to-digital converter of the tactile sensing unit to latch the current sampled value; a second synchronous trigger signal for triggering the inertial measurement unit to latch the current attitude data; and a third synchronous trigger signal for triggering the vision unit to expose.

[0099] The first synchronous trigger signal is connected to the sample-and-hold trigger input of the analog-to-digital converter (ADC) of the haptic sensing unit. The haptic sensing unit typically includes a flexible piezoresistive array sensor and an ADC. Without a trigger, the ADC continuously samples at a fixed sampling rate. When the first synchronous trigger signal arrives, the ADC's sample-and-hold circuit latches the current analog input value at the trigger edge, and then performs analog-to-digital conversion, ensuring that the timestamp of the frame of haptic data strictly corresponds to the trigger moment. The latching delay is ≤200ns.

[0100] The second synchronization trigger signal is connected to the synchronization trigger pin of the inertial measurement unit (IMU). The IMU typically includes a three-axis accelerometer and a three-axis gyroscope, and has internal data registers to store the latest measurement values. When the second synchronization trigger signal arrives, the IMU latches the current acceleration and angular velocity data into the register at the trigger edge for subsequent reading. The latch response time is ≤1µs.

[0101] The third synchronous trigger signal is connected to the external trigger exposure pin of the vision unit (binocular camera). Miniature cameras typically support hardware-triggered exposure mode; when the external trigger signal arrives, the camera initiates the exposure process. The third synchronous trigger signal triggers the binocular camera's exposure start signal, with a trigger delay ≤500µs.

[0102] Since the three synchronous trigger signals originate from the same synchronous clock reference signal, they share a common trigger time reference. Therefore, at this trigger time, the tactile data, inertial data, and visual data are in one-to-one temporal correspondence, forming a set of time-aligned multimodal data frames. The time deviation of the three modes within a single frame is ≤0.5ms. This time alignment is directly implemented by the hardware synchronization link, without involving software timestamp matching or post-processing interpolation, thus avoiding the timing uncertainties in software alignment schemes.

[0103] It should be noted that among the three synchronous trigger signals mentioned above, only any two can be selected, or all three can be selected simultaneously, depending on the combination of sensor types required for synchronous acquisition in the application scenario.

[0104] In some embodiments, the main control unit 200 further includes a bus arbitration module for allocating three-level priority time slots for communication of the bus harness. Specifically: the highest priority is allocated to synchronization events and / or joint angle mutation events corresponding to the synchronization clock reference signal, preemptively occupying the bus; medium priority is allocated to periodically sampled joint angle data and tactile data, allocating fixed time slots for transmission; and low priority is allocated to visual image data, utilizing idle time slots on the bus for fragmented transmission.

[0105] In cascaded bus communication, different types of data have different real-time requirements and transmission bandwidth demands. Synchronization events and angle change events are critical real-time events, requiring extremely low end-to-end latency; periodically sampled data requires stable transmission bandwidth to avoid data loss; image data has a large data volume but lower real-time requirements and can be transmitted during idle periods. To accommodate the transmission needs of different data types, this implementation adopts a priority time slot scheduling mechanism.

[0106] The highest priority (P0) is allocated to synchronization events (i.e., the reference position event corresponding to the second electrical signal) and joint angle mutation events (i.e., events where the rate of angle change exceeds a preset threshold). Both types of events are random, real-time events that need to be reported to the main control unit 200 as soon as possible to trigger corresponding synchronization operations or motion tracking responses. When a P0 event occurs, the bus arbitration module immediately interrupts the currently ongoing transmission and hands over bus control to the data frame corresponding to the P0 event, thus preemptively occupying the bus. The P0 event does not need to wait for the current frame to complete transmission; the end-to-end delay is ≤5ms.

[0107] Medium priority (P1) is allocated to periodically sampled joint angle data and tactile data. This data is continuously generated at a fixed frequency and requires a fixed transmission time slot to ensure that data from each sampling period is transmitted before the next sampling period arrives, avoiding data overwrite and loss. P1 data is transmitted within a fixed time slot window and is unaffected by P2 data. However, P1 transmission may be preempted by P0 events.

[0108] Low priority (P2) data is allocated to visual image data. Image data is large (a single frame can reach hundreds of KB to several MB), but its real-time requirements are lower than those of P0 and P1 data. P2 data is transmitted in fragments only during bus idle time slots (i.e., time slots where there is no P0 and P1 data transmission), without blocking high-priority data. When P0 or P1 data needs to be transmitted, P2 transmission is immediately paused and resumes when the bus becomes idle.

[0109] The three-level priority scheduling mechanism is implemented in pure hardware by the bus arbitration module in the main control unit 200, without the need for software involvement in arbitration decisions, thus ensuring the real-time performance and determinism of scheduling. Through this scheduling mechanism, the overall bus utilization rate is improved by more than 35% compared to a static allocation scheme without priority scheduling.

[0110] In some embodiments, the main control unit 200 further includes a clock compensation module, which is used to perform linear drift calibration and interval interpolation compensation on the local clock of each acquisition unit in the data acquisition device using two adjacent second electrical signals as hardware references, so as to correct the timestamp deviation of each acquisition unit between two adjacent hardware references.

[0111] As mentioned earlier, each acquisition unit (including tactile sensing unit, inertial measurement unit, and vision unit) has its own independent local clock crystal oscillator. Due to manufacturing tolerances, temperature drift, and aging effects of the crystal oscillator frequency, there is a cumulative time deviation between the local clocks and between the local clock and the global clock. Even if the sampling time is aligned through hardware synchronization trigger signals, each device will still sample independently at its own clock frequency during the period between two triggers, resulting in sampling time drift.

[0112] The clock compensation module addresses this issue by using two consecutive second electrical signals (i.e., the time points when the joint passes through the reference position twice) as two hardware reference times. Between these two hardware reference times, the sampling time of each acquisition unit is determined by its respective local clock. The global time of the two consecutive reference times (provided by the system clock of the main control unit 200) and the local clock readings of each acquisition unit at these two reference times are recorded.

[0113] Assuming the local clock drift of each acquisition unit is linear between two adjacent reference times, the drift rate (i.e., clock correction factor) of each acquisition unit relative to the global clock can be calculated. Substituting the local timestamp of each frame of sampled data from each device within the interval into the linear interpolation formula, it can be mapped to a unified global timestamp, thereby correcting the timestamp deviation between two adjacent hardware references for each acquisition unit. After compensation, the data synchronization error across the entire range is ≤1ms.

[0114] In one specific embodiment, the glove, the joint sliding electrical contact node device 100, the data acquisition device, and the main control unit 200 work together to complete the hardware-level synchronous acquisition of tactile, posture, and visual multimodal data. Specifically, the glove includes a wearable glove body, which serves as the physical support structure for each functional module. It is made of flexible fabric or elastic material and can be worn on the operator's hand. The glove body has mounting bases at the corresponding finger joints to fix the joint sliding electrical contact node device 100; tactile sensing units from the data acquisition device are located at the fingertips and fingertips; and the main control unit 200 is located on the back of the hand.

[0115] Please see Figure 7The diagram sequentially shows the proximal finger node (power / signal / encoding / zero position), middle finger node (power / signal / encoding / zero position), and distal finger node (power / signal / encoding / zero position) along the finger link. These three nodes correspond to three sliding electrical contact node devices 100, located at the proximal, middle, and distal finger joints, respectively. Each sliding electrical contact node device 100 simultaneously performs four functions on the same physical interface: power transmission (via a third type of track), tactile analog signal transmission (via a fourth type of track), angle encoding reading (via a first type of track), and zero-position synchronous triggering (via a second type of track). The sliding electrical contact node devices 100 are serially cascaded via a 6-core cascade bus (i.e., bus harness). The node hardware address code is pre-set in the local processing circuit 140 of each sliding electrical contact node device 100, used to respond to polling by the main control unit 200 during cascaded communication.

[0116] The data acquisition device includes a tactile sensing unit, which comprises a fingertip tactile sensor (analog output) positioned at the fingertip and a fingertip tactile sensor (analog output) positioned at the fingertip. Both are flexible piezoresistive arrays (the source of the tactile analog signal), consisting of multiple flexible piezoresistive array sensors. When pressure is applied, their resistance changes accordingly, generating an analog voltage signal that varies with pressure through resistive voltage division or constant current source excitation. This analog signal is transmitted without a chip to the analog-to-digital converter (e.g., ADC acquisition module) of the tactile sensing unit via the corresponding analog signal line (used for transmitting the tactile analog signal) in a 6-pin cascaded bus. When the ADC acquisition module receives the first synchronization trigger signal from the zero-position synchronization trigger module, it latches the current analog input value on the trigger edge and performs analog-to-digital conversion, converting the analog tactile quantity into a digital quantity.

[0117] The data acquisition device also includes an inertial measurement unit (such as an IMU attitude unit), which is a six-axis IMU sensor (integrating a three-axis accelerometer and a three-axis gyroscope) used to collect acceleration and angular velocity data of the hand, output hand attitude information, support hardware-triggered latching, and have a latching response time ≤1μs.

[0118] The data acquisition device also includes a vision unit (such as a binocular miniature camera connected to a binocular vision interface), which is connected to the main control unit through a MIPI interface, supports hardware-triggered exposure, has a trigger latency of ≤500μs, and a frame rate of 30~60fps, and is used to acquire binocular image information of hand operation scenes or hand postures (i.e., vision unit exposure).

[0119] The 6-core cascaded bus (i.e., bus harness) serves as the communication and power supply harness carrier connecting the main control unit and the joint sliding electrical contact node device 100, simultaneously carrying three types of signals: power supply, tactile analog signals, and digital communication signals. The bus interface module (i.e....) Figure 7The cascaded bus interface is located inside the main control unit and serves as the port module for communication and interaction between the main control unit and the sliding electrical contact node devices 100 of each joint. It is responsible for polling and collecting angle data of each joint node through the 6-core cascaded bus and transmitting the electrical energy output by the power management unit to each node through the bus.

[0120] The bus is a high-speed interconnect bus inside the main control unit, serving as a data interaction channel between various hardware modules within the main control unit, enabling data transmission and reception between the cascaded bus interface, ADC acquisition module, IMU attitude unit, binocular vision interface, processor, and MCU unit.

[0121] The processor is the core of the main control unit, comprising a synchronization trigger module, a clock compensation module, and a bus arbitration module, each implementing its function through pure hardware logic or a hardware-software co-processing approach. The synchronization trigger module receives the second electrical signal output from the joint sliding electrical contact node device 100, performs metastable synchronization processing in two levels of registers (i.e., cascaded first and second registers), generates a synchronous clock reference signal, and then outputs multiple synchronous trigger signals of the same origin through the parallel synchronization trigger signal lines of the hardware trigger chain. These signals are sent to the ADC acquisition module, the IMU attitude unit, and the binocular vision interface, respectively, to achieve hardware-level synchronous sampling of touch, attitude, and vision. The clock compensation module uses two adjacent second electrical signals as hardware references to perform linear drift calibration and interval interpolation compensation on the local clock of each acquisition unit, correcting the timestamp deviation between two adjacent hardware references for each acquisition unit. The bus arbitration module is used to schedule the communication time slots of the 6-core cascaded bus and the main control internal bus, and to perform three-level priority time slot allocation to ensure that synchronization events and joint change signals are transmitted first. That is, the highest priority is allocated to the synchronization events and / or joint angle change events corresponding to the synchronization trigger signal, and the bus is occupied in a preemptive manner.

[0122] The MCU (Microcontroller Unit) is the core control chip of the main control unit. It is electrically connected to the processor and the aforementioned modules, and undertakes tasks such as upper-layer business logic, data frame packaging, and protocol stack processing. The MCU unit receives multimodal acquisition data (tactile data, joint angle data, IMU data, and image data) output by the processor, summarizes and packages the data according to a preset frame format, and then transmits the packaged multimodal data to the host computer / data terminal through the connection cable.

[0123] The host computer / data terminal is a backend device that receives and processes multimodal data output from the data acquisition system, including but not limited to industrial robot controllers, embodied intelligent model training servers, rehabilitation assessment workstations, or VR / AR hosts. The host computer / data terminal is connected to the main control unit via wired or wireless communication and is used to receive data frames and perform subsequent processing or display.

[0124] The tactile sensing unit, inertial measurement unit, and vision unit achieve synchronous sampling under the control of three hardware trigger chains, ensuring the temporal consistency of multimodal data. After clock compensation and inherent delay compensation, the synchronization error is ≤1ms across the entire measurement range.

[0125] As another technical solution, this application embodiment also provides a multimodal synchronous data acquisition method, which can be executed based on the multimodal data acquisition system in the above system implementation.

[0126] In some embodiments, please refer to Figure 8 The multimodal synchronous data acquisition method includes: S101. Obtain the zero-position electrical signal generated by the joint sliding electrical contact node device set at the finger joint when the joint rotates through the reference position.

[0127] As described above, the joint sliding electrical contact node device 100 includes an insulating substrate 110, a conductive track 120, and a contact assembly 130. When the joint rotates through a reference position (e.g., a zero-angle position), the second track 122 makes contact with the second contact portion 132, generating a transient jump signal, which is the zero-position electrical signal. This zero-position electrical signal is debounced and edge extracted by the detection circuit in the local processing circuit 140, and then outputs a detection signal with a jump edge, which is transmitted to the main control unit 200.

[0128] In some embodiments, "acquisition" refers to the process by which the main control unit 200 receives a zero-position electrical signal through its input interface. The input interface of the main control unit 200 can be configured in an edge-triggered mode, that is, when a transition edge (rising edge or falling edge) is detected on the input signal line, the event is immediately recorded and subsequent processing is triggered.

[0129] S102. Generate a synchronous clock reference signal based on the zero-position electrical signal.

[0130] After receiving the zero-position electrical signal (the detection signal after preprocessing by the local processing circuit 140), the main control unit 200 performs synchronous processing through two-stage registers in the synchronous trigger module to generate a synchronous clock reference signal synchronized with the system clock. This synchronous clock reference signal has a deterministic timing relationship with the transition edge of the zero-position electrical signal and can be used as a hardware reference for subsequent synchronization operations.

[0131] S103. Control the data acquisition equipment to perform synchronous sampling according to the synchronous clock reference signal.

[0132] The main control unit 200, based on a synchronous clock reference signal, outputs multiple synchronous trigger signals (such as a first synchronous trigger signal, a second synchronous trigger signal, and a third synchronous trigger signal) through a synchronous trigger module, which are then transmitted to each acquisition unit. Each acquisition unit, in response to the received synchronous trigger signal, performs a sampling operation and latches the data at the current moment.

[0133] The data acquisition equipment includes multiple different types of acquisition units, specifically including two or more of the following: tactile sensing units, inertial measurement units, and vision units. That is, the system can select any two or three of these devices for simultaneous data acquisition, depending on the application scenario.

[0134] The zero-position electrical signal is directly generated by hardware (the contact between the second type of track and the second contact part) when the joint passes through the reference position during physical movement, rather than being generated by software timers or operating system scheduling. This zero-position electrical signal is processed by a pure hardware link to generate a synchronous clock reference signal, which is then used to trigger synchronous sampling by each acquisition unit. The entire process, from signal generation to trigger output, is implemented at the hardware level, with no software scheduling delay, ensuring high accuracy and low jitter in synchronous sampling.

[0135] In some embodiments, the reference position is the zero-angle position of the joint, that is, the angular position of the joint when the finger is fully extended. The zero-angle position is the natural starting point of joint movement, and using it as a synchronization reference has universal applicability. However, the reference position can also be set to other preset angular positions, not limited to zero.

[0136] In some embodiments, the data acquisition device includes at least two of a tactile sensing unit, an inertial measurement unit, and a vision unit. This means that the synchronization method can flexibly adapt to different multimodal combinations: in scenarios requiring tactile and posture synchronization, a tactile sensing unit and an inertial measurement unit can be selected; in scenarios requiring tactile and visual synchronization, a tactile sensing unit and a vision unit can be selected; in scenarios requiring posture and visual synchronization, an inertial measurement unit and a vision unit can be selected; and in all-modal scenarios, all three can be selected simultaneously. This embodiment does not limit the type and number of data acquisition devices; this method can be applied as long as the system is configured with two or more devices.

[0137] In some embodiments, the joint sliding electrical contact node device 100 includes an insulating substrate 110, a conductive track 120 disposed on the surface of the insulating substrate 110, and a contact assembly 130. The conductive track 120 includes a second type track 122. The contact assembly 130 is configured to output a second electrical signal characterizing a reference position event when the joint rotates through a reference position. The zero-position electrical signal is the second electrical signal generated when the second type track 122 and the contact assembly 130 make contact and conduct when the joint rotates through the reference position.

[0138] The zero-position electrical signal, at the physical level, corresponds to the transient electrical signal generated when the second type of track 122 and the second contact part 132 make contact and conduct at the reference position. This signal is a hardware-generated transition signal, and the rise time of its transition edge is determined by the physical process of contact conduction (≤100ns). This signal is the original electrical characterization of the zero-position event, and subsequent edge detection, two-stage register synchronization, and other processing are all based on this original signal.

[0139] In some embodiments, the second type of track 122 is disposed in a local area on the surface of the insulating substrate 110 corresponding to the reference position, preferably a convex conductive structure. When the joint rotates through the reference position, the second contact portion 132 makes instantaneous contact with the convex conductive structure to generate the zero-position electrical signal.

[0140] In some embodiments, the method further includes: outputting multiple synchronous trigger signals based on a synchronous clock reference signal, wherein the synchronous trigger signals include at least two of the following: a first synchronous trigger signal for triggering the analog-to-digital converter of the tactile sensing unit to latch the current sampled value; a second synchronous trigger signal for triggering the inertial measurement unit to latch the current attitude data; and a third synchronous trigger signal for triggering the vision unit to expose.

[0141] The synchronization trigger module in the main control unit 200 replicates the synchronization clock reference signal into multiple outputs through a fan-out buffer or clock distribution network. Each synchronization trigger signal corresponds to a type of acquisition unit. All synchronization trigger signals have the same timing source, therefore they are synchronized.

[0142] The first synchronization trigger signal is connected to the sample-and-hold trigger terminal of the analog-to-digital converter (ADC) of the haptic sensing unit. When the first synchronization trigger signal arrives, the ADC latches the analog input value at the current moment and starts the conversion, thus ensuring that the haptic data strictly corresponds to the trigger moment. The second synchronization trigger signal is connected to the external synchronization trigger pin of the inertial measurement unit (IMU). When the second synchronization trigger signal arrives, the IMU latches the acceleration and angular velocity data into its internal register at the current moment, ensuring that the attitude data strictly corresponds to the trigger moment. The third synchronization trigger signal is connected to the external trigger exposure pin of the vision unit. When the third synchronization trigger signal arrives, the camera starts the exposure at the current moment, ensuring that the image acquisition moment strictly corresponds to the trigger moment.

[0143] In practical applications, any two or all three channels can be output depending on the system configuration. For example, in a scenario where only haptic and inertial measurement are configured, only the first and second synchronous trigger signals can be output.

[0144] In some embodiments, the method further includes the following steps: S104. Record the number of sampling frames and the local clock time between two adjacent zero-position electrical signals for each acquisition unit.

[0145] The zero-position electrical signal is generated by hardware when the joint passes through the reference position. The time interval between two consecutive zero-position electrical signals is defined as follows: Within this interval, each acquisition unit (tactile sensing unit, inertial measurement unit, and vision unit) continuously samples and generates data frames according to its respective local clock frequency. The main control unit 200 records the number of sampling frames generated by each acquisition unit within this interval and the local clock time corresponding to each frame of data.

[0146] S105. Using the generation times of two consecutive zero-position electrical signals as hardware references, perform linear drift calibration on the local clock of each acquisition unit.

[0147] Because the local clock crystals of each acquisition unit have frequency deviations and drift, there is a cumulative deviation between their local clock times and the global clock times of the main control unit 200. Using the generation times of two consecutive zero-position electrical signals and the global time recorded by the main control unit 200 as two reference points, the drift rate of each acquisition unit's clock relative to the global clock can be calculated by comparing the global time difference with the local time difference.

[0148] S106. Based on the calibrated local clock time, map the sampling data of each acquisition unit between two adjacent hardware references to a unified global timestamp.

[0149] The clock interpolation compensation method in this application uses the zero-position electrical signal generated by the hardware each time the joint passes through the reference position as the calibration reference point to linearly correct the local clock drift between the two references. Since the reference position event occurs naturally each time the joint passes through the zero position, the calibration is performed continuously and automatically without manual intervention or external triggering. The compensation process does not involve operating system scheduling or software interrupts, ensuring the real-time nature and determinism of the compensation.

[0150] In some embodiments, the step of performing linear drift calibration on the local clock of each acquisition unit can be performed by calculating the clock correction coefficient of the i-th type of data acquisition device according to the following formula: , Among them, T k and T k+1 These are the global timestamps of two consecutive zero-position electrical signals, t and t'. i,k and t i,k+1 These are the local clock times of the i-th type of data acquisition device during two consecutive zero-position electrical signals.

[0151] The clock correction factor is stored in a register and used to timestamp the sampled data frames of each acquisition unit.

[0152] Specifically, the clock correction factor k i This represents the offset coefficient of the local clock frequency of the i-th type of data acquisition device relative to the global clock frequency between two adjacent reference events. When k... i When k = 1, it indicates that the local clock is synchronized with the global clock; when k i When k > 1, it indicates that the local clock is ahead; when k i When the value is less than 1, it indicates that the local clock is running slow.

[0153] For the nth frame of sampled data within the interval, its local time is t. i,n Then it is mapped to the global timestamp T. i,n It can be obtained through the following relation: This mapping linearly maps the local timestamp of each frame of data within the interval to the global timestamp, thereby eliminating the cumulative deviation caused by local clock drift.

[0154] In some embodiments, the clock correction factor k i The data is stored in the registers of the main control unit 200 in fixed-point or floating-point format to facilitate the timestamp mapping calculation of subsequent sampled data frames.

[0155] In some embodiments, the method further includes: S107. When a sudden change in the zero-position electrical signal or joint angle is detected, the corresponding synchronization event or angle change event will be transmitted to the bus with the highest priority. S108. The periodically sampled joint angle data and tactile data are allocated to fixed time slots for transmission; S109. The visual image data is transmitted in segments using the idle time slots of the bus.

[0156] In master-slave polling communication mode, the efficiency and real-time performance of bus communication are key considerations. Different types of data have different transmission priority requirements. Specifically, the "synchronization event" refers to the reference position event corresponding to the zero-position electrical signal generated by the contact and conduction between the second type of track and the second contact part when the joint passes the reference position. This event needs to be immediately reported to the master control unit 200 to trigger synchronization sampling, and therefore has the highest priority.

[0157] Angle abrupt events refer to events in which the joint angle changes rapidly. These events can be identified by the angle change rate (i.e., angular velocity) exceeding a preset threshold. Angle abrupt events may correspond to rapid grasping or releasing movements of the fingers and require timely reporting for real-time response. Therefore, angle abrupt events, along with synchronization events, belong to the highest priority category.

[0158] When any of the above events is detected, the bus arbitration module immediately preemptively occupies the bus and transmits the corresponding event data frame to the master control unit 200 without waiting for the current bus frame transmission to complete.

[0159] The periodically sampled joint angle and tactile data are continuously generated and updated at a fixed frequency. While these data do not require microsecond-level real-time response, they need to be transmitted before the next sampling period to avoid data overwriting. Therefore, fixed time slots are allocated for transmission to ensure at least one transmission opportunity within each sampling period.

[0160] Visual image data is large in volume, but its update frequency is relatively low, and its real-time requirements are lower than those for critical events and periodic data. Therefore, image data is transmitted in segments using bus idle time slots. That is, one image segment is transmitted at a time, and multiple segments are completed successively in multiple idle time slots until the entire image is transmitted. The granularity of segment transmission can be dynamically adjusted according to the bus bandwidth and the duration of available idle time slots.

[0161] The aforementioned data transmission steps can be executed by the bus arbitration module of the main control unit 200 in conjunction with the communication interface module. Arbitration and scheduling are implemented purely in hardware logic, without the need for software scheduling.

[0162] It should be noted that the "bus preemption," "fixed time slot allocation," and "idle time slot fragmentation transmission" mentioned above are all specific implementation methods of bus scheduling strategies. "Preemption" means that a high-priority event can interrupt the currently ongoing low-priority transmission and take priority in using bus resources; "fixed time slot" means allocating a fixed time window in the bus communication cycle for transmitting a specific type of data; "fragmentation transmission" means dividing a large data packet into multiple small data fragments and transmitting them at different times.

[0163] In some embodiments, the method further includes: S110. Measure the inherent delay from receiving the synchronization trigger signal to data output in each acquisition unit; S111: Write the inherent delay to the delay compensation register; S112. During operation, the timestamps of the sampled data output by each acquisition unit are offset and corrected according to the inherent delay stored in the delay compensation register.

[0164] There is an inherent hardware processing delay between the receipt of the synchronization trigger signal and the actual output of valid data by each acquisition unit; this delay is called inherent latency. This latency includes, but is not limited to: the conversion time of the analog-to-digital converter (for tactile sensing units), the digital signal processing time within the inertial measurement unit (IMU decoding delay), and the exposure delay and readout time of the image sensor (for vision units). The magnitude of these delays depends on the specific hardware specifications and configuration of each device and is usually known during the system design phase or can be measured during factory calibration.

[0165] During the system calibration phase, using the transition edge of the zero-position electrical signal as the time reference, the time interval from receiving the synchronization trigger signal to the valid data output of each acquisition unit is measured; this is the inherent delay. This measurement can be performed using a high-speed oscilloscope to simultaneously monitor the synchronization trigger signal and the data output signals of each device, measuring the time difference between the two.

[0166] The measured inherent delay value is written digitally into the delay compensation register of the main control unit 200 for storage. During operation, the main control unit 200 corrects the timestamps of the sampled data output by each acquisition unit based on the inherent delay value stored in the delay compensation register. Specifically, for the sampled data read from the data acquisition device, its timestamp is shifted forward and the inherent delay is subtracted to recover the actual sampling time (i.e., the trigger time) corresponding to the data. The corrected timestamp is aligned with the occurrence time of the synchronization trigger signal, thereby eliminating the time deviation introduced by the inherent processing delay of each device, with a compensation accuracy of ±0.1ms.

[0167] The inherent delay compensation and the aforementioned interval clock interpolation compensation together constitute a complete time synchronization calibration system: the interval clock interpolation compensation solves the local clock drift problem between two reference times, and the inherent delay compensation solves the fixed time offset problem between the trigger time and the data output time. The combination of these two methods ensures a full-range synchronization error ≤1ms.

[0168] It should be noted that the measurement and compensation of inherent delay can be performed separately by device type (i.e., all devices of the same type use the same inherent delay value), or it can be measured and compensated separately for each individual device (i.e., each device is calibrated independently). Compensation by device type is suitable for situations where there is good consistency among devices of the same type, while compensation for each individual device is suitable for scenarios with higher accuracy requirements. In some embodiments, compensation by device type is preferred to balance accuracy and calibration complexity.

[0169] Please see Figure 9 This timing diagram illustrates a system-level hardware synchronization scheme based on zero-bit input (T0). All links are completed by pure hardware logic, with no software scheduling delays. The sampling, triggering, and data output timings of each peripheral are as follows: Zero-position input signal (i.e., zero-position electrical signal): At time T0, the finger joint rotates through the zero-position reference position (i.e., the zero-position angle position of the joint), and the second type of track (e.g., a convex conductor) conducts with the second contact, generating a zero-position input signal with a rising edge (i.e., a second electrical signal with a rising edge), with a signal rise time ≤ 100ns. This signal is sent to the synchronization trigger module inside the main control unit. The pulse width is maintained at 1µs (T0-50ns to T0+950ns).

[0170] Synchronization reference signal (i.e., synchronous clock reference signal): The zero-position input signal enters a two-stage register (i.e., the cascaded first and second registers) timing circuit to complete cross-clock domain metastable synchronization processing; after the two-stage register synchronization processing, a metastable synchronous clock reference signal is output at time T0+20ns. This signal has a pulse width of 1µs (lasting until T0+1020ns) and a jitter of ≤50ns. This signal serves as the sole source hardware time reference for the entire multi-mode acquisition system, and its time accuracy is guaranteed by both physical contact conduction and pure hardware logic circuitry.

[0171] ADC latch signal (i.e., the first synchronous trigger signal): The synchronous reference signal fans out to generate the first synchronous trigger signal, which is output to the ADC acquisition module (i.e., the analog-to-digital converter of the tactile sensing unit). This signal generates a valid trigger edge (rising edge) at T0+100ns, and the latch window lasts for 500ns (until T0+600ns). During this period, the ADC sample-and-hold circuit is triggered to latch the current tactile analog input value. The latch delay is ≤200ns, completing the instantaneous capture of the tactile pressure signal. It should be noted that the high level of the ADC latch signal lasting for 500ns provides sufficient sampling window time for the ADC sample-and-hold circuit, ensuring that the analog signal is fully established before latching, thereby guaranteeing the accuracy and reliability of the analog-to-digital conversion.

[0172] IMU latch signal (i.e., the second synchronous trigger signal): The synchronous reference signal fans out to generate the second synchronous trigger signal, which is output to the inertial measurement unit. This signal generates a valid trigger edge (rising edge) at T0+300ns, and the latch window lasts for 1µs (until T0+1300ns). During this period, the IMU's internal registers are triggered to latch the current acceleration and angular velocity attitude data. The latch response time is ≤1µs, completing the instantaneous acquisition of hand attitude information.

[0173] Visual trigger signal (i.e., the third synchronous trigger signal): The synchronous reference signal is fanned out to generate a third homogeneous trigger signal, which is output to the binocular vision interface. This signal generates a valid trigger edge (rising edge) at T0+200ns, with a pulse width of 10µs (lasting until T0+10200ns). Within this window, the exposure trigger signal is output to start the hardware exposure of the binocular camera, with a trigger delay ≤500µs. After the camera exposure is completed, a certain readout time is required, so the image data is not immediately ready. It should be noted that the ADC latch signal, IMU latch signal, and visual trigger signal all originate from the same synchronous reference signal. The only difference between them is the nanosecond-level inherent offset caused by gate circuit fan-out, thus achieving hardware homogeneous triggering. Unlike schemes that use independent clock sources and software timers for each sensor, homogeneous triggering ensures a deterministic alignment relationship between the trigger signals in time, eliminating the problems of clock asynchrony and trigger timing uncertainty between different trigger sources.

[0174] Tactile data output: After the ADC latch is completed, the valid sampling point is T0+200ns (i.e., during the ADC latch window of the third waveform); the ADC performs analog-to-digital conversion, and after the conversion is completed, it outputs a valid tactile data frame. The data output starts at T0+10µs, and the output process lasts for 40µs (ending at T0+50µs). The time difference between the start of data output and the valid sampling point is the inherent hardware delay of the tactile sensing unit.

[0175] IMU Data Output: After the IMU latches the data, the latch validity point is T0+200ns (i.e., the IMU latching time). The IMU internally performs digital signal processing and computation, then outputs attitude data frames. The data output begins at T0+50µs and lasts for 50µs (ending at T0+100µs). The time difference between the data output start time (T0+300µs) and the latch validity point is the IMU's inherent hardware latency.

[0176] Frame exchange signal: After waiting for all tactile data, IMU data and visual image data to be ready, perform multimodal frame exchange operation (that is, pack the tactile data, posture data and image data obtained by each acquisition unit synchronously sampled into target data frames), pack the tactile data, posture data and image data triggered by the same zero-position reference into a group of synchronous frames, ready for external transmission.

[0177] From the transition edge of the zero-position input signal to the output of the three co-source trigger signals, the total link delay is ≤1μs; within a single reference trigger frame, the physical sampling time deviation of the tactile, IMU, and visual three-modal approaches is ≤0.5ms; after superimposed interval clock interpolation compensation, the system's full-range multimodal synchronization error is ≤1ms. The above delay and deviation indicators are inherent electrical characteristics of the pure hardware synchronization link implementation, requiring no software intervention, thus ensuring the time accuracy and determinism of synchronous acquisition.

[0178] Please see Figure 10 This is a three-level priority bus scheduling time slot allocation diagram according to an embodiment of this application. From top to bottom, the diagram shows the P0 event request signal, P1 period request signal, P2 image request signal, and actual bus occupancy. The three priority levels, P0, P1, and P2, are marked in white, black, and gray, respectively.

[0179] The highest priority P0 (white): Zero-position events (i.e., the reference position event corresponding to the second electrical signal, corresponding to the synchronization event corresponding to the synchronization trigger signal) and tactile abrupt change events (i.e., events where the rate of change of the tactile signal exceeds a preset threshold, corresponding to the joint angle abrupt change event) have the highest priority. When the P0 event request signal arrives, the bus arbitration module immediately occupies the bus in a preemptive manner, without waiting for the current frame to complete transmission, with an end-to-end delay ≤5ms. The P0 event is requested at time t0, and the bus responds immediately at time t0. The P0 data frame releases the bus after completing transmission.

[0180] Medium Priority P1 (Black): Joint angle data and tactile data sampled periodically have medium priority. P1 data is continuously generated at a fixed frequency. The bus arbitration module allocates fixed time slots for P1 data to ensure that data from each sampling cycle is transmitted before the arrival of the next sampling cycle. The P1 cycle request signal arrives at a fixed time interval. The bus responds to the P1 request in an idle time slot without a P0 event and completes data transmission within the fixed time slot window. P1 transmission may be preempted by a P0 event (when a P0 event request and P1 transmission occur simultaneously, the bus responds to P0 first, and P1 transmission is delayed, but the delay range is controlled within P1 cycle jitter ≤10%).

[0181] Lowest Priority P2 (Gray): Visual image data has the lowest priority. This image data is large in volume but has lower real-time requirements than P0 and P1 data. The P2 image request signal is continuously present (continuously pending transmission, segmented transmission). The bus only allows P2 data to occupy the bus for transmission when it detects that the bus is idle (i.e., there is currently no P0 or P1 data transmission). P2 transmission uses a segmented method, transmitting a fixed-size data segment (i.e., the "P2" time period) at a time. After transmitting one data segment, the bus is immediately released, waiting for the next idle time slot before transmitting the next data segment, until the entire image is transmitted. When a P0 or P1 event occurs during P2 transmission, P2 transmission is immediately paused, and the bus responds to the P0 or P1 request first.

[0182] Actual bus occupancy: The actual occupancy of the bus on the time axis is determined by the arbitration result. During the time interval t0 to t1, if a P0 event request arrives, the bus is actually occupied by P0. During the time interval t1 to t2, if there is no P0 event and a P1 periodic request arrives, the bus is actually occupied by P1. During the time interval t2 to t3, if there are no P0 or P1 events and a P2 image request is transmitted using an idle time slot, the bus is actually occupied by P2. At time t3, if a new P0 event request arrives, the bus is immediately preempted by P0, and P2 transmission is paused. During the time interval t4 to t5, P1 transmission is completed within a fixed time slot.

[0183] The highest priority is allocated to the synchronization events and / or joint angle mutation events corresponding to the synchronization trigger signal, and the bus is occupied in a preemptive manner; the medium priority is allocated to the periodically sampled joint angle data and tactile data, and fixed time slots are allocated for transmission; the low priority is allocated to the visual image data, and the data is transmitted in segments using the idle time slots of the bus.

[0184] The P0 end-to-end latency is ≤5ms, ensuring real-time response to synchronization and angle change events; the overall bus utilization is ≥85%, far exceeding the static allocation scheme without priority scheduling (approximately 50%), balancing the real-time nature of critical events with the efficiency of full data transmission; the P1 period jitter is ≤10%, guaranteeing stable transmission bandwidth for periodically sampled data and preventing data overwriting and loss. The above scheduling and arbitration processes are implemented purely in hardware by the bus arbitration module, requiring no software involvement in arbitration decisions, thus ensuring the real-time nature and determinism of scheduling.

[0185] In one specific embodiment, please refer to Figure 11 Multimodal synchronous data acquisition methods include: Step S201: After the system powers on and starts up, the MCU is configured and the hardware peripherals are initialized.

[0186] The MCU unit (i.e., the main control unit) begins to execute the startup configuration, including loading system parameters, configuring the working mode of each interface module, initializing the cascade bus interface, ADC acquisition module, IMU attitude unit and the register parameters of the binocular vision interface.

[0187] Step S202: Calibration of reference parameters.

[0188] After the hardware peripherals are initialized, the system enters the reference parameter calibration stage, which includes three sub-steps: delay calibration (measuring the inherent delay of each data acquisition device, i.e., measuring the inherent delay of each acquisition unit from receiving the synchronous trigger signal to data output), template calibration (establishing abrupt change detection template for tactile signals and angle signals for subsequent event determination), and spatial calibration (determining the zero-position reference position and range of motion of each joint).

[0189] Step S203: Configure runtime parameters.

[0190] After the baseline parameters are calibrated, the system performs a consistency check, verifying the completeness and reasonableness of the calibration results. If the data is normal, the system proceeds to the runtime parameter configuration phase, writing the calibration results into the configuration registers of each module (including writing the inherent delay into the delay compensation register). Step S204: Normal asynchronous data acquisition.

[0191] After initialization, the system enters the normal asynchronous acquisition mode. Each acquisition unit operates independently according to its own sampling frequency, continuously acquiring joint angle data (generated by the first type of track and the first contact part of the sliding electrical contact node device 100 of each joint, and obtained by decoding circuit), tactile data (generated by tactile sensing unit and transmitted to ADC acquisition module), IMU data (acquired by inertial measurement unit) and visual image data (acquired by binocular vision unit).

[0192] Step S205: Pre-trigger monitoring cache.

[0193] The main controller (i.e., the synchronous trigger module) performs pre-trigger monitoring and buffering of data from each channel, with a buffer depth sufficient to accommodate all data frames between two adjacent zero-bit events.

[0194] Step S206: Determine whether to enter the warning window.

[0195] During normal data acquisition, the main controller continuously monitors the zero-position electrical signal and the joint angle change rate. When a zero-position electrical signal transition edge or the joint angle change rate exceeds a preset threshold is detected, the system determines that it is entering an early warning window, that is, it executes step S207.

[0196] Step S207: Initiate Level 1 Hysteresis Detection.

[0197] Initiate pre-triggered buffer locking, marking the data in the current buffer as frames to be synchronized; perform hysteresis comparison verification on the trigger signal to confirm whether the signal has indeed crossed the reference position, avoiding false triggering caused by signal noise. Determine whether the determination passes; if so, proceed to step S208.

[0198] Step S208, Secondary morphological matching.

[0199] The local shape of the current angle / tactile waveform is matched and verified with a pre-stored template to confirm whether it is a valid event. If the match is valid, proceed to step S209; if the match is invalid, return to step S204.

[0200] Step S209: Level 3 hardware synchronization trigger.

[0201] The zero-position synchronization trigger module in the main controller generates a synchronization clock reference signal based on the zero-position electrical signal, and then outputs three synchronous trigger signals of the same source, which are used to trigger the tactile ADC to latch the sampled value, trigger the IMU to latch the current posture data, and trigger the binocular camera to expose, thereby realizing hardware-level synchronous sampling of tactile, IMU and vision three modalities.

[0202] Meanwhile, during the synchronization triggering process, the bus arbitration module performs a three-level priority time slot allocation. The highest priority P0 is allocated to synchronization events and angle change events (preemptively occupying the bus), the middle priority P1 is allocated to periodically sampled joint angle data and tactile data (allocated to fixed time slots for transmission), and the low priority P2 is allocated to visual image data (using idle time slots on the bus for fragmented transmission), ensuring low-latency transmission of synchronization frames.

[0203] Step S210: Inherent delay compensation.

[0204] Subtract the inherent delay from the timestamps of the sampled data output by each data acquisition device, and push back the timestamps of all modes to the actual triggering time of the synchronization reference signal.

[0205] Step S211: Generate a reference synchronization frame.

[0206] The main controller fuses and packages the tactile data, joint angle data, IMU attitude data and visual image data obtained by latching under the same zero-position reference trigger into a set of reference synchronization frames according to a preset frame format. The reference synchronization frame takes the zero-position trigger time T0 as the initial time reference.

[0207] Step S212, Interval timing reconstruction (i.e., clock interpolation compensation) and global clock calibration Interval timing reconstruction: Record the number of sampling frames and local clock time of each data acquisition device between two adjacent zero-position electrical signals; use the generation time of two adjacent zero-position electrical signals as the hardware reference, and perform linear drift calibration on the local clock of each acquisition unit respectively; based on the calibrated local clock time, map the sampling data of each acquisition unit between two adjacent hardware references to a unified global timestamp.

[0208] Global clock calibration: Based on the above compensation results, a reference synchronization frame is generated, and the tactile, joint angle, IMU and visual data triggered by the same zero-position reference are packaged into a group of synchronization frames.

[0209] Step S213: Three-level priority scheduling.

[0210] After global clock calibration, the calibrated multimodal data enters the bus arbitration module, which performs a three-level priority time slot allocation. Priority in P1 is allocated to periodically sampled joint angle data and tactile data, with fixed time slots for transmission. Lower priority in P2 is allocated to visual image data, which is transmitted in segments using idle bus time slots. This three-level priority scheduling ensures that the baseline synchronization frame (containing key synchronization event data) after interval timing reconstruction is transmitted with the lowest latency, guaranteeing an end-to-end latency of ≤5ms for synchronization events. Simultaneously, it balances the stable transmission bandwidth of periodically sampled data and the large-capacity transmission requirements of image data, achieving an overall bus utilization rate of ≥85%.

[0211] Step S214: Pack and output the data.

[0212] After the three-level priority scheduling is completed, the multimodal data (tactile data, joint angle data, IMU data, and visual image data) output by the scheduling are packaged into target data frames according to the preset frame format. The target data frame contains timestamp fields, tactile data fields, angle data fields, IMU data fields, and image data fields.

[0213] Step S215: Consistency check.

[0214] After the data is packaged, the system performs a consistency check on the packaged target data frame, including checking whether the data frame format is complete, whether the timestamps are continuous, whether the time deviation of each modality data is within the allowable range (≤1ms), and whether the checksum of the data frame is correct. If the consistency check passes, step S216 is executed; if the consistency check fails, step S217 is executed.

[0215] Step S216: Transmit the target data frame outward.

[0216] For example, the target data frame is transmitted to the host computer / data terminal through the data output module. After receiving the data, the host computer / data terminal completes the multimodal synchronous data acquisition for this round, and then returns to step S204 to continue monitoring the next round of zero-position events, realizing full closed-loop continuous operation.

[0217] Step S217, Error Assessment and Calibration.

[0218] The specific content of error assessment and calibration includes: analyzing the reasons for failure. After calibration is completed, return to step S204 and start the next round of data acquisition.

[0219] Other configurations of the above embodiments can be adopted from various technical solutions that are now and will be known to those skilled in the art, and will not be described in detail here.

[0220] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0221] It should be noted that although the steps of the method in this application are described in a specific order in the accompanying drawings, this does not require or imply that these steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps. The above drawings are merely illustrative of the processes included in the method according to exemplary embodiments of this application and are not intended to be limiting. It is readily understood that the processes shown in the above drawings do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may be performed synchronously or asynchronously in multiple modules, for example.

[0222] The foregoing disclosure provides many different implementations or examples for carrying out different structures of this application. To simplify the disclosure, specific examples of components and arrangements are described above. Of course, these are merely examples and are not intended to limit the scope of this application. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various implementations and / or arrangements discussed.

[0223] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A glove-based multimodal data acquisition system, characterized in that, include: A joint sliding electrical contact node device, comprising: Insulating substrate; Conductive tracks are disposed on the surface of the insulating substrate, the conductive tracks comprising first-class tracks and second-class tracks electrically isolated from each other; A contact assembly, comprising a first contact portion that slides in contact with a first type of track and a second contact portion that slides in contact with a second type of track, wherein the first contact portion is configured to output a first electrical signal characterizing a rotation angle of a joint based on the contact position between the first contact portion and the first type of track, and the second contact portion is configured to output a second electrical signal characterizing a reference position event when the joint rotates through a reference position; and The main control unit is electrically connected to the joint sliding electrical contact node device. The main control unit is configured to receive the first electrical signal and the second electrical signal, and generate a synchronous clock reference signal based on the second electrical signal.

2. The system according to claim 1, characterized in that, The insulating substrate includes at least a spherical portion, and the conductive tracks are arranged circumferentially on the surface of the insulating substrate.

3. The system according to claim 1, characterized in that, The first type of track is arranged circumferentially along the insulating substrate and has multiple segments, the multiple segments including conductive segments and insulating segments, the conductive segments and the insulating segments are alternately distributed circumferentially along the insulating substrate; the first contact portion sweeps through each of the segments sequentially as the insulating substrate rotates, and the first electrical signal is the on / off signal sequence generated when the first contact portion contacts each of the segments.

4. The system according to claim 3, characterized in that, The conductive segments and the insulating segments are distributed circumferentially along the insulating substrate according to Gray code encoding, and the on / off signal sequence is a Gray code encoded signal.

5. The system according to claim 1, characterized in that, The second type of track is disposed on the surface of the insulating substrate in a region corresponding to the reference position, wherein the reference position is a preset angle position of the joint.

6. The system according to claim 5, characterized in that, The second type of track is a raised conductor formed on the surface of the insulating substrate; the second contact portion makes contact with the raised conductor when the joint rotates through the preset angle position of the joint to generate the second electrical signal with a transition edge based on the contact conduction.

7. The system according to claim 1, characterized in that, The conductive track also includes a third type of track, which includes a power supply positive track and a common ground track, both of which are continuous circular conductive tracks used to transmit electrical energy.

8. The system according to claim 1, characterized in that, The conductive track also includes a fourth type of track, which is a continuous ring-shaped conductive track used to transmit tactile analog signals.

9. The system according to claim 1, characterized in that, An insulating groove is provided between each of the conductive tracks to electrically isolate adjacent conductive tracks.

10. The system according to claim 6, characterized in that, The joint sliding electrical contact node device further includes a local processing circuit, which includes a decoding circuit and / or a detection circuit. The decoding circuit is electrically connected to the contact component and is used to decode the first electrical signal to obtain joint angle data; The detection circuit is electrically connected to the contact component and is used to perform debouncing and edge extraction processing on the second electrical signal to output a detection signal with a jumping edge. The main control unit is also configured to receive the detection signal and generate the synchronous clock reference signal based on the detection signal.

11. The system according to claim 1, characterized in that, The contact component includes multiple sets of elastic metal brushes, each set of elastic metal brushes making contact with each of the conductive tracks.

12. The system according to claim 1, characterized in that, The joint sliding electrical contact node device is multiple, and is respectively set at different joints of the same finger. Each joint sliding electrical contact node device is cascaded in series through a set of bus harnesses.

13. The system according to claim 10, characterized in that, The main control unit includes a synchronization trigger module, which includes a cascaded first register and a second register. The data input terminal of the first register is used to receive the detection signal, the clock input terminal of the first register is used to connect to the system clock, the data output terminal of the first register is connected to the data input terminal of the second register, the clock input terminal of the second register is used to connect to the system clock, and the data output terminal of the second register is used to output the synchronization clock reference signal.

14. The system according to claim 13, characterized in that, The main control unit also includes a data acquisition device, which includes multiple acquisition units of different types. The synchronization trigger module outputs multiple synchronous trigger signals of the same source based on the synchronization clock reference signal, which are used to trigger the different types of acquisition units of the data acquisition device to perform synchronous sampling.

15. The system according to claim 14, characterized in that, The plurality of different types of acquisition units include a tactile sensing unit, an inertial measurement unit, and a vision unit; the multiple synchronous trigger signals originating from the same source include at least two of the following: The first synchronous trigger signal is used to trigger the analog-to-digital converter of the tactile sensing unit to latch the current sampled value; The second synchronous trigger signal is used to trigger the inertial measurement unit to latch the current attitude data; The third synchronous trigger signal is used to trigger the exposure of the vision unit.

16. The system according to claim 12, characterized in that, The main control unit also includes a bus arbitration module, used to allocate three-level priority time slots for the communication of the bus harness, wherein: The highest priority is assigned to the synchronization events and / or joint angle change events corresponding to the synchronization clock reference signal, and the bus is occupied in a preemptive manner; The joint angle data and tactile data, which are periodically sampled, are allocated medium priority and transmitted in fixed time slots. Low priority is allocated to visual image data, which is then transmitted in segments using idle time slots on the bus.

17. The system according to claim 12, characterized in that, The main control unit also includes a clock compensation module and a data acquisition device. The data acquisition device includes multiple acquisition units of different types. The clock compensation module is used to perform linear drift calibration and interval interpolation compensation on the local clock of each acquisition unit using two adjacent second electrical signals as hardware references, so as to correct the timestamp deviation of each acquisition unit between two adjacent hardware references.

18. A multimodal synchronous data acquisition method, characterized in that, include: Acquire the zero-position electrical signal generated by the joint sliding electrical contact node device set at the finger joint when the joint rotates through the reference position. A synchronous clock reference signal is generated based on the zero-position electrical signal; The data acquisition device is controlled to perform synchronous sampling according to the synchronous clock reference signal; The data acquisition device includes multiple different types of acquisition units, which include two or more of the following: a tactile sensing unit, an inertial measurement unit, and a vision unit.

19. The method according to claim 18, characterized in that, The joint sliding electrical contact node device includes an insulating substrate, a conductive track disposed on the surface of the insulating substrate, and a contact assembly; the conductive track includes a second type of track; the contact assembly is configured to output a second electrical signal characterizing the reference position event when the joint rotates through a reference position; The zero-position electrical signal is the second electrical signal generated when the second type of track and the contact component make contact and conduct when the joint rotates through the reference position.

20. The method according to claim 18, characterized in that, Also includes: Based on the synchronous clock reference signal, multiple synchronous trigger signals originating from the same source are output, wherein the synchronous trigger signals include at least two of the following: The first synchronous trigger signal is used to trigger the analog-to-digital converter of the tactile sensing unit to latch the current sampled value; The second synchronous trigger signal is used to trigger the inertial measurement unit to latch the current attitude data; The third synchronous trigger signal is used to trigger the exposure of the visual unit.

21. The method according to claim 18, characterized in that, Also includes: Record the number of sampling frames and the local clock time between two consecutive zero-position electrical signals for each of the data acquisition devices; Using the generation times of two consecutive zero-position electrical signals as hardware references, the local clocks of each data acquisition device are linearly drift calibrated. Based on the calibrated local clock time, the sampled data of each data acquisition device between two adjacent hardware references are mapped to a unified global timestamp.

22. The method according to claim 21, characterized in that, The linear drift calibration of the local clocks of each of the data acquisition devices includes: Calculate the clock correction factor for the i-th type of data acquisition device according to the following formula: ， Among them, T k and T k+1 These are the global timestamps of two consecutive zero-position electrical signals, t and t', respectively. i,k and t i,k+1 These are the local clock times of the i-th type of data acquisition device during two consecutive zero-position electrical signals; The clock correction coefficient is stored in a register for timestamp mapping of the sampled data frames of each acquisition unit.

23. The method according to claim 20, characterized in that, Also includes: When a sudden change in the zero-position electrical signal or joint angle is detected, the corresponding synchronization event or angle change event will be transmitted on the bus with the highest priority. The periodically sampled joint angle data and tactile data are transmitted in fixed time slots; Visual image data is transmitted in segments using idle time slots on the bus.

24. The method according to claim 23, characterized in that, Also includes: Measure the inherent delay of each of the data acquisition devices from receiving the synchronization trigger signal to data output; Write the inherent delay into the delay compensation register; During runtime, the timestamps of the sampled data output by each of the data acquisition devices are offset and corrected according to the inherent delay stored in the delay compensation register.

Images