Wearable centaur robot rigidity adjusting method and device, electronic equipment, storage medium, robot and system

Abstract

The application provides a soft elastic coupling mechanism stiffness adjustment method and device of a wearable centaur robot, an electronic device, a computer readable storage medium, a wearable centaur robot and a system. The method determines walking gait phase data of a human object and load material counterweight data of the wearable centaur robot, extracts leg support period landing characteristics, determines a load working condition level, determines a target pre-compression displacement parameter of an antagonistic spring assembly, generates a displacement adjustment instruction and drives a micro lead screw execution assembly to change an initial deformation scale to update an equivalent physical stiffness.

Description

Technical Field

[0001] This application relates to the field of wearable robot control technology, and more specifically, to a method, device, electronic device, computer-readable storage medium, wearable centaur robot and system for adjusting the stiffness of a softened elastic coupling mechanism of a wearable centaur robot. Background Technology

[0002] With the development of exoskeleton assistive devices, wearable load-bearing robots, and human-robot collaborative handling equipment, wearable robots designed for navigating complex terrains, carrying heavy loads, and providing human support are increasingly being applied in scenarios such as material handling, field operations, rehabilitation training, and emergency response. In these scenarios, the robot needs to bear additional loads while following the human body's movements, and maintain good collaborative stability under conditions of changes in human walking posture, ground undulations, and load variations, in order to reduce the impact of robot motion disturbances on the human body's support status.

[0003] Existing wearable load-bearing robots typically employ rigid supports, fixed hinges, or passive elastic connectors at the human-robot connection point to achieve mechanical connection between the human body and the robot body. This type of solution generally first connects the human-side support to the robot's load-bearing body via connectors, then utilizes the connector's own fixed structure or fixed elastic parameters to transmit the human's traction force and the robot's load-bearing reaction force. Finally, the robot's chassis or leg mechanisms complete the following movement and load transfer.

[0004] However, the above solutions still have technical drawbacks. Rigid connection structures tend to directly transmit the swaying of the robot during movement to the human body, resulting in low human-robot coordination. Although passive connectors with fixed elastic parameters can provide some cushioning, their connection stiffness is difficult to adapt to changes in the human walking state and the robot's load state. This can lead to a connection response that is too stiff during light-load walking, while the connection support capacity may be insufficient under heavy load or impact scenarios, thus affecting the cooperative stability between the human support side and the robot body. Summary of the Invention

[0005] This application provides a method, device, electronic device, computer-readable storage medium, wearable centaur robot, and system for adjusting the stiffness of a softened elastic coupling mechanism, in order to at least alleviate the aforementioned technical problems.

[0006] A method for adjusting the stiffness of a softened elastic coupling mechanism in a wearable centaur robot includes: Determine the walking gait phase data of the human object in the current detection cycle; Determine the weight data of the material carried by the wearable Centaur robot during the current detection cycle; Based on the walking gait phase data, extract the leg support phase ground contact characteristics of the human object; Based on the weight data of the load-bearing material, the load condition level of the wearable Centaur robot is determined; Based on the load condition level and the leg support period ground landing characteristics, the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism is determined. The softened elastic coupling mechanism connects the human body object and the main skeleton of the wearable centaur robot through a diamond-shaped linkage transmission component. Based on the target pre-compression displacement parameter, a displacement adjustment command matching the target pre-compression displacement parameter is generated; The displacement adjustment command is sent to the micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism.

[0007] A stiffness adjustment device for a softened elastic coupling mechanism of a wearable centaur robot includes: The gait phase determination module is configured to determine the walking gait phase data of a human object in the current detection cycle; The counterweight data determination module is configured to determine the counterweight data of the material carried by the wearable Centaur robot in the current detection cycle; The ground landing feature extraction module is configured to extract the leg support phase ground landing features of the human object based on the walking gait phase data; The working condition level determination module is configured to determine the load working condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The pre-compression parameter determination module is configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism based on the load condition level and the leg support period ground landing characteristics. The softened elastic coupling mechanism connects the human object and the main skeleton of the wearable centaur robot through a diamond-shaped linkage transmission component. The adjustment command generation module is configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The stiffness update control module is configured to send the displacement adjustment command to the micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism.

[0008] An electronic device, comprising a memory and a processor; The memory stores computer programs; When the processor executes the computer program, it implements the above-described method for adjusting the stiffness of the softened elastic coupling mechanism.

[0009] A computer-readable storage medium storing a computer program; When the computer program is executed by the processor, it implements the above-mentioned method for adjusting the stiffness of the softened elastic coupling mechanism.

[0010] A wearable centaur robot includes a main skeleton, a softened elastic coupling mechanism, a micro-screw actuator, and a stiffness adjustment controller; The softened elastic coupling mechanism includes a rhomboid connecting rod transmission component and an antagonistic spring assembly. The diamond-shaped connecting rod is used to connect the human body object to the main skeleton. The stiffness adjustment controller is configured to determine the walking gait phase data of the human object in the current detection cycle; The stiffness adjustment controller is also configured to determine the weight data of the material carried by the wearable Centaur robot in the current detection cycle; The stiffness adjustment controller is also configured to extract the leg support phase ground contact characteristics of the human object based on the walking gait phase data; The stiffness adjustment controller is also configured to determine the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The stiffness adjustment controller is also configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the load condition level and the leg support period ground contact characteristics; The stiffness adjustment controller is further configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The micro-screw actuator is configured to receive the displacement adjustment command and change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command to update the equivalent physical stiffness of the softened elastic coupling mechanism.

[0011] A stiffness adjustment system for a wearable centaur robot with a softened elastic coupling mechanism includes a gait phase sensing unit, a counterweight data sensing unit, a stiffness adjustment control unit, a micro-screw actuation component, and a softened elastic coupling mechanism. The softened elastic coupling mechanism includes a rhomboid connecting rod transmission component and an antagonistic spring assembly. The rhomboid linkage is used to connect the human body object to the main skeleton of the wearable centaur robot. The gait phase sensing unit is configured to determine the walking gait phase data of the human object in the current detection cycle and transmit the walking gait phase data to the stiffness adjustment control unit. The counterweight data sensing unit is configured to determine the counterweight data of the material carried by the wearable Centaur robot in the current detection cycle, and transmit the counterweight data of the material carried to the stiffness adjustment control unit. The stiffness adjustment control unit is configured to extract the leg support phase ground contact characteristics of the human object based on the walking gait phase data; The stiffness adjustment control unit is also configured to determine the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The stiffness adjustment control unit is also configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the load condition level and the leg support period ground contact characteristics; The stiffness adjustment control unit is further configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The micro-screw actuator is configured to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, so as to update the equivalent physical stiffness of the softened elastic coupling mechanism.

[0012] The technical advantages of the technical solution provided in this application are: This application presents a method for adjusting the stiffness of a softened elastic coupling mechanism in a wearable centaur robot. Addressing the technical shortcomings of existing wearable load-bearing robots that use fixed connection stiffness at the human-machine interface, making it difficult to simultaneously account for changes in human walking state and robot load state, this method first determines the human gait phase data during the current detection cycle and then determines the weight data of the load-bearing material on the wearable centaur robot during the same detection cycle. This allows subsequent stiffness adjustment to not only rely on fixed structural parameters but also incorporate both human motion state and robot load state data. Compared to traditional rigid or fixed elastic connection schemes, this application enables the adjustment of connection stiffness to correspond to the human walking state and load-bearing material state within the current detection cycle, thereby reducing the possibility of the connection response becoming disconnected from the actual human-machine collaborative scenario.

[0013] This application extracts the leg support phase ground contact characteristics of a human object based on walking gait phase data, and determines the load condition level of the wearable Centaur robot based on the load-bearing material weight data. This allows the human support side ground contact state and the robot load state to form technical parameters that can participate in stiffness adjustment. Traditional fixed connection schemes usually cannot distinguish the human force requirements under different support phase ground contact states, nor can they distinguish the robot support requirements under different load-bearing material weights. This application uses the leg support phase ground contact characteristics and load condition level to jointly participate in the determination of subsequent target pre-compression displacement parameters, enabling the adjustment direction of the softened elastic coupling mechanism to simultaneously respond to changes in human side support and robot side load, thereby improving the condition adaptability of the stiffness configuration at the human-machine connection.

[0014] This application determines the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism based on the load level and the landing characteristics during the leg support phase. The human body is then connected to the main skeleton of the wearable centaur robot via a diamond-shaped linkage in the softened elastic coupling mechanism. Compared to solutions that rely solely on passive buffering with fixed elastic components, this application uses the target pre-compression displacement parameter as an intermediate control target to change the initial deformation scale of the antagonistic spring assembly. This allows the connection stiffness to form an adjustable elastic coupling state along the mechanical force transmission path between the human body and the main skeleton, thereby reducing the degree to which robot sway is directly transmitted to the human support side via the connection structure and improving the adaptability of the connection structure to different loads and asynchronous states.

[0015] This application generates a matching displacement adjustment command based on the target pre-compression displacement parameter and sends the displacement adjustment command to a micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism. Compared with traditional passive connectors with fixed elastic parameters, this application establishes an execution channel from the calculated target pre-compression displacement parameter to the physical deformation state of the antagonistic spring assembly through the displacement adjustment command and the micro-screw actuator. This allows the equivalent physical stiffness to be updated according to the walking and load states within the current detection cycle, thereby reducing the risk of decreased human-machine coordination caused by fixed stiffness connections in scenarios such as walking, carrying loads, and obstacle crossing of wearable centaur robots. Attached Figure Description

[0016] Figure 1 This application provides an embodiment of a scenario for adjusting the stiffness of a wearable centaur robot's softened elastic coupling mechanism. Figure 2 This application provides a method for adjusting the stiffness of a softened elastic coupling mechanism in a wearable centaur robot. Figure 3This application provides an embodiment of a wearable centaur robot with a softened elastic coupling mechanism stiffness adjustment device. Figure 4 An electronic device is described in an embodiment of this application; Figure 5 This application provides an embodiment of a computer-readable storage medium. Figure 6 This application provides an embodiment of a mobile robot. Figure 7 This application provides an embodiment of a wearable centaur robot with a softened elastic coupling mechanism stiffness adjustment system. Detailed Implementation

[0017] like Figure 1 The image shows a scenario for adjusting the stiffness of a softened elastic coupling mechanism in a wearable centaur robot according to an embodiment of this application; as shown... Figure 2 As shown, this embodiment of the present application discloses a method for adjusting the stiffness of a softened elastic coupling mechanism of a wearable centaur robot, comprising: Determine the walking gait phase data of the human object in the current detection cycle; Determine the weight data of the material carried by the wearable Centaur robot during the current detection cycle; Based on the walking gait phase data, extract the leg support phase ground contact characteristics of the human object; Based on the weight data of the load-bearing material, the load condition level of the wearable Centaur robot is determined; Based on the load condition level and the leg support period ground landing characteristics, the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism is determined. The softened elastic coupling mechanism connects the human body object and the main skeleton of the wearable centaur robot through a diamond-shaped linkage transmission component. Based on the target pre-compression displacement parameter, a displacement adjustment command matching the target pre-compression displacement parameter is generated; The displacement adjustment command is sent to the micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism.

[0018] Optionally, determining the walking gait phase data of the human object in the current detection period includes: The spatial posture of the lower limbs of the human object is collected during the current detection period by an inertial measurement unit deployed on the lower limbs of the human object. The distribution of plantar pressure of the human body during the current detection period is collected by a thin-film array sensor applied to the sole of the foot. Perform coordinate system transformation on the lower limb spatial posture to determine the human lower limb posture data in the human body base reference frame; The human lower limb pose data is appended to the plantar pressure distribution to generate a multi-source sensing fusion feature vector; Based on the multi-source sensing fusion feature vector, the walking gait phase data of the human object in the current detection cycle is determined.

[0019] Preferably, the specific implementation process of collecting the lower limb spatial posture of the human object during the current detection cycle by using an inertial measurement unit deployed on the lower limb part of the human object is as follows: At the beginning of the current detection cycle, the installation position mark of the inertial measurement unit corresponding to the lower limb part of the human object is read first, and a lower limb spatial posture acquisition and bearing record is established based on the installation position mark of the inertial measurement unit. The installation position mark of the inertial measurement unit is used to express the installation source of the inertial measurement unit fixed on the thigh, calf, or instep of the human object, so that the lower limb spatial posture subsequently collected can be corresponding to the actual lower limb movement part of the human object, rather than being interpreted as an ordinary spatial posture unrelated to the human object; the lower limb spatial posture acquisition and bearing record is used to carry the sampling data after the installation source is defined by the installation position mark of the inertial measurement unit, so that the sampling data in the lower limb spatial posture acquisition and bearing record can be continuously traced back to the lower limb part of the human object. Subsequently, within the sampling time range defined by the current detection cycle, the inertial measurement unit continuously acquires angular velocity, linear acceleration, and attitude angle sampling values ​​of the lower limbs of the human object, and writes these values ​​into the lower limb spatial attitude acquisition and bearing record according to their sampling time sequence. The sampling time is used to express the chronological order of the angular velocity, linear acceleration, and attitude angle sampling values ​​within the current detection cycle. The sampling time is subsequently processed with timestamp alignment to ensure that the angular velocity, linear acceleration, and attitude angle sampling values ​​can be combined to the same sampling moment. The angular velocity sampling value expresses the velocity changes of the lower limbs of the human object as they swing around various spatial axes; the linear acceleration sampling value expresses the inertial changes of the lower limbs during swinging, support transitions, and trunk traction; and the attitude angle sampling value expresses the tilt state of the lower limbs of the human object relative to the local reference direction of the inertial measurement unit. Furthermore, the angular velocity sample value, linear acceleration sample value, and attitude angle sample value in the lower limb spatial attitude acquisition and bearing record are timestamped to form a lower limb spatial attitude data frame at the same sampling time; then, the lower limb spatial attitude is formed according to the sequential relationship of the lower limb spatial attitude data frames in the current detection period.The lower limb spatial posture is not a single posture angle value, but rather a human lower limb movement state expressed by multiple consecutive lower limb spatial posture data frames. The lower limb spatial posture is subsequently processed by coordinate system transformation, so that the swing direction, support transition direction, and posture change amplitude of the human object's lower limbs can be interpreted in a unified human basic reference frame. The swing direction, support transition direction, and posture change amplitude are all derived from the angular velocity sample value, linear acceleration sample value, and posture angle sample value in the lower limb spatial posture data frame, and continue to participate in the subsequent determination of the human lower limb pose data, thereby providing a human motion side data source for the subsequent determination of the human object's walking gait phase data in the current detection period.

[0020] Preferably, the specific implementation process of collecting the plantar pressure distribution of the human object within the current detection cycle using a thin-film array sensor installed on the sole of the foot is as follows: Within the same current detection cycle at the start of the lower limb spatial posture acquisition, the array installation position mark of the thin-film array sensor on the sole of the human object is first read, and a plantar pressure distribution acquisition and load record is established based on the array installation position mark. The array installation position mark is used to express the plantar position of each pressure sampling node in the thin-film array sensor corresponding to the heel region, arch region, and toe region, respectively, so that the plantar pressure distribution can reflect the contact migration of the sole of the human object during walking, rather than just reflecting the total pressure on the sole; the plantar pressure distribution acquisition and load record is used to incorporate the array installation position mark, each pressure sampling node, and the subsequently formed plantar pressure distribution data frame into the same acquisition caliber, so that the node pressure sampling values ​​in the plantar pressure distribution data frame can correspond to the heel region, arch region, and toe region. Subsequently, within the sampling time range defined by the current detection cycle, the thin-film array sensor continuously collects nodal pressure sampling values ​​of multiple pressure sampling nodes according to the array laying position markings, and writes the multiple nodal pressure sampling values ​​at each sampling moment into the plantar pressure distribution acquisition and bearing record. The nodal pressure sampling values ​​are used to express the local pressure degree formed when the corresponding pressure sampling node of the human foot comes into contact with the ground; multiple nodal pressure sampling values ​​at the same sampling moment form the plantar pressure distribution data frame according to the arrangement diameter of the heel area, arch area, and toe area; multiple consecutive plantar pressure distribution data frames together form the plantar pressure distribution. Further, the plantar pressure distribution data frames undergo effective contact filtering processing. Node pressure sampling values ​​below the contact recognition threshold are marked as non-contact node pressure sampling values, while node pressure sampling values ​​reaching the contact recognition threshold are marked as contact node pressure sampling values. The contact recognition threshold is configured based on the natural foot contact state when the thin-film array sensor completes wear initialization. This threshold distinguishes the node pressure sampling values ​​into non-contact node pressure sampling values ​​and contact node pressure sampling values. Non-contact node pressure sampling values ​​exclude pressure sampling nodes where effective foot bearing has not been achieved, while contact node pressure sampling values ​​retain pressure sampling nodes where effective foot bearing has been achieved. Subsequently, a plantar contact migration sequence is formed based on the positional changes of the contact node pressure sampling values ​​in the plantar region. The non-contact node pressure sampling values ​​are included as non-bearing reference content in the plantar contact migration sequence and written into the plantar pressure distribution acquisition bearing record.The plantar contact migration sequence is stored in the plantar pressure distribution acquisition and bearing record, and participates in the subsequent generation of the multi-source sensing fusion feature vector as part of the plantar pressure distribution. In this way, the plantar pressure distribution can express support contact states such as heel contact, plantar bearing transition, and toe lift-off. This allows the subsequent determination of the walking gait phase data to utilize not only the changes in motion posture in the lower limb spatial posture, but also the changes in ground contact state in the plantar pressure distribution. This establishes a more direct data correlation between the walking gait phase data and the stiffness adjustment requirements of the softened elastic coupling mechanism.

[0021] Preferably, the specific implementation process of performing coordinate system transformation processing on the lower limb spatial posture to determine the human lower limb pose data under the human basic reference frame is as follows: After the lower limb spatial posture has been formed by multiple lower limb spatial posture data frames, the human basic reference frame configuration record formed by the human object during the wearing initialization phase is first read, and a correspondence is established between the human basic reference frame configuration record and the lower limb spatial posture acquisition and bearing record. The human basic reference frame configuration record is used to express the reference directions of the human object's torso orientation, pelvic lateral orientation, and vertical support direction in the wearing state. The human basic reference frame is not an arbitrary coordinate system in the external environment, but a basic reference system used to describe the human object's own walking motion. The human basic reference frame configuration record is formed by the human object maintaining an upright and static posture during the wearing initialization phase. The torso orientation, pelvic lateral orientation, and vertical support direction in the human basic reference frame configuration record are subsequently used to determine the directional correspondence between the local reference direction of the inertial measurement unit and the human basic reference frame. Subsequently, the lower limb spatial attitude data frames are read frame by frame from the lower limb spatial attitude acquisition and bearing record. Based on the human body basic reference frame configuration record, coordinate system transformation processing is performed on the angular velocity sample value, linear acceleration sample value, and attitude angle sample value in each lower limb spatial attitude data frame. The technical essence of this coordinate system transformation processing is to convert the lower limb motion quantities formed under the local reference direction of the inertial measurement unit into the lower limb motion quantities under the human body basic reference frame, so that the lower limb spatial attitudes of the same human object acquired under different inertial measurement unit installation directions can be processed using the same human motion caliber. In specific implementation, firstly, the directional correspondence between the local reference direction of the inertial measurement unit and the human body basic reference frame is determined based on the installation position mark of the inertial measurement unit; then, based on the directional correspondence, the angular velocity sample values ​​are axially rearranged, the linear acceleration sample values ​​are directionally projected, and the attitude angle sample values ​​are corrected according to the reference direction; finally, the axially rearranged angular velocity sample values, the directionally projected linear acceleration sample values, and the reference direction corrected attitude angle sample values ​​are combined to form a human lower limb pose data frame. The axially rearranged angular velocity sample values ​​are used to express the swing velocity change of the lower limbs of the human object under the human body basic reference frame, the directionally projected linear acceleration sample values ​​are used to express the inertial change of the lower limbs of the human object under the human body basic reference frame, and the reference direction corrected attitude angle sample values ​​are used to express the tilt state of the lower limbs of the human object under the human body basic reference frame; the axially rearranged angular velocity sample values, the directionally projected linear acceleration sample values, and the reference direction corrected attitude angle sample values ​​together constitute the human lower limb pose data frame.Multiple consecutive frames of human lower limb pose data collectively form the human lower limb pose data. This lower limb pose data is subsequently appended to the plantar pressure distribution, enabling the multi-source sensing fusion feature vector to simultaneously retain the spatial posture changes of the lower limbs and the pressure contact changes of the plantar surface. Therefore, the human lower limb pose data serves as an intermediate result in converting the data acquired by the inertial measurement unit (IMU) into a human motion analysis framework, allowing the gait phase data to be determined based on the actual walking motion of the human subject.

[0022] Preferably, the specific implementation process of attaching the human lower limb pose data to the plantar pressure distribution to generate a multi-source sensing fusion feature vector is as follows: After the human lower limb pose data and the plantar pressure distribution both correspond to the current detection period, the human lower limb pose data frame in the human lower limb pose data is read first, and the plantar pressure distribution data frame in the plantar pressure distribution is read simultaneously. Then, according to the sampling time, frame-level correspondence processing is performed on the human lower limb pose data frame and the plantar pressure distribution data frame to form a correspondence between the human lower limb pose data frame and the plantar pressure distribution data frame at the same sampling time. The frame-level correspondence processing is used to eliminate the differences in sampling frequency and sampling time between the inertial measurement unit and the thin-film array sensor, so that when the human lower limb pose data is appended to the plantar pressure distribution, it can be completed under the same current detection cycle and the same sampling time caliber. The correspondence between the human lower limb pose data frame and the plantar pressure distribution data frame is subsequently used to generate a fused feature data frame, so that the fused feature data frame retains the source of both the human lower limb pose data frame and the plantar pressure distribution data frame. Subsequently, using the plantar pressure distribution data frame as the main carrier data of the contact state, the corresponding human lower limb pose data frame is appended after the plantar pressure distribution data frame, thereby forming the fused feature data frame. The main carrier data of the contact state is used to ensure that the node pressure sampling value, the contact node pressure sampling value, the non-contact node pressure sampling value, and the plantar contact migration sequence in the plantar pressure distribution data frame retain their original physical meaning. The main carrier data of the contact state is subsequently combined with the human lower limb pose data frame to form the fused feature data frame. The additional data is not a simple addition of two data points, nor does it change the physical meaning of the node pressure sampling values ​​in the plantar pressure distribution data frame. Instead, it involves adding the posture change field, swing velocity field, and inertia change field from the human lower limb pose data frame to the plantar pressure distribution data frame according to a field arrangement specification. This allows the same fused feature data frame to simultaneously express the plantar contact state and the lower limb spatial posture state. The field arrangement specification is configured after the formation of the human baseline reference frame configuration record. This specification includes the field start position of the plantar pressure distribution data frame, the field addition position of the human lower limb pose data frame, and the field association position of the sampling time. This specification ensures that the fused feature data frame can read the plantar pressure distribution data frame and the human lower limb pose data frame according to fixed field positions when subsequently determining the walking gait phase data. Furthermore, multiple fused feature data frames formed at consecutive sampling times are sequentially concatenated to obtain the multi-source sensing fusion feature vector.The elements in the multi-source sensing fusion feature vector are arranged according to the sampling time, foot region, and lower limb posture fields. The sampling time expresses the sequential relationship of data within the current detection period; the foot region expresses the pressure changes corresponding to the heel, arch, and toe regions; and the lower limb posture field expresses the posture changes, swing speed, and inertial changes in the human lower limb pose data. This multi-source sensing fusion feature vector subsequently serves as a direct input for determining the walking gait phase data, enabling the foot contact migration sequence and the human lower limb pose data to jointly participate in walking gait phase determination. Through this additional method, the multi-source sensing fusion feature vector can maintain the correspondence between contact source and posture source within the same data structure, avoiding the inability to distinguish between dragging contact and normal support contact based solely on the foot pressure distribution, and also avoiding the inability to identify the actual grounding state based solely on the human lower limb pose data.

[0023] Preferably, the specific implementation process for determining the walking gait phase data of the human object in the current detection cycle based on the multi-source sensing fusion feature vector is as follows: After the multi-source sensing fusion feature vector is formed, each fusion feature data frame is read in the order of the sampling time in the multi-source sensing fusion feature vector, and the plantar contact state field corresponding to the plantar pressure distribution data frame and the lower limb posture change field corresponding to the human lower limb pose data frame are extracted from each fusion feature data frame. The plantar contact state field is derived from the plantar pressure distribution, and the lower limb posture change field is derived from the human lower limb pose data. The two types of fields maintain a correspondence at the same sampling time in the fusion feature data frame. The plantar contact state field includes the pressure sampling values ​​of the contact nodes and the pressure sampling values ​​of the non-contact nodes corresponding to the heel region, arch region, and toe region. The lower limb posture change field includes the angular velocity sampling value after axial rearrangement processing, the linear acceleration sampling value after directional projection processing, and the posture angle sampling value after reference direction correction processing. Subsequently, the foot contact state field undergoes contact area migration recognition processing to form a foot contact stage marker. This marker indicates that the human object's foot is in one of three contact stages at the current sampling time: heel contact, full foot bearing, or toe-off transition. The foot contact stage marker, along with the lower limb posture stage marker, participates in the temporal consistency determination. Next, the lower limb posture change field undergoes swing amplitude change recognition processing to form a lower limb posture stage marker. This marker indicates that the human object's lower limb is in one of three posture stages at the current sampling time: stable support, support transition, or swing preparation. This lower limb posture stage marker, along with the foot contact stage marker, participates in the temporal consistency determination. Furthermore, the temporal consistency determination is performed on the foot contact phase marker and the lower limb posture phase marker at the same sampling time. When the foot contact phase marker shows that the heel area begins to bear weight and the lower limb posture phase marker shows that the lower limb swing amplitude decreases, the corresponding sampling time is written into the support start field of the walking gait phase data. When the foot contact phase marker shows that the toe area bears weight decreases and the lower limb posture phase marker shows that the lower limb swing amplitude increases, the corresponding sampling time is written into the support end field of the walking gait phase data. When the foot contact phase marker and the lower limb posture phase marker are both in a continuous bearing state for multiple consecutive sampling times, the corresponding time range is written into the support maintenance field of the walking gait phase data.The resulting walking gait phase data includes a support start field, a support maintenance field, and a support end field. These three fields together express the support phase range of the human object within the current detection cycle. The walking gait phase data then proceeds to the extraction process of leg support phase ground contact features, enabling these features to be formed based on the actual ground contact state and the actual lower limb posture state. This provides a technical basis directly related to the human support phase action for subsequently determining the target pre-compression displacement parameter.

[0024] Optionally, extracting the leg stance phase landing features of the human object based on the walking gait phase data includes: Extract the heel strike time parameter from the walking gait phase data; The toe-off time parameter is extracted from the walking gait phase data; Calculate the duration span from heel contact to toe lift-off based on the heel contact time parameter and the toe lift-off time parameter; Based on the duration span, select motion trajectory slices from the walking gait phase data; The trajectory extreme value change point features are extracted from the motion trajectory slice data, and then the leg support phase landing features of the human object are determined based on the trajectory extreme value change point features.

[0025] Preferably, the specific implementation process of extracting the heel contact time parameter from the walking gait phase data is as follows: After the walking gait phase data has been determined by the multi-source perception fusion feature vector, the support start field, support maintenance field, support end field, foot contact phase marker and lower limb posture phase marker in the walking gait phase data are read first, and a heel contact time extraction bearing record is established. The heel strike moment extraction bearing record is used to incorporate the support phase phase source in the walking gait phase data, the plantar contact source in the plantar contact phase marker, and the lower limb posture source in the lower limb posture phase marker into the same heel strike moment extraction caliber. This allows the subsequently extracted heel strike moment parameters to be traced back to the actual plantar contact changes and actual lower limb posture changes of the human object within the current detection cycle. The support phase phase source is used to express the formation source of the support start field, the support maintenance field, and the support end field in the walking gait phase data. The plantar contact source is used to express the formation source of the plantar contact phase marker in the walking gait phase data. The lower limb posture source is used to express the formation source of the lower limb posture phase marker in the walking gait phase data. The support phase phase source, the plantar contact source, and the lower limb posture source together define the reading range of the heel strike moment extraction bearing record. Subsequently, each fused feature data frame is read in the sampling time sequence of the walking gait phase data, and the contact node pressure sampling value, non-contact node pressure sampling value, and lower limb posture stage marker corresponding to the heel region are identified from the fused feature data frames. The contact node pressure sampling value is used to express the pressure state in which effective surface contact has been formed in the heel region, the non-contact node pressure sampling value is used to express the pressure state in which effective surface contact has not yet been formed in the heel region, and the lower limb posture stage marker is used to express the posture stage of the lower limb of the human object at the current sampling time. Furthermore, the boundary positioning process is performed on the change process of the heel region from the pressure sampling value of the non-contact node to the pressure sampling value of the contact node at continuous sampling times, and the lower limb posture stage marker at the same sampling time is simultaneously checked to see if it has changed from swing preparation to support transition; the boundary positioning process is used to determine the position where the pressure state of the heel region changes continuously in the sampling time sequence, so that the change process of the heel region from the pressure sampling value of the non-contact node to the pressure sampling value of the contact node can participate in the formation of the heel contact candidate time field together with the posture state change of the lower limb posture stage marker.When the pressure change in the heel region corresponds to the posture change of the lower limb posture stage marker within the same sampling time neighborhood, the sampling time is written into the heel contact time extraction bearing record's heel contact candidate time field. The sampling time neighborhood consists of adjacent sampling times in the walking gait phase data, and is used to accommodate the small time difference between the contact node pressure sampling value and the lower limb posture stage marker caused by the sampling order. Subsequently, the heel contact candidate time field undergoes continuous filtering to remove isolated candidate times caused by brief sole friction or localized shoe sole contact, and retains heel contact candidate times that can guide the formation of the support initiation field, thereby forming the heel contact time parameter. The technical essence of the heel-to-ground contact time parameter is the starting time boundary of the transition from swinging preparation to support within the current detection cycle. The heel-to-ground contact time parameter, together with the toe-to-ground departure time parameter, participates in the calculation of the duration span, so that the duration span can be formed around the actual support period of the human body, rather than inferred from a fixed duration or a single posture change. Thus, the heel-to-ground contact time parameter provides a starting time boundary for the subsequent extraction of leg support period landing features, so that the leg support period landing features can correspond to the landing process that the softened elastic coupling mechanism needs to respond to during the human body's support period.

[0026] Preferably, the specific implementation process for extracting the toe-off time parameter from the walking gait phase data is as follows: After the heel-off time parameter has been written into the heel-off time extraction bearing record, the support maintenance field, support end field, plantar contact phase marker, and lower limb posture phase marker in the same walking gait phase data are read again, and a toe-off time extraction bearing record is established. The toe-off time extraction bearing record is used to include the support maintenance process and support end process after the heel-off time parameter into the same toe-off time extraction caliber, so that the toe-off time parameter can maintain the same current detection cycle, the same walking gait phase data source, and the same sampling time caliber as the heel-off time parameter; the toe-off time extraction caliber and the heel-off time extraction caliber are both limited by the sampling time order in the walking gait phase data, so that the toe-off time parameter and the heel-off time parameter can participate in the calculation of the same duration span. Subsequently, each fused feature data frame is read sequentially according to the sampling time from the support maintenance field to the support end field. The contact node pressure sampling value, non-contact node pressure sampling value, and lower limb posture stage marker corresponding to the toe region are identified from the fused feature data frames. The contact node pressure sampling value corresponding to the toe region is used to express the pressure state where the toe region of the human object still forms effective surface contact at the end of the support phase. The non-contact node pressure sampling value corresponding to the toe region is used to express the pressure state where the toe region of the human object has left effective surface contact. The lower limb posture stage marker is used to express whether the lower limb of the human object has transitioned from support to swing preparation. Furthermore, the process of changing pressure in the toe region from the contact node pressure sampling value to the non-contact node pressure sampling value at continuous sampling times is subjected to boundary positioning processing, and the lower limb posture stage marker at the same sampling time is simultaneously checked to see if it transitions from support to swing preparation. This boundary positioning processing is used to determine the location of continuous change in pressure state in the toe region within the sampling time sequence, ensuring that the process of changing pressure in the toe region from the contact node pressure sampling value to the non-contact node pressure sampling value can participate in the formation of the toe-off candidate time field along with the posture state change of the lower limb posture stage marker. When the pressure state change in the toe region corresponds to the posture state change of the lower limb posture stage marker within the same sampling time neighborhood, this sampling time is written into the toe-off candidate time field in the toe-off time extraction bearing record. Subsequently, the toe-off candidate time field undergoes a bearing end filtering process, eliminating isolated candidate times where the toe region is briefly unloaded and then re-bears, while retaining toe-off candidate times that can guide the formation of the support end field, thereby forming the toe-off time parameter.The technical essence of the toe-off time parameter is the end time boundary of the human body transitioning from support to swing preparation within the current detection cycle. The toe-off time parameter is subsequently used in conjunction with the heel-on-the-ground time parameter to calculate the duration span, so that the duration span can cover the complete human support period from the heel touching the ground to the toe-off time. Thus, the toe-off time parameter can define the end time boundary of the motion trajectory slice data, so that the subsequently selected motion trajectory slice data can centrally reflect the lower limb movement changes directly related to the force adjustment of the softened elastic coupling mechanism during the support period.

[0027] Preferably, the specific implementation process for calculating the duration span from heel touchdown to toe liftoff based on the heel touchdown time parameter and the toe liftoff time parameter is as follows: After the heel touchdown time parameter and the toe liftoff time parameter have been formed, a duration span calculation record is first established, and the heel touchdown time parameter, the toe liftoff time parameter, the sampling time sequence in the walking gait phase data, the support start field, the support maintenance field, and the support end field are written into the duration span calculation record. The duration span calculation record is used to include the start time boundary source of the heel touchdown time parameter and the end time boundary source of the toe liftoff time parameter into the same duration span calculation caliber, so that the subsequently formed duration span can correspond to the same human support phase time link; the human support phase time link is used to express the continuous sampling time relationship of the human object from heel touchdown, through support maintenance, to toe liftoff, and the human support phase time link is subsequently used to participate in the selection of motion trajectory slice data through the duration span. Subsequently, the sampling time order of the heel-to-ground contact time parameter and the toe-to-ground departure time parameter is verified. When the heel-to-ground contact time parameter precedes the toe-to-ground departure time parameter, and the sampling time between them covers the support maintenance field, the heel-to-ground contact time parameter is used as the start time boundary of the duration span, and the toe-to-ground departure time parameter is used as the end time boundary of the duration span. When there is an isolated candidate moment corresponding to a short-term pressure jitter between the heel-to-ground contact time parameter and the toe-to-ground departure time parameter, the isolated candidate moment is excluded based on the continuous bearing state in the support maintenance field, so that the duration span is not interrupted by local pressure jitter. Further, the sampling time difference between the end time boundary and the start time boundary is calculated in the duration span calculation bearing record, and this sampling time difference is correlated with the sampling time order in the walking gait phase data to form the duration span. The technical essence of the duration span is the time range within which the human body is in the leg support phase during the current detection cycle. This duration span is not a rough truncation of the complete walking cycle, but rather an effective support phase time period defined by the heel-to-toe contact time parameter and the toe-to-toe lift-off time parameter. This effective support phase time period has the same start and end time boundaries as the human body's support phase timeline. The effective support phase time period is used to describe the effective time range of the duration span within the selected data. The duration span is subsequently used to select motion trajectory slices from the walking gait phase data, ensuring that the motion trajectory slices only retain the lower limb movement state and foot contact state from heel-to-toe contact toe-toe lift-off.By separating the support phase time range from the walking gait phase data through the duration span, the interference of swing phase posture changes on the leg support phase landing characteristics can be reduced, and the subsequent extracted trajectory extreme value change point features can more centrally reflect the motion changes corresponding to landing impact, load transfer and ground transition during the support phase.

[0028] Preferably, the specific implementation process of selecting motion trajectory slice data from the walking gait phase data according to the duration span is as follows: After the duration span has been calculated from the heel contact time parameter and the toe lift-off time parameter, a motion trajectory slice data selection carrier record is first established, and the duration span, the walking gait phase data, the support start field, the support maintenance field, the support end field, and the fusion feature data frame in the multi-source perception fusion feature vector are written into the motion trajectory slice data selection carrier record. The motion trajectory slice data selection carrier record is used to correspond the time range defined by the duration span with the data frame source in the walking gait phase data, so that the subsequently selected motion trajectory slice data can simultaneously retain the start time boundary, the end time boundary, the foot contact state, and the lower limb posture state; the data frame source is used to express the source of the fusion feature data frame in the multi-source perception fusion feature vector, and the data frame source is subsequently used to participate in the source tracing of the motion trajectory slice data through the motion trajectory slice data selection carrier record. Subsequently, according to the start time boundary and the end time boundary of the duration span, the fused feature data frames located within the duration span are read from the walking gait phase data, and the fused feature data frames located outside the duration span are excluded from the motion trajectory slice data. For each fused feature data frame located within the duration span, the plantar contact state field, the lower limb posture change field, and the sampling time field are read respectively; the plantar contact state field is used to express the pressure migration in the heel region, arch region, and toe region during the support phase; the lower limb posture change field is used to express the posture changes, swing speed, and inertial changes of the human lower limbs during the support phase; and the sampling time field is used to maintain the chronological order of each fused feature data frame within the duration span. Further, the fused feature data frames located within the duration span are spliced ​​in time according to the sampling time field, and the spliced ​​fused feature data frames are used as the motion trajectory slice data. The technical essence of the motion trajectory slice data is that it is a segment of local motion data during the support phase extracted from the walking gait phase data. This motion trajectory slice data includes both the contact migration of the sole of the human object during the support phase and the postural changes of the lower limbs during the support phase, enabling it to express a continuous motion trajectory from heel strike, sole bearing, to toe lift-off. The local motion data segment during the support phase is used to define the effective data range of the motion trajectory slice data, and the continuous motion trajectory is used to express the temporal correlation between the various fused feature data frames in the motion trajectory slice data.The motion trajectory slice data is then used to extract the trajectory extreme value mutation point features, so that the trajectory extreme value mutation point features can be formed from the local motion data segments of the support phase, rather than being extracted from the entire walking cycle. This reduces the impact of the swing phase posture changes on the leg support phase landing features.

[0029] Preferably, the specific implementation process of extracting trajectory extreme change point features from the motion trajectory slice data, and then determining the leg support phase landing features of the human object based on the trajectory extreme change point features is as follows: After the motion trajectory slice data has been formed, a trajectory extreme change point feature extraction bearing record is first established, and the foot contact state field, lower limb posture change field, sampling time field, and duration span from the motion trajectory slice data are written into the trajectory extreme change point feature extraction bearing record. The trajectory extreme change point feature extraction record is used to incorporate the support period time source, foot contact source, and lower limb posture source of the motion trajectory slice data into the same trajectory extreme change point feature extraction caliber. This allows the subsequently formed trajectory extreme change point features to correspond to the landing impact, load transfer, and ground transition of the human object during the support period. The support period time source is derived from the duration span, the foot contact source is derived from the foot contact state field, and the lower limb posture source is derived from the lower limb posture change field. The support period time source, the foot contact source, and the lower limb posture source together define the extraction range of the trajectory extreme change point features. Subsequently, each fused feature data frame is read along the sampling time field in the motion trajectory slice data, and the local change amplitude recognition processing is performed on the posture change, swing speed, and inertial change in the lower limb posture change field to form a lower limb trajectory change sequence. Simultaneously, the contact node pressure sampling value and the non-contact node pressure sampling value in the foot contact state field are subjected to contact migration amplitude recognition processing to form a foot contact change sequence. The lower limb trajectory change sequence is used to express the movement trajectory changes of the lower limbs during the support phase, and the foot contact change sequence is used to express the load-bearing migration changes in the foot region during the support phase. Both the lower limb trajectory change sequence and the foot contact change sequence participate in the trajectory extreme value mutation identification processing. Further, within the duration span, local extreme points in the lower limb trajectory change sequence that transition from decreasing to stable, from stable to increasing, or from increasing to decreasing are identified, and these local extreme points are correlated with the foot contact change sequence near the same sampling time. These local extreme points express the position of the change direction transformation in the lower limb trajectory change sequence, and the correlation verification is used to temporally correlate the local extreme points with the load-bearing migration changes in the foot contact change sequence. When the local extreme point corresponds to increased contact in the heel region, transitional load-bearing in the arch region, or decreased load-bearing in the toe region, the local extreme point is written into the trajectory extreme value mutation point feature.The technical essence of the trajectory extreme value mutation point feature is that the trajectory point information in the motion trajectory slice data can reflect the change of the landing state during the support phase. The trajectory extreme value mutation point feature is not simply a maximum or minimum value, but a support phase point feature that simultaneously satisfies the correspondence between the lower limb trajectory change sequence and the foot contact change sequence. Subsequently, based on the positional order of the trajectory extreme value mutation point feature within the duration span, the support phase landing process of the human object is marked in stages; wherein, the trajectory extreme value mutation point feature near the heel contact time parameter is used to express the landing impact initiation state, the trajectory extreme value mutation point feature located in the support maintenance field is used to express the load transfer state, and the trajectory extreme value mutation point feature near the toe lift-off time parameter is used to express the lift-off transition state. The staged markers are used to sequentially convert the positional characteristics of the trajectory extreme value mutation points within the duration span into the landing impact initiation state, load transfer state, and ground transition state during the support phase landing process. The staged markers and the trajectory extreme value mutation points are written together into the leg support phase landing feature record to form the leg support phase landing feature of the human object. The leg support phase landing feature record carries the staged markers and the trajectory extreme value mutation points, allowing the leg support phase landing feature of the human object to retain the positional and stage origins of the trajectory extreme value mutation points. The leg support phase landing feature subsequently participates in the determination of the target pre-compression displacement parameter along with the load condition level, enabling the target pre-compression displacement parameter to simultaneously reflect the landing state of the human object during the support phase and the load state of the wearable centaur robot, thereby allowing the equivalent physical stiffness update of the softened elastic coupling mechanism to conform to the landing process on the human support side.

[0030] Optionally, determining the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material includes: Read the baseline unloaded weight value of the wearable centaur robot when it is in an unloaded state; Based on the weight data of the load-bearing material and the baseline unloaded weight value, calculate the weight increment of the wearable Centaur robot. The weight increment value is compared with the load critical node parameters of multiple levels one by one to locate the load matching interval into which the weight increment value falls. The load condition level of the wearable centaur robot is generated based on the identifier code corresponding to the load matching interval.

[0031] Preferably, the specific implementation process for reading the reference empty weight value of the wearable Centaur robot in an unloaded state is as follows: After the weight data of the carrying material has been mapped to the current detection cycle, the empty weight calibration load record formed by the wearable Centaur robot during the assembly calibration stage is first read, and the reference empty weight value is read from the empty weight calibration load record. The empty weight calibration load record is used to record the weight source of the wearable Centaur robot when it is not loaded with carrying material, has no external handling load, and all basic structural components are in a normal installation state; wherein, the external handling load is an external weight source that, in addition to the carrying material, would change the overall weight state of the wearable Centaur robot, and the external handling load is excluded when the unloaded state is formed, so that the empty weight calibration load record can only express the structural weight of the wearable Centaur robot itself. The empty weight calibration load record includes a main frame weight field, a softened elastic coupling mechanism weight field, a micro screw actuator component weight field, a stiffness adjustment controller weight field, and a wearable connection accessory weight field. The main frame weight field expresses the weight contribution of the main frame itself to the reference unloaded weight value. The softened elastic coupling mechanism weight field expresses the weight contribution of the rhomboid linkage and the antagonistic spring assembly to the reference unloaded weight value. The micro-screw actuator weight field expresses the weight contribution of the micro-screw actuator to the reference unloaded weight value. The stiffness adjustment controller weight field expresses the weight contribution of the stiffness adjustment controller to the reference unloaded weight value. The wearable connection accessory weight field expresses the weight contribution of the fixed wearing structure when the human body is connected to the main frame to the reference unloaded weight value. The fixed wearing structure is the structural source corresponding to the wearable connection accessory weight field, and the weight contribution of the fixed wearing structure enters the reference unloaded weight value through the wearable connection accessory weight field. Furthermore, before reading the reference unloaded weight value, the weight fields of the main frame, the softened elastic coupling mechanism, the micro-screw actuator, the stiffness adjustment controller, and the wearable connection accessory in the unloaded weight calibration load record are subjected to field integrity verification processing to form an unloaded weight field verification result. When the unloaded weight field verification result shows that the weight fields of the main frame, the softened elastic coupling mechanism, the micro-screw actuator, the stiffness adjustment controller, and the wearable connection accessory are all in a readable state, the weight fields of the main frame, the softened elastic coupling mechanism, the micro-screw actuator, the stiffness adjustment controller, and the wearable connection accessory are written into the reference unloaded weight reading load record according to the same weight unit caliber.The reference empty weight reading record is used to include the structural weight source of the wearable Centaur robot in an unloaded state within the same reading caliber, so that when calculating the weight increment value, the weight data of the load-bearing material and the reference empty weight value can be processed under the same weight reference. The technical essence of the wearable Centaur robot being in an unloaded state is that the wearable Centaur robot is in a calibration state that includes only its own structural weight and does not include the weight of the load-bearing material and the weight of the external handling load. The unloaded state is not any frame of weight data during operation, but a reference state formed during the assembly calibration stage and used as the basis for calculating the weight difference in the subsequent current detection cycle. The baseline unloaded weight value read therefrom is subsequently used in conjunction with the weight data of the carrying material to calculate the weight increment value. This allows the weight increment value to represent the additional load introduced by the carrying material, rather than misjudging the weight of the wearable Centaur robot's own structure as an external load. This provides a basic weight reference for subsequently determining the load condition level of the wearable Centaur robot. The basic weight reference is expressed by the baseline unloaded weight value and continues to be used as the difference calculation benchmark in the subsequent calculation of the weight increment value.

[0032] Preferably, the specific implementation process for calculating the weight increment of the wearable Centaur robot based on the load-bearing material counterweight data and the baseline empty weight value is as follows: After the baseline empty weight value has been written into the baseline empty weight reading load record, the load-bearing material counterweight data corresponding to the current detection cycle is first read, and a weight increment calculation load record is established. The weight increment calculation load record is used to include the load-bearing material counterweight data, the baseline empty weight value, the current detection cycle, and the subsequently formed weight increment value into the same calculation caliber, so that the weight increment value can be traced back to the load-bearing material changes within the current detection cycle. The load-bearing material counterweight data is used to express the overall weight state of the wearable Centaur robot after carrying material within the current detection cycle, and the baseline empty weight value is used to express the self-structural weight state of the wearable Centaur robot when it is in an unloaded state; the overall weight state includes the self-structural weight state and the additional load corresponding to the load-bearing material changes, the self-structural weight state is entered into the weight increment calculation load record through the baseline empty weight value, and the load-bearing material changes are entered into the weight increment calculation load record through the load-bearing material counterweight data. After the load-bearing material counterweight data and the baseline empty weight value are entered into the same weight increment value calculation record, a weight unit consistency processing is first performed on the load-bearing material counterweight data and the baseline empty weight value to obtain weight unit consistent load-bearing material counterweight data and weight unit consistent baseline empty weight value. The weight unit consistency processing is used to eliminate differences in weight unit caliber between the load-bearing material counterweight data and the baseline empty weight value in the data acquisition source or storage source, avoiding inconsistencies in dimensions during subsequent weight difference calculations. The data acquisition source indicates the origin of the load-bearing material counterweight data within the current detection cycle, and the storage source indicates the origin of the baseline empty weight value stored in the empty weight calibration load record. Both the data acquisition source and the storage source are unified to the same weight unit caliber through the weight unit consistency processing. Subsequently, a weight difference calculation processing is performed on the weight unit consistent load-bearing material counterweight data and the weight unit consistent baseline empty weight value to form the initial weight increment value. The initial weight increment value is used to express the incremental change of the overall weight state relative to the no-load state within the current detection cycle. The initial weight increment value is further written into the weight increment value calculation load record and a corresponding relationship is established with the current detection cycle.Furthermore, a non-negativity check is performed on the initial weight increment value. When the initial weight increment value is non-negative, it is established as the weight increment value. When the initial weight increment value is lower than the allowable fluctuation range corresponding to the no-load state, the initial weight increment value is zeroed based on the reference no-load weight value in the no-load weight calibration record, and the zeroed initial weight increment value is established as the weight increment value. The allowable fluctuation range is configured according to the no-load weight calibration record during the assembly calibration stage. The allowable fluctuation range is used to absorb the small acquisition fluctuations of the structural weight state under the no-load state. The allowable fluctuation range subsequently participates in the non-negativity check process, ensuring that the initial weight increment value is not interpreted as a negative weight difference due to small acquisition fluctuations. The non-negativity check process is used to prevent negative weight differences caused by sensor drift or short-term unloading from entering the load condition level determination process, ensuring that the weight increment value maintains the expression of the additional load introduced to the load-bearing material. The technical essence of the weight increment value is the increase in load relative to the baseline empty weight value of the carrying material's counterweight data. This weight increment value is subsequently compared one by one with load critical node parameters at multiple levels, allowing the load condition level of the wearable Centaur robot to be formed based on the additional load introduced by the carrying material, rather than directly based on the total weight of the entire machine or a fixed weight level. The total weight of the entire machine refers to the overall weight state before deducting the baseline empty weight value, and this total weight is not directly included in the load condition level determination process. Through this processing, the weight increment value transforms the carrying material counterweight data into a data object that can participate in interval determination, providing a numerical basis for subsequently locating the load matching interval into which the weight increment value falls. This numerical basis is expressed by the weight increment value and continues to serve as a data source for comparison during the subsequent load matching interval positioning process.

[0033] Preferably, the specific implementation process of comparing the weight increment value with multiple levels of load critical node parameters to locate the load matching interval into which the weight increment value falls is as follows: After the weight increment value has been calculated from the weight data of the load-bearing material and the baseline no-load weight value, the load critical node parameter configuration load record is first read, and multiple levels of load critical node parameters are read from the load critical node parameter configuration load record. The load critical node parameter configuration load record is formed when the wearable centaur robot completes structural calibration or load capacity calibration. The load critical node parameter configuration load record is used to record the weight boundary source corresponding to different load ranges; the structural calibration is used to form the load range of the main frame and the elastic coupling range of the softened elastic coupling mechanism, and the load capacity calibration is used to form the adjustment range in which the micro-screw actuator can drive the antagonistic spring assembly to change the initial deformation scale. The load range of the main frame, the elastic coupling range of the softened elastic coupling mechanism, and the adjustment range in which the micro-screw actuator can drive the antagonistic spring assembly to change the initial deformation scale all participate in the configuration of the load critical node parameters. The load critical node parameters at multiple levels are arranged in ascending order of weight increment, with each pair of adjacent load critical node parameters defining a load matching interval. These load critical node parameters are not arbitrary thresholds, but rather weight boundary data configured based on the load-bearing range of the main frame, the elastic coupling range of the softened elastic coupling mechanism, and the adjustment range by which the micro-screw actuator can drive the antagonistic spring assembly to change its initial deformation scale. This ensures that the subsequently formed load matching interval corresponds to the stiffness adjustment capability of the softened elastic coupling mechanism. The weight boundary data is reflected by the load critical node parameters, forming the load matching interval between adjacent load critical node parameters. Subsequently, a load matching interval positioning load record is established, and the weight increment value, the multiple levels of load critical node parameters, and the candidate load matching intervals formed by adjacent load critical node parameters are written into the load matching interval positioning load record. The candidate load matching intervals are used to bear the range into which the weight increment value can fall before comparison. These candidate load matching intervals are subsequently selected as the load matching intervals through a step-by-step comparison process. Furthermore, according to the arrangement order of the load critical node parameters of the multiple levels from low to high, the weight increment value is compared with each of the load critical node parameters. When the weight increment value is not lower than the load critical node parameter of the previous level and is lower than the load critical node parameter of the next level, the candidate load matching interval between the load critical node parameter of the previous level and the load critical node parameter of the next level is located as the load matching interval into which the weight increment value falls.If the weight increment value is the same as a certain load critical node parameter, then according to the boundary attribution caliber configured in the load critical node parameter configuration load record, the weight increment value is assigned to the candidate load matching interval corresponding to the adjacent higher level or the adjacent lower level. The boundary attribution caliber is synchronously written into the load critical node parameter configuration load record when configuring the load critical node parameter, and the boundary attribution caliber is used to eliminate interval attribution ambiguity when the weight increment value is at the boundary position. The load matching interval obtained in this way is used to express the load range formed by the wearable Centaur robot carrying materials in the current detection cycle. The load matching interval is subsequently mapped to the corresponding identification code, so that the load condition level can be converted from the continuously changing weight increment value into condition level data that can participate in stiffness adjustment judgment. By comparing the weight increment value with the load critical node parameters of the multiple levels one by one, this application avoids directly using unranged weight data to drive subsequent stiffness adjustment, so that the load condition level can maintain stable data expression under different load material weight changes; the unranged weight data belongs to the weight increment value that has not undergone the load matching interval positioning process, and the unranged weight data will not directly enter the identification code reading process.

[0034] Preferably, the specific implementation process for generating the load condition level of the wearable Centaur robot based on the identifier code corresponding to the load matching interval is as follows: After the load matching interval has been located by comparing the weight increment value with the load critical node parameters of the multiple levels one by one, the load interval code mapping load record is first read, and the identifier code corresponding to the load matching interval is read from the load interval code mapping load record. The load interval code mapping load record is formed synchronously when configuring the load critical node parameters. The load interval code mapping load record is used to record the correspondence between each load matching interval and its respective identifier code, so that the load matching interval can be converted into load condition level data that can be directly read in the subsequent stiffness adjustment process. The identifier code is not used to describe the classification information of the load-bearing material, but to express the load interval state after the weight increment value falls into the corresponding load matching interval; the classification information does not participate in the formation of the load condition level, and the load interval state enters the load condition level through the identifier code to generate the load record. For example, in a general configuration, the lower weight increment range, the medium weight increment range, and the higher weight increment range can each be configured with different identifier codes. The specific encoding format can be consecutive integers, segmented characters, or stored field numbers, as long as each identifier code maintains a unique correspondence with each load matching range. The lower weight increment range, the medium weight increment range, and the higher weight increment range are all hierarchical representations of the load matching ranges. Each of these ranges enters the load condition level to generate a load-bearing record through its corresponding identifier code. Subsequently, the load matching range and the identifier code are written into the load condition level to generate a load-bearing record, and a correspondence verification process is performed between the load matching range and the identifier code to form an identifier code verification result. This identifier code verification result is used to confirm whether the currently read identifier code comes from the code entry in the load range code mapping load-bearing record that corresponds to the load matching range. The identifier code verification result subsequently participates in the formation of the load condition level, preventing code entries that do not correspond to the load matching range from entering the subsequent stiffness adjustment process. Furthermore, when the identification code verification result shows that the load matching interval corresponds to the identification code, the identification code is written into the load condition level field in the load condition level generation record, and the load condition level field is established as the load condition level of the wearable Centaur robot. The load condition level field is used to carry the identification code confirmed by the identification code verification result, and the load condition level field subsequently serves as the field carrier of the load condition level in determining the target pre-compression displacement parameter.The technical essence of the load condition level is discrete load condition data mapped from the load matching interval of the weight increment value. The load condition level is used to transform the continuously changing weight of the load-bearing material into a load state that can be directly read and matched when determining the target pre-compression displacement parameter. The load condition level subsequently participates in the determination of the target pre-compression displacement parameter together with the leg support phase grounding characteristics, enabling the target pre-compression displacement parameter to respond not only to the human object's support phase grounding state but also to the load-bearing state of the wearable centaur robot within the current detection cycle. Through the conversion from the load matching interval to the identification code and then to the load condition level, the continuous weight change of the load-bearing material counterweight data is organized into load condition level data adapted to the stiffness adjustment process of the softened elastic coupling mechanism, thereby providing data basis on the robot load side for subsequent updates to the equivalent physical stiffness of the softened elastic coupling mechanism. This data basis on the robot load side is expressed by the load condition level and subsequently participates in the determination of the target pre-compression displacement parameter together with the leg support phase grounding characteristics.

[0035] Optionally, determining the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism based on the load condition level and the leg support period ground contact characteristics includes: Extract the basic connection stiffness parameters corresponding to the load condition level from the multidimensional stiffness mapping association library; Extract the peak landing impact value from the leg support phase landing characteristics; Calculate the ratio between the peak impact value and the standard flat ground walking impact reference value to obtain the stiffness dynamic adjustment compensation factor; The stiffness parameters of the basic connection are numerically corrected using the stiffness dynamic adjustment compensation factor to determine the target required stiffness value. Based on the target stiffness value and the inherent elastic coefficient of the antagonistic spring assembly, the corresponding displacement conversion value is calculated, and then the target pre-compression displacement parameter of the antagonistic spring assembly is determined based on the displacement conversion value.

[0036] Preferably, the specific implementation process for extracting the basic connection stiffness parameters corresponding to the load condition level in the multidimensional stiffness mapping association library is as follows: After the load condition level has been generated by the identifier code corresponding to the load matching interval, a basic connection stiffness parameter extraction load record is first established, and the load condition level, the condition level field, the identifier code, the load matching interval, and the weight increment value are written into the basic connection stiffness parameter extraction load record. The basic connection stiffness parameter extraction load record is used to include the encoding source, interval source, and weight source of the load condition level into the same reading caliber, so that the subsequently extracted basic connection stiffness parameters can be traced back to the load state of the wearable Centaur robot in the current detection cycle; the encoding source is expressed by the identifier code, the interval source is expressed by the load matching interval, and the weight source is expressed by the weight increment value. The encoding source, the interval source, and the weight source together define the retrieval basis of the basic connection stiffness parameter extraction load record. Subsequently, the multidimensional stiffness mapping association library is read. The multidimensional stiffness mapping association library is a stiffness data carrying structure formed after the wearable centaur robot completes structural calibration, unloaded weight calibration, and load critical node parameter configuration. This multidimensional stiffness mapping association library is not a single stiffness table, but rather a data mapping relationship formed by placing the load condition level, the load matching interval, the elastic coupling range of the softened elastic coupling mechanism, the force transmission state of the rhomboid linkage, and the pre-compressible range of the antagonistic spring assembly under the same search caliber. The force transmission state is used to express the force transmission caliber of the rhomboid linkage when transmitting human traction force and robot load reaction force in the softened elastic coupling mechanism. The pre-compressible range is used to express the displacement adjustment range that the antagonistic spring assembly can form by changing its initial deformation scale. The force transmission state and the pre-compressible range jointly participate in the configuration of the basic connection stiffness parameters in the multidimensional stiffness mapping association library. Each mapping record in the multidimensional stiffness mapping association library includes at least a load condition level field, a load matching interval field, a basic connection stiffness parameter field, and an adjustable state field. The load condition level field is used to read the corresponding load condition level in the basic connection stiffness parameter extraction bearing record. The load matching interval field is used to verify whether the load condition level still corresponds to the load matching interval into which the current weight increment value falls. The basic connection stiffness parameter field is used to store the basic connection stiffness parameter corresponding to the load condition level. The adjustable state field is used to express whether the basic connection stiffness parameter is within the adjustment range that the softened elastic coupling mechanism can achieve by changing the initial deformation scale through the antagonistic spring assembly.Further, the multidimensional stiffness mapping association library is matched and read according to the load condition level field to obtain candidate stiffness mapping records. Then, the load matching interval field is used to perform interval consistency verification on the candidate stiffness mapping records to extract the basic connection stiffness parameters from the candidate stiffness mapping records. The candidate stiffness mapping records are used to carry the mapping records to be verified that match the load condition level field. After the interval consistency verification, the candidate stiffness mapping records can retain the basic connection stiffness parameters that are consistent with the load matching interval. The technical essence of the basic connection stiffness parameters is that, when only considering the load-side bearing state of the wearable centaur robot, the softened elastic coupling mechanism should present an initial connection stiffness reference value. The basic connection stiffness parameters are not directly used as the final adjustment target, but continue to be subject to numerical correction by the stiffness dynamic adjustment compensation factor so that the target required stiffness value can simultaneously include the wearable centaur robot load-side information expressed by the load condition level and the human support-side information expressed by the leg support period ground landing characteristics. Through this extraction method, the source of the basic connection stiffness parameter maintains a corresponding relationship with the load condition level, avoiding the use of the same fixed connection stiffness reference value between light load, medium load and high load, thereby providing a load-side stiffness basis for the subsequent determination of the target pre-compression displacement parameter; the load-side stiffness basis is expressed by the basic connection stiffness parameter, and continues to participate in the numerical correction process in the subsequent determination of the target required stiffness value.

[0037] Preferably, the specific implementation process for extracting the landing impact peak from the leg support phase landing characteristics is as follows: After the basic connection stiffness parameters have been written into the basic connection stiffness parameter extraction load record, the leg support phase landing characteristics of the human object are read, and a landing impact peak extraction load record is established. The landing impact peak extraction load record is used to include the leg support phase landing characteristics, the trajectory extreme value change point characteristics, the staged markers, the heel contact time parameter, the toe lift-off time parameter, and the duration span into the same extraction caliber, so that the subsequently formed landing impact peak can be traced back to the human object's support phase landing process within the current detection period. The leg support phase landing characteristics include the trajectory extreme value change point characteristics and the staged markers; wherein, the trajectory extreme value change point characteristics are used to express the trajectory point information in the motion trajectory slice data that can reflect the change in the support phase landing state, and the staged markers are used to convert the position sequence of the trajectory extreme value change point characteristics within the duration span into the landing impact initiation state, load transfer state, and lift-off transition state. Subsequently, the trajectory extreme value mutation point features are read according to the sampling time sequence within the duration span, and trajectory point information corresponding to the ground impact initiation state is filtered from the trajectory extreme value mutation point features. The ground impact initiation state is close to the heel contact time parameter, and the ground impact initiation state is used to express the impact change on the support side caused when the heel area of ​​the human object transitions from a non-contact state to an effective bearing state; the impact change on the support side continues to be processed by the trajectory point information to read the impact amplitude, so that the ground impact peak can originate from the actual grounding process at the beginning of the human object's support period. Further, the trajectory point information corresponding to the initial state of the ground impact is processed by impact amplitude reading to read the lower limb posture change field, the sole contact state field, and the contact node pressure sampling value at the same or adjacent sampling time as the trajectory point information. The lower limb posture change field is used to express the inertial change of the lower limb of the human body in the initial state of the ground impact, the sole contact state field is used to express the contact enhancement state of the heel area in the initial state of the ground impact, and the contact node pressure sampling value is used to express the local pressure degree of the heel area in the initial state of the ground impact. Then, based on the inertial change in the lower limb posture change field, the contact enhancement state of the heel area in the sole contact state field, and the change amplitude of the contact node pressure sampling value, a candidate ground impact peak value is formed. The candidate ground impact peak value is not an arbitrary maximum sole pressure value, but a candidate impact amplitude that simultaneously originates from the initial state of the ground impact, the trajectory extreme value change point characteristics, and the contact enhancement state of the heel area.Subsequently, the candidate ground impact peak values ​​are verified for their support period range. Candidate ground impact peak values ​​that fall within the duration span and correspond to the support start field are retained, and these retained candidate ground impact peak values ​​are established as the ground impact peak values. The technical essence of the ground impact peak value is the impact intensity parameter on the support side caused by the heel area landing at the beginning of the leg support phase. The ground impact peak value is then compared with the standard flat-ground walking impact benchmark value to form the stiffness dynamic adjustment compensation factor. Thus, the ground impact peak value enables the stiffness adjustment of the softened elastic coupling mechanism to respond not only to the load-bearing state of the wearable centaur robot but also to the ground impact changes of the human object during the current leg support phase, providing data basis for the human support side of the subsequent numerical correction of the basic connection stiffness parameters. This data basis on the human support side is expressed by the ground impact peak value and continues to be used as the ratio calculation object in the subsequent calculation of the stiffness dynamic adjustment compensation factor.

[0038] Preferably, the specific implementation process for calculating the ratio between the peak impact value and the standard flat-ground walking impact reference value to obtain the stiffness dynamic adjustment compensation factor is as follows: After the peak impact value has been written into the peak impact value extraction load record, the standard flat-ground walking impact reference value is read, and a stiffness dynamic adjustment compensation factor calculation load record is established. The standard flat-ground walking impact reference value is the support period ground impact reference value formed when the human object is in a stable flat-ground walking state. The standard flat-ground walking impact reference value is used to express the ground impact reference caliber when the human object does not experience obstacle crossing swaying, does not experience high load traction disturbance, and has a relatively smooth gait support transition. The standard flat-ground walking impact benchmark value can be formed during the wear initialization phase or the initial flat-ground trial phase. The configuration is as follows: Within multiple consecutive current detection cycles, the landing impact peak value in the leg support phase landing characteristics is read, and multiple landing impact peak values ​​in a stable flat-ground walking state are subjected to benchmark filtering processing to form the standard flat-ground walking impact benchmark value. The wear initialization phase is used to establish the basic wear state between the human object and the wearable centaur robot. The initial flat-ground trial phase is used to collect multiple landing impact peak values ​​in a stable flat-ground walking state. Both the wear initialization phase and the initial flat-ground trial phase provide traceable data sources for the standard flat-ground walking impact benchmark value. The benchmark filtering processing is used to remove abnormally high or low landing impact peak values, enabling the standard flat-ground walking impact benchmark value to serve as a reference for the landing impact peak values ​​in subsequent current detection cycles. Subsequently, the peak impact value and the standard flat-ground walking impact reference value are written into the stiffness dynamic adjustment compensation factor calculation load record, and a dimensional consistency check is performed on the peak impact value and the standard flat-ground walking impact reference value. The dimensional consistency check confirms that both the peak impact value and the standard flat-ground walking impact reference value use the same impact intensity caliber, preventing dimensional misuse issues in subsequent ratio calculations. The impact intensity caliber expresses that both the peak impact value and the standard flat-ground walking impact reference value belong to the impact intensity parameter during the support period, and this impact intensity caliber is subsequently used in the ratio calculation of the stiffness dynamic adjustment compensation factor. Further, after the dimensional consistency check passes, the ratio between the peak impact value and the standard flat-ground walking impact reference value is calculated, and this ratio is written into the stiffness dynamic adjustment compensation factor calculation load record to form the stiffness dynamic adjustment compensation factor.The technical essence of the stiffness dynamic adjustment compensation factor is a dimensionless adjustment parameter that normalizes the ground impact intensity of the human object within the current detection period relative to the ground impact benchmark under stable walking conditions on flat ground. When the peak ground impact is higher than the standard flat walking impact benchmark, the stiffness dynamic adjustment compensation factor indicates a higher ground impact during the current leg support phase of the human object; when the peak ground impact is close to the standard flat walking impact benchmark, the stiffness dynamic adjustment compensation factor indicates a ground impact during the current leg support phase of the human object is close to that of stable walking conditions on flat ground. The stiffness dynamic adjustment compensation factor is subsequently used to numerically correct the basic connection stiffness parameters, enabling the basic connection stiffness parameters extracted from the load condition level to incorporate the dynamic impact changes in the leg support phase ground characteristics. The landing characteristics during the leg support period are converted into an adjustment parameter that can be numerically corrected with the basic connection stiffness parameter by calculating the ratio between the peak impact value and the standard flat ground walking impact benchmark value. This allows the subsequent target stiffness value to take into account the changes in the load side and the human support side of the wearable Centaur robot. The adjustment parameter is expressed by the stiffness dynamic adjustment compensation factor and continues to participate in the numerical correction process of the basic connection stiffness parameter during the subsequent determination of the target stiffness value.

[0039] Preferably, the specific implementation process of using the stiffness dynamic adjustment compensation factor to numerically correct the basic connection stiffness parameter to determine the target required stiffness value is as follows: After the stiffness dynamic adjustment compensation factor has been calculated from the ratio between the peak ground impact and the standard flat ground walking impact benchmark value, a target required stiffness value determination load record is first established, and the basic connection stiffness parameter, the stiffness dynamic adjustment compensation factor, the load condition level, the leg support period ground landing characteristics, and the peak ground impact are written into the target required stiffness value determination load record. The target stiffness value determination record is used to incorporate the wearable Centaur robot load-side source of the basic connection stiffness parameter and the support-side source of the stiffness dynamic adjustment compensation factor into the same stiffness correction caliber. This ensures that the subsequently formed target stiffness value can simultaneously correspond to the load state of the wearable Centaur robot and the ground impact state of the human object in the current detection cycle. The wearable Centaur robot load-side source is expressed by the load condition level and the basic connection stiffness parameter, while the support-side source is expressed by the leg support period ground characteristics, the ground impact peak value, and the stiffness dynamic adjustment compensation factor. Subsequently, the basic connection stiffness parameter is processed to read the adjustable state field corresponding to the basic connection stiffness parameter in the multidimensional stiffness mapping association library, and the range within which the basic connection stiffness parameter can be corrected is limited based on the adjustable state field. The adjustable state field is used to express the range of connection stiffness that the softened elastic coupling mechanism can achieve when the initial deformation scale is changed by the antagonistic spring assembly. The adjustable state field subsequently participates in the boundary processing of the target required stiffness value, ensuring that the target required stiffness value does not deviate from the actual adjustment capabilities of the antagonistic spring assembly and the micro-screw actuation assembly. The connection stiffness range is used to limit the allowable variation range of the target required stiffness value. The connection stiffness range is expressed by the adjustable state field and continues to participate in subsequent adjustable range matching processing. Further, the stiffness dynamic adjustment compensation factor is applied to the basic connection stiffness parameter for numerical correction to form the initial required stiffness value. The initial required stiffness value is used to express the stiffness requirement after the impact change of the human object's current leg support period, superimposed on the basic connection stiffness parameter corresponding to the load condition level.Subsequently, an adjustable range matching process is performed on the initial required stiffness value. When the initial required stiffness value is within the connection stiffness range defined by the adjustable state field, the initial required stiffness value is established as the target required stiffness value. When the initial required stiffness value exceeds the connection stiffness range defined by the adjustable state field, boundary constraint processing is performed on the initial required stiffness value according to the boundary stiffness value corresponding to the adjustable state field, and the initial required stiffness value after boundary constraint processing is established as the target required stiffness value. The boundary stiffness value is used to express the upper or lower stiffness boundary of the connection stiffness range. The boundary stiffness value participates in the formation of the target required stiffness value through the boundary constraint processing, so that the target required stiffness value can correspond to the adjustable capability of the softened elastic coupling mechanism. The technical essence of the target required stiffness value is that the connection stiffness target is obtained by dynamically correcting the basic connection stiffness parameters given at the load condition level based on the landing impact peak value in the leg support period landing characteristics. The target required stiffness value is subsequently used in conjunction with the inherent elastic coefficient of the antagonistic spring assembly in the calculation of the displacement conversion value, enabling the target required stiffness value to be further converted into a displacement adjustment target that the antagonistic spring assembly can perform. Through this numerical correction method, the basic connection stiffness parameter no longer remains a static stiffness reference determined solely by the load condition level, but further incorporates the impact change of the human body's leg support period, thereby allowing the target pre-compression displacement parameter to be formed around both the current load state and the current human body support period state.

[0040] Optionally, the step of calculating the corresponding displacement conversion value based on the target required stiffness value and the inherent elastic coefficient of the antagonistic spring assembly, and then determining the target pre-compression displacement parameter of the antagonistic spring assembly based on the displacement conversion value, includes: Determine the four-bar included angle parameter of the rhomboid linkage transmission component at the current operating moment; Based on the included angle parameter of the four-bar linkage, the nonlinear transmission coefficient of the rhomboid linkage transmission component for physical forces is determined; The target stiffness requirement value is converted using the nonlinear transmission coefficient to determine the converted stiffness requirement value transmitted to the end of the antagonistic spring assembly. Calculate the conversion quotient between the reduced stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly, and then determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the conversion quotient.

[0041] Preferably, the specific implementation process for determining the four-bar angle parameter of the rhombic linkage transmission component at the current operating moment is as follows: After the target required stiffness value has been determined by the basic connection stiffness parameters and the stiffness dynamic adjustment compensation factor, a load-bearing record for determining the four-bar angle parameter is first established. The target required stiffness value, the softened elastic coupling mechanism, the rhombic linkage transmission component, the antagonistic spring assembly, the current operating moment, and the previously formed load-bearing record for determining the target required stiffness value are written into this load-bearing record. This load-bearing record for determining the four-bar angle parameter is used to include the stiffness source of the target required stiffness value, the mechanism source of the rhombic linkage transmission component, the execution object source of the antagonistic spring assembly, and the triggering source of the current operating moment into the same determining caliber, so that the subsequently formed four-bar angle parameter can be traced back to the specific time point when the target required stiffness value is converted into the target pre-compression displacement parameter. The current operating time is used to express the specific control point within the current detection cycle at which the target required stiffness value needs to be converted into the target pre-compression displacement parameter of the antagonistic spring assembly. The current operating time is not an arbitrary point in time, but rather the displacement conversion trigger point after the walking gait phase data, the load-bearing material counterweight data, the leg support phase ground landing characteristics, and the load condition level have all participated in the formation of the target required stiffness value. The displacement conversion trigger point is then stored as part of the current operating time in the four-bar linkage angle parameter determination load record. Subsequently, the linkage attitude acquisition record of the rhomboid linkage transmission component at the current operating time is read. The linkage attitude acquisition record is used to save the node position relationship of each linkage node in the rhomboid linkage transmission component at the current operating time. The node position relationship originates from the connection attitude between the rhomboid linkage transmission component and the human body object, the main skeleton, and the softened elastic coupling mechanism. The connection attitude is used to express the actual connection state formed by the rhomboid linkage transmission component between the human body object and the main skeleton. The linkage attitude acquisition record includes at least a human body-side connection node position field, a main skeleton-side connection node position field, a first oblique link node position field, and a second oblique link node position field. The human body-side connection node position field is used to express the position of one side node of the rhomboid linkage transmission component connected to the human body object. The main skeleton-side connection node position field is used to express the position of one side node of the rhomboid linkage transmission component connected to the main skeleton. The first oblique link node position field and the second oblique link node position field are used to express the positions of two oblique link nodes that form a rhomboid transmission posture inside the rhomboid linkage transmission component.Further, based on the human body-side connection node position field, the main skeleton-side connection node position field, the first oblique link node position field, and the second oblique link node position field, link direction data corresponding to adjacent links of the rhombic link transmission component are constructed respectively, and the link direction data is written into the four-link included angle parameter determination load record. The link direction data is used to express the actual geometric unfolding state of the rhombic link transmission component at the current working moment. The actual geometric unfolding state is used to express the degree of link unfolding of the rhombic link transmission component formed by the node position relationship. The link direction data continues to participate in the included angle calculation process to form the four-link included angle parameter. Specifically, the direction difference reading processing of adjacent link direction data is performed to obtain the four-link included angle parameter of the rhombic link transmission component at the current working moment. The technical essence of the four-bar included angle parameter is the geometric force transmission angle data formed by the positional relationship of each node in the rhombic link transmission component at the current operating moment. This geometric force transmission angle data is used to express the angular influence of the actual geometric deployment state of the rhombic link transmission component on the subsequent force transmission conversion. The four-bar included angle parameter is subsequently used to determine the nonlinear transmission coefficient of the rhombic link transmission component for physical forces, ensuring that the target required stiffness value can be converted according to the actual geometric deployment state of the rhombic link transmission component before entering the antagonistic spring assembly. Therefore, the four-bar included angle parameter introduces the actual geometric deployment state in the softened elastic coupling mechanism into the displacement conversion process, avoiding direct conversion based solely on the target required stiffness value and the inherent elastic coefficient of the antagonistic spring assembly while ignoring the force transmission differences of the rhombic link transmission component at different deployment angles.

[0042] Preferably, the specific implementation process for determining the nonlinear transmission coefficient of the rhomboid linkage transmission component to the physical force based on the four-bar angle parameter is as follows: After the four-bar angle parameter has been written into the four-bar angle parameter determination load record, a nonlinear transmission coefficient determination load record is first established, and the four-bar angle parameter, the rhomboid linkage transmission component, the softened elastic coupling mechanism, the target required stiffness value, and the linkage direction data are written into the nonlinear transmission coefficient determination load record. The nonlinear transmission coefficient determination load record is used to incorporate the geometric source of the four-bar angle parameter, the actual geometric unfolding state source of the linkage direction data, and the stiffness source of the target required stiffness value into the same transmission conversion caliber, so that the subsequently determined nonlinear transmission coefficient can be used to correct the force transmission difference before the target required stiffness value is transmitted to the end of the antagonistic spring assembly. In this application, the physical force is used to express the traction force applied by the human body to the softened elastic coupling mechanism through the rhomboid linkage, and the bearing reaction force fed back from the main skeleton of the wearable centaur robot to the softened elastic coupling mechanism through the rhomboid linkage. Both the traction force and the bearing reaction force are transmitted along the mechanical force transmission path between the human body, the rhomboid linkage, the softened elastic coupling mechanism, and the main skeleton. Therefore, the physical force is not an independent external concept, but rather a force parameter formed along the mechanical force transmission path. Subsequently, the transmission coefficient configuration bearing record is read according to the four-bar angle parameter. The transmission coefficient configuration bearing record is a configuration record formed after the rhomboid linkage completes mechanism calibration. The transmission coefficient configuration bearing record is used to record the force transmission ratio corresponding to different four-bar angle parameters. The force transmission ratio is derived from the directional decomposition and displacement amplification relationship of the physical force by the rhomboid linkage at different deployment angles. The force transmission ratio is subsequently used to form the nonlinear transmission coefficient. Further, the included angle parameter of the four-link is matched with the included angle interval field in the transmission coefficient configuration load record to obtain candidate transmission coefficient records. The included angle interval field is used to express different included angle ranges pre-configured in the transmission coefficient configuration load record, and the candidate transmission coefficient records are used to support transmission coefficient records that are in the same included angle interval field as the included angle parameter of the four-link. Then, the candidate transmission coefficient records are geometrically consistent with the link direction data to ensure that the candidate transmission coefficient records not only correspond to the included angle parameter of the four-link, but also correspond to the actual geometric deployment state of the rhombic link transmission component at the current operating moment. After the verification is completed, the nonlinear transmission coefficient is read from the candidate transmission coefficient records.The technical essence of the nonlinear transmission coefficient is to express the non-proportional transmission relationship generated when the rhombic link transmission component transmits the physical force to the end of the antagonistic spring assembly under the current four-bar angle parameter. Since the effect of the same physical force at the end of the antagonistic spring assembly does not remain linearly consistent when the actual geometric deployment state of the rhombic link transmission component changes, the nonlinear transmission coefficient is needed to convert the geometric force transmission difference corresponding to the four-bar angle parameter into a conversion parameter that can participate in the conversion of the target required stiffness value. The nonlinear transmission coefficient is subsequently used to convert the target required stiffness value to determine the converted stiffness requirement value transmitted to the end of the antagonistic spring assembly, so that the target required stiffness value can match the current force transmission state of the rhombic link transmission component.

[0043] Preferably, the specific implementation process of using the nonlinear transmission coefficient to convert the target required stiffness value to determine the converted stiffness requirement value transmitted to the end of the antagonistic spring assembly is as follows: After the nonlinear transmission coefficient has been determined by the four-bar linkage angle parameter, a converted stiffness requirement value determination load record is first established, and the nonlinear transmission coefficient, the target required stiffness value, the four-bar linkage angle parameter, the rhomboid linkage transmission component, the antagonistic spring assembly, and the nonlinear transmission coefficient determination load record are written into the converted stiffness requirement value determination load record. The converted stiffness requirement value determination load record is used to incorporate the connection stiffness target expressed by the target required stiffness value and the mechanism force conversion relationship expressed by the nonlinear transmission coefficient into the same processing caliber, so that the subsequently formed converted stiffness requirement value can correspond to the end of the antagonistic spring assembly, rather than remaining at the overall connection stiffness level of the softened elastic coupling mechanism. The target stiffness requirement value is derived from the load condition level and the leg support landing characteristics. This target stiffness requirement value expresses the connection stiffness target that the softened elastic coupling mechanism should exhibit within the current detection cycle. The nonlinear transmission coefficient is derived from the four-bar linkage angle parameter. This nonlinear transmission coefficient expresses the non-proportional transmission relationship of the rhomboid linkage transmission component to the physical force at the current operating moment. Subsequently, a conversion caliber consistency verification process is performed on the target stiffness requirement value and the nonlinear transmission coefficient to form a conversion caliber consistency verification result. This conversion caliber consistency verification result confirms that the target stiffness requirement value corresponds to the connection stiffness target of the softened elastic coupling mechanism, and that the nonlinear transmission coefficient corresponds to the conversion parameter when the rhomboid linkage transmission component transmits the physical force to the end of the antagonistic spring assembly. The conversion caliber consistency verification result continues to participate in subsequent conversion processing to avoid incorrectly applying the nonlinear transmission coefficient to stiffness parameters unrelated to the end of the antagonistic spring assembly, ensuring that the converted stiffness requirement value maintains a clear force transmission object. Furthermore, if the consistency verification result of the conversion caliber indicates that the target required stiffness value and the nonlinear transmission coefficient are within the same conversion caliber, the target required stiffness value is converted using the nonlinear transmission coefficient to form an initial converted stiffness requirement value. This initial converted stiffness requirement value expresses the stiffness requirement that the end of the antagonistic spring assembly needs to bear after considering the current geometric force transmission state of the rhomboid linkage. Subsequently, an end-bearing range verification process is performed on the initial converted stiffness requirement value to form an end-bearing range verification result. The end-bearing range originates from the elastic response range corresponding to the end of the antagonistic spring assembly in the parameter load record of the antagonistic spring assembly, and the end-bearing range verification result expresses whether the initial converted stiffness requirement value falls within the end-bearing range.When the end-bearing capacity verification result shows that the initial equivalent stiffness requirement value is within the end-bearing capacity range, the initial equivalent stiffness requirement value is established as the equivalent stiffness requirement value. When the end-bearing capacity verification result shows that the initial equivalent stiffness requirement value exceeds the end-bearing capacity range, the initial equivalent stiffness requirement value is subjected to end-bearing boundary constraint processing based on the end-bearing capacity boundary stiffness value, and the initial equivalent stiffness requirement value after end-bearing boundary constraint processing is established as the equivalent stiffness requirement value. The technical essence of the equivalent stiffness requirement value is to convert the target required stiffness value through the nonlinear transmission coefficient of the rhomboid linkage transmission component and then transmit it to the stiffness requirement at the end of the antagonistic spring assembly. The equivalent stiffness requirement value is subsequently used in conjunction with the inherent elastic coefficient of the antagonistic spring assembly to calculate the conversion quotient, so that the equivalent stiffness requirement value can be further converted into the target pre-compression displacement parameter of the antagonistic spring assembly. Through this process, the target required stiffness value is not directly converted into displacement, but is first converted through the current force transmission state of the rhomboid linkage, so that the target pre-compression displacement parameter can reflect the actual mechanism transmission relationship in the softened elastic coupling mechanism.

[0044] Preferably, the specific implementation process for calculating the conversion quotient between the reduced stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly, and then determining the target pre-compression displacement parameter of the antagonistic spring assembly based on the conversion quotient, is as follows: After the reduced stiffness requirement value has been written into the reduced stiffness requirement value determination load record, the antagonistic spring assembly parameter load record is first read, and the inherent elastic coefficient of the antagonistic spring assembly is read from the antagonistic spring assembly parameter load record. The antagonistic spring assembly parameter load record is a parameter record formed after the antagonistic spring assembly has been assembled and calibrated. The antagonistic spring assembly parameter load record is used to record the source of the elastic response of the antagonistic spring assembly in the pre-compression direction. The inherent elastic coefficient of the antagonistic spring assembly is used to express the structural elastic reference of the change in elastic reaction force formed by the antagonistic spring assembly under a unit pre-compression displacement change. The inherent elastic coefficient of the antagonistic spring assembly and the reduced stiffness requirement value together constitute the conversion basis for converting the end stiffness requirement into the pre-compression displacement target. Subsequently, a conversion quotient calculation load record is established, and the converted stiffness requirement value, the inherent elastic coefficient of the antagonistic spring assembly, the parameter load record of the antagonistic spring assembly, the load record for determining the converted stiffness requirement value, and the antagonistic spring assembly are written into the conversion quotient calculation load record. The conversion quotient calculation load record is used to incorporate the end stiffness source of the converted stiffness requirement value and the structural elasticity source of the inherent elastic coefficient of the antagonistic spring assembly into the same conversion caliber, so that the subsequently calculated conversion quotient value can correspond to the pre-compression displacement change of the antagonistic spring assembly. Further, a physical quantity correspondence verification process is performed on the converted stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly to form a physical quantity correspondence verification result. The physical quantity correspondence verification result is used to confirm that the converted stiffness requirement value expresses the stiffness requirement transmitted to the end of the antagonistic spring assembly, and the inherent elastic coefficient of the antagonistic spring assembly expresses the elastic response capability of the antagonistic spring assembly itself in the pre-compression direction. The physical quantity correspondence verification result continues to participate in the calculation process of the conversion quotient, ensuring that the converted stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly are calculated under the same pre-compression conversion caliber. This avoids directly converting the overall connection stiffness target of the softened elastic coupling mechanism with the inherent elastic coefficient of the antagonistic spring assembly. After the physical quantity correspondence verification result shows that the converted stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly are under the same pre-compression conversion caliber, the conversion quotient between the converted stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly is calculated, and the conversion quotient is written into the conversion quotient calculation load record.The technical essence of the conversion quotient is to convert the calculated stiffness requirement value into a candidate displacement result after converting it into the pre-compression displacement amount that the antagonistic spring assembly needs to achieve. This conversion quotient is not the final control command issued to the micro-screw actuator, but rather intermediate displacement data used to determine the target pre-compression displacement parameter. Subsequently, the pre-compressible range in the antagonistic spring assembly parameter load record is read, and the drivable displacement range of the micro-screw actuator is read. The pre-compressible range and the drivable displacement range are written into the target pre-compression displacement parameter determination load record. The pre-compressible range expresses the pre-compression displacement range within which the antagonistic spring assembly can maintain a stable elastic response, and the drivable displacement range expresses the axial displacement range within which the micro-screw actuator can drive the antagonistic spring assembly to change its initial deformation scale. The pre-compressible range and the drivable displacement range together define the executable range of the target pre-compression displacement parameter. Further, an executable range verification process is performed on the conversion quotient value to form an executable range verification result. When the executable range verification result shows that the conversion quotient value is within the executable range, the conversion quotient value is established as the target pre-compression displacement parameter. When the executable range verification result shows that the conversion quotient value exceeds the executable range, boundary constraint processing is performed on the conversion quotient value based on the boundary displacement value corresponding to the executable range, and the boundary-constrained conversion quotient value is established as the target pre-compression displacement parameter. The technical essence of the target pre-compression displacement parameter is the pre-compression displacement target that the antagonistic spring assembly needs to achieve at the current operating moment in order to bear the equivalent stiffness requirement value. The target pre-compression displacement parameter is subsequently used to generate a displacement adjustment command that matches the target pre-compression displacement parameter, so that the micro-screw actuator can change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism. Therefore, the conversion quotient connects the nonlinear transmission coefficient of the rhombic link transmission component, the equivalent stiffness requirement value, and the inherent elastic coefficient of the antagonistic spring assembly, so that the target pre-compression displacement parameter can simultaneously reflect the technical relationship between the target required stiffness value, the four-bar included angle parameter, the nonlinear transmission coefficient, the equivalent stiffness requirement value, and the inherent elastic coefficient of the antagonistic spring assembly.

[0045] Optionally, generating a displacement adjustment command matching the target pre-compression displacement parameter based on the target pre-compression displacement parameter includes: Read the actual deformation displacement data of the antagonistic spring assembly at the control trigger moment; Calculate the displacement deviation correction amount based on the target pre-compression displacement parameter and the actual deformation displacement data; The target number of rotations of the micro lead screw actuator is determined based on the displacement deviation correction amount. The target rotation direction of the micro-screw actuator is determined based on the positive and negative polarity parameters of the displacement deviation correction amount. The target rotation number is compiled into a drive pulse count value; The target rotation direction is compiled into a level polarity state control character; The displacement adjustment command is generated based on the drive pulse count value and the level polarity state control symbol.

[0046] Preferably, the specific implementation process for reading the actual deformation displacement data of the antagonistic spring assembly at the control trigger moment is as follows: Before the target pre-compression displacement parameter has been determined by the conversion quotient and enters the displacement adjustment command generation process, an actual deformation displacement data reading and bearing record is first established, and the target pre-compression displacement parameter, the antagonistic spring assembly, the softened elastic coupling mechanism, the micro-screw actuator assembly, and the control trigger moment are written into the actual deformation displacement data reading and bearing record. The actual deformation displacement data reading and bearing record is used to include the target source of the target pre-compression displacement parameter, the deformation object source of the antagonistic spring assembly, the stiffness adjustment source of the softened elastic coupling mechanism, the execution source of the micro-screw actuator assembly, and the reading time point source of the control trigger moment into the same reading caliber, so that the subsequently read actual deformation displacement data can form a homologous correspondence with the current target pre-compression displacement parameter. The control trigger moment is used to express the reading point when the target pre-compression displacement parameter has been formed and needs to be compared with the current physical deformation state of the antagonistic spring assembly. The control trigger moment is not an arbitrary sampling moment, but a command preparation point located after the formation of the target pre-compression displacement parameter and before the generation of the displacement adjustment command. Subsequently, the deformation detection load record corresponding to the antagonistic spring assembly is read; the deformation detection load record is used to save the spring end position, spring neutral reference position, current position of the lead screw nut, and historical adjustment held position of the antagonistic spring assembly at the control trigger moment. The spring end position is used to express the end position state of the antagonistic spring assembly after being compressed or released by the micro lead screw actuator assembly; the spring neutral reference position is used to express the reference position state of the antagonistic spring assembly when it is not pre-compressed; the current position of the lead screw nut is used to express the axial displacement state formed by the micro lead screw actuator assembly driving the antagonistic spring assembly; and the historical adjustment held position is used to express the held position state of the antagonistic spring assembly after the end of the previous adjustment cycle. Furthermore, the spring end position, the spring neutral reference position, the current position of the lead screw nut, and the historical adjusted holding position are processed to correspond their positions within the same axial diameter to form the actual deformation displacement data. The technical essence of the actual deformation displacement data is the pre-compression displacement state already formed by the antagonistic spring assembly at the control trigger moment. The actual deformation displacement data is neither the target pre-compression displacement parameter nor the displacement target that the micro-lead screw actuator is about to execute, but rather a representation of the initial deformation scale already present in the antagonistic spring assembly.The actual deformation displacement data, along with the target pre-compression displacement parameter, is subsequently used in the calculation of the displacement deviation correction amount. This ensures that the displacement deviation correction amount reflects the difference between the current physical deformation state of the antagonistic spring assembly and the target pre-compression displacement parameter, rather than directly generating the adjustment amount from a zero displacement state. By reading the actual deformation displacement data at the control trigger moment, the displacement adjustment command can be formed based on the current deformation of the antagonistic spring assembly, avoiding repeated compression or over-release of the antagonistic spring assembly. This provides a current position reference when the subsequent micro-screw actuator changes the initial deformation scale of the antagonistic spring assembly.

[0047] Preferably, the specific implementation process for calculating the displacement deviation correction amount based on the target pre-compression displacement parameter and the actual deformation displacement data is as follows: After the actual deformation displacement data has been written into the actual deformation displacement data reading and bearing record, a displacement deviation correction amount calculation and bearing record is first established, and the target pre-compression displacement parameter, the actual deformation displacement data, the control trigger time, the antagonistic spring assembly, the micro-screw actuator assembly, and the actual deformation displacement data reading and bearing record are written into the displacement deviation correction amount calculation and bearing record. The displacement deviation correction amount calculation and bearing record is used to include the target pre-compression displacement state expressed by the target pre-compression displacement parameter and the current pre-compression displacement state expressed by the actual deformation displacement data into the same calculation caliber, so that the subsequently formed displacement deviation correction amount can directly reflect the displacement difference that the antagonistic spring assembly needs to continue to compress or release. Subsequently, the target pre-compression displacement parameter and the actual deformation displacement data are subjected to displacement direction consistency processing, so that the target pre-compression displacement parameter and the actual deformation displacement data both adopt the same direction along the axial direction of the micro-screw actuator assembly pushing the antagonistic spring assembly as the positive displacement caliber. The displacement direction consistency processing is used to avoid inconsistencies in direction between the target pre-compression displacement parameter and the actual deformation displacement data due to different reading sources. The target pre-compression displacement parameter is derived from the conversion quotient between the calculated stiffness requirement value and the inherent elastic coefficient of the antagonistic spring assembly, while the actual deformation displacement data is derived from the deformation detection load record of the antagonistic spring assembly at the control trigger moment. Further, under the same displacement direction, the difference between the target pre-compression displacement parameter and the actual deformation displacement data is calculated to form an initial displacement deviation correction. This initial displacement deviation correction is used to express the axial displacement change required for the antagonistic spring assembly to adjust from the current pre-compression displacement state to the target pre-compression displacement state. Subsequently, a small displacement suppression process is performed on the initial displacement deviation correction amount. When the initial displacement deviation correction amount is lower than the displacement adjustment resolution lower limit, the initial displacement deviation correction amount is processed into a zero displacement deviation correction amount, and the zero displacement deviation correction amount is written into the displacement deviation correction amount calculation load record as the maintenance state of the displacement deviation correction amount. When the initial displacement deviation correction amount reaches the displacement adjustment resolution lower limit, the initial displacement deviation correction amount is established as the displacement deviation correction amount. The displacement adjustment resolution lower limit is derived from the minimum drivable axial displacement of the micro-screw actuator. The displacement adjustment resolution lower limit is written into the execution resolution load record when the micro-screw actuator completes the execution calibration and is read from the displacement deviation correction amount calculation load record.The technical essence of the displacement deviation correction is the executable displacement difference between the target pre-compression displacement parameter and the actual deformation displacement data. This displacement deviation correction is subsequently used to determine the target number of rotations and the target rotation direction of the micro-screw actuator, ensuring that the displacement adjustment command no longer only includes the target displacement result, but also the executable correction required to move from the current pre-compression displacement state to the target pre-compression displacement state. Through this process, the current pre-compression displacement state, the target pre-compression displacement state, and the driveable displacement resolution capability of the antagonistic spring assembly form a corresponding relationship in the same data link, providing a displacement difference basis for subsequent instruction compilation.

[0048] Preferably, the specific implementation process for determining the target number of rotations of the micro-screw actuator based on the displacement deviation correction amount is as follows: After the displacement deviation correction amount has been calculated from the target pre-compression displacement parameter and the actual deformation displacement data, a target rotation number determination load record is first established, and the displacement deviation correction amount, the micro-screw actuator, the antagonistic spring assembly, the control triggering time, and the displacement deviation correction amount calculation load record are written into the target rotation number determination load record. The target rotation number determination load record is used to include the displacement source of the displacement deviation correction amount and the axial transmission source of the micro-screw actuator in the same conversion caliber, so that the subsequently formed target rotation number can correspond to the rotational motion required when the micro-screw actuator drives the antagonistic spring assembly to change the initial deformation scale. Subsequently, the screw transmission parameter load record of the micro-screw actuator is read; the screw transmission parameter load record is formed after the micro-screw actuator completes assembly calibration, and is used to record the axial displacement conversion caliber of the screw nut mechanism in the micro-screw actuator. The screw drive parameter load record includes at least a screw lead field, a transmission backlash compensation field, and a maximum single rotation circle number segment. The screw lead field expresses the axial linear translation distance that the micro screw actuator can drive the screw nut mechanism to produce when the micro screw actuator rotates one revolution. The transmission backlash compensation field expresses the mechanical backlash displacement that the screw nut mechanism needs to compensate for during reversal or small displacement adjustment. The maximum single rotation circle number segment expresses the range of rotation circles that the micro screw actuator is allowed to execute in one displacement adjustment command. Further, the displacement deviation correction amount and the screw lead field are converted from displacement to rotation to form the initial number of rotation circles; simultaneously, the direction record of the previous adjustment cycle is read, which is used to store the target rotation direction corresponding to the micro screw actuator in the previous adjustment cycle. If the adjustment direction corresponding to the displacement deviation correction is different from the target rotation direction in the previous adjustment cycle direction record, then the initial number of rotations is compensated for based on the transmission clearance compensation field to form a compensated number of rotations. If the adjustment direction corresponding to the displacement deviation correction is the same as the target rotation direction in the previous adjustment cycle direction record, then the initial number of rotations is used as the compensated number of rotations. The compensated number of rotations is used to express the number of rotations that the miniature lead screw actuator needs to perform after considering the mechanical clearance of the lead screw and nut mechanism.Subsequently, a rotation range verification process is performed on the compensated number of rotations. When the compensated number of rotations is within the range defined by the maximum single rotation number segment, the compensated number of rotations is established as the target number of rotations. When the compensated number of rotations exceeds the range defined by the maximum single rotation number segment, a single rotation boundary processing is performed on the compensated number of rotations based on the maximum single rotation number segment, and the compensated number of rotations after the single rotation boundary processing is established as the target number of rotations. The technical essence of the target number of rotations is the rotational motion quantity that the micro-screw actuator needs to execute to eliminate the displacement deviation correction amount. The target number of rotations is subsequently compiled into a drive pulse count value, enabling the micro-screw actuator to perform rotational control in the form of pulse counting. By converting the displacement deviation correction amount into the target number of rotations, the target pre-compression displacement parameter can be transformed from a displacement-level adjustment target into a rotation-level control quantity that the micro-screw actuator can execute.

[0049] Preferably, the specific implementation process for determining the target rotation direction of the micro-screw actuator based on the positive and negative polarity parameters of the displacement deviation correction amount is as follows: After the displacement deviation correction amount has been written into the displacement deviation correction amount calculation load record, a target rotation direction determination load record is first established, and the displacement deviation correction amount, the positive and negative polarity parameters, the target pre-compression displacement parameter, the actual deformation displacement data, the antagonistic spring assembly, and the micro-screw actuator are written into the target rotation direction determination load record. The positive and negative polarity parameters are derived from the sign state of the displacement deviation correction amount and are used to express whether the target pre-compression displacement parameter needs to be increased or decreased relative to the actual deformation displacement data. Subsequently, the rotation direction mapping load record of the micro-screw actuator is read; the rotation direction mapping load record is formed when the micro-screw actuator completes the execution calibration and is used to record the correspondence between the rotor rotation direction of the servo motor inside the micro-screw actuator and the axial linear translation direction of the screw nut mechanism. The rotation direction mapping record includes at least a forward rotation field, a reverse rotation field, a compression direction field, and a release direction field; the forward rotation field is used to express the direction mark when the servo motor runs in a first rotation direction, the reverse rotation field is used to express the direction mark when the servo motor runs in a second rotation direction, the compression direction field is used to express the axial movement direction in which the lead screw and nut mechanism pushes the antagonistic spring assembly to increase the pre-compression displacement, and the release direction field is used to express the axial movement direction in which the lead screw and nut mechanism drives the antagonistic spring assembly to decrease the pre-compression displacement. Further, polarity determination processing is performed on the positive and negative polarity parameters. When the positive and negative polarity parameters indicate that the target pre-compression displacement parameter is greater than the actual deformation displacement data, the compression direction field is used as the direction matching object, and the rotation direction corresponding to the compression direction field is determined according to the rotation direction mapping load record. When the positive and negative polarity parameters indicate that the target pre-compression displacement parameter is less than the actual deformation displacement data, the release direction field is used as the direction matching object, and the rotation direction corresponding to the release direction field is determined according to the rotation direction mapping load record. When the positive and negative polarity parameters indicate that there is no executable displacement difference between the target pre-compression displacement parameter and the actual deformation displacement data, the stop direction state is written into the target rotation direction determination load record. Subsequently, the rotation direction corresponding to the compression direction field or the release direction field in the rotation direction mapping load record is established as the target rotation direction. When the stop direction state has been written into the target rotation direction determination load record, the stop direction state is established as the target rotation direction.The technical essence of the target rotation direction is that the miniature lead screw actuator needs to adopt a rotation direction state to adjust the antagonistic spring assembly from the actual deformation displacement data to the target pre-compression displacement parameter. The target rotation direction is subsequently compiled into a level polarity state control character, enabling the miniature lead screw actuator to first determine the rotor rotation orientation after receiving the displacement adjustment command, and then execute rotation according to the drive pulse count value. Through this processing, the displacement deviation correction amount is not only used to determine the target number of rotations, but also to determine the target rotation direction through the positive and negative polarity parameters, so that the displacement adjustment process simultaneously possesses both adjustment amplitude and adjustment direction execution information.

[0050] Preferably, the specific implementation process of compiling the target rotation number into a drive pulse count value is as follows: After the target rotation number has been written into the target rotation number determination record, a drive pulse count value compilation record is first established, and the target rotation number, the micro-screw actuator, the screw transmission parameter record, the control trigger time, and the target rotation number determination record are written into the drive pulse count value compilation record. The drive pulse count value compilation record is used to include the target rotation number and the pulse drive aperture of the micro-screw actuator in the same compilation aperture, so that the target rotation number can be converted into pulse count information that the micro-screw actuator can read. Subsequently, the pulse conversion record of the micro-screw actuator is read; the pulse conversion record is formed when the micro-screw actuator completes the execution calibration, and is used to record the number of pulses, the subdivision drive aperture, and the upper limit of the pulse count corresponding to each rotation number of the servo motor. The number of pulses corresponding to each rotation of the servo motor is used to express the basic number of pulses required for the servo motor to rotate one revolution. The subdivision drive aperture is used to express the drive aperture of the micro-screw actuator after subdividing the basic number of pulses during low-speed or small-displacement adjustment. The pulse count upper limit is used to express the maximum number of pulses allowed to be issued in a single displacement adjustment command. Further, a conversion from rotation revolutions to pulse count is performed based on the target number of rotations and the number of pulses corresponding to each rotation of the servo motor to form an initial drive pulse count value. Then, the initial drive pulse count value is further subdivided according to the subdivision drive aperture to form a subdivision drive pulse count value. The subdivision drive pulse count value is used to express the pulse count result required to drive the servo motor to complete the target number of rotations under the current drive aperture of the micro-screw actuator. Subsequently, a pulse count range verification process is performed on the subdivided drive pulse count value. When the subdivided drive pulse count value does not exceed the upper limit of the pulse count, the subdivided drive pulse count value is established as the drive pulse count value. When the subdivided drive pulse count value exceeds the upper limit of the pulse count, the subdivided drive pulse count value is segmented according to the upper limit of the pulse count to form the first segment pulse count result, and the first segment pulse count result is established as the drive pulse count value. The technical essence of the drive pulse count value is the pulse quantity control field obtained by converting the target rotation number. The drive pulse count value and the level polarity state control character subsequently participate in the generation of the displacement adjustment command, so that the displacement adjustment command can simultaneously contain two types of execution information: how much the micro-screw actuator needs to rotate and in which direction it needs to rotate.Through this compilation process, the target number of rotations is converted into a pulse count field that can be read by the drive control terminal of the micro-screw actuator, thereby enabling the target pre-compression displacement parameter of the antagonistic spring assembly to be implemented in the pulse drive process of the servo motor.

[0051] Preferably, the specific implementation process of compiling the target rotation direction into a level polarity state control character is as follows: After the target rotation direction has been written into the target rotation direction determination bearing record, a level polarity state control character compilation bearing record is first established, and the target rotation direction, the positive and negative polarity parameters, the micro-screw actuator, the rotation direction mapping bearing record, and the target rotation direction determination bearing record are written into the level polarity state control character compilation bearing record. The level polarity state control character compilation bearing record is used to incorporate the target rotation direction and the direction control aperture of the micro-screw actuator into the same compilation aperture, so that the target rotation direction can be converted into a direction state control field that can be read by the drive control terminal of the micro-screw actuator. Subsequently, the direction control field configuration bearing record of the micro-screw actuator is read; the direction control field configuration bearing record is formed when the micro-screw actuator completes the execution calibration, and is used to record the level polarity representation corresponding to different direction states. The direction control field configuration record includes at least a forward rotation level field, a reverse rotation level field, and a stop-and-hold level field. The forward rotation level field expresses the level polarity state that needs to be written to the drive control terminal when the target rotation direction is forward; the reverse rotation level field expresses the level polarity state that needs to be written to the drive control terminal when the target rotation direction is reverse; and the stop-and-hold level field expresses the hold state that needs to be written to the drive control terminal when the displacement deviation correction amount does not have an executable displacement difference. Further, the target rotation direction is matched with the forward rotation level field, the reverse rotation level field, and the stop-and-hold level field in the direction control field configuration record. When the target rotation direction corresponds to the forward rotation field, the forward rotation level field is read to form a forward level polarity state control character; when the target rotation direction corresponds to the reverse rotation field, the reverse rotation level field is read to form a reverse level polarity state control character; and when the target rotation direction corresponds to the stop direction state, the stop-and-hold level field is read to form a stop level polarity state control character. Subsequently, the level polarity state control character corresponding to the target rotation direction among the positive level polarity state control character, the reverse level polarity state control character, or the stop level polarity state control character is established as the level polarity state control character, and the level polarity state control character is written into the level polarity state control character compilation and bearing record. The technical essence of the level polarity state control character is a direction control field obtained by converting the target rotation direction. The level polarity state control character is used to instruct the drive control end of the micro-screw actuator to first lock the rotor rotation orientation of the servo motor in the subsequent displacement adjustment command.The level polarity state control symbol, together with the drive pulse count value, participates in the generation of the displacement adjustment command, enabling the displacement adjustment command to establish a correspondence between the direction field and the pulse count field. This prevents the antagonistic spring assembly from being compressed or released in the wrong direction simply by issuing the number of pulses without specifying the rotation direction. Through this compilation process, the positive and negative polarity parameters of the displacement deviation correction amount are converted into level polarity state control symbols executable by the micro-screw actuator, providing a directional control basis for subsequently changing the initial deformation scale of the antagonistic spring assembly.

[0052] Preferably, the specific implementation process of generating the displacement adjustment command based on the drive pulse count value and the level polarity state control character is as follows: After the drive pulse count value and the level polarity state control character have been formed, a displacement adjustment command generation record is first established, and the drive pulse count value, the level polarity state control character, the target pre-compression displacement parameter, the actual deformation displacement data, the displacement deviation correction amount, the target number of rotations, the target rotation direction, the micro-screw actuator, and the antagonistic spring assembly are written into the displacement adjustment command generation record. The displacement adjustment command generation record is used to incorporate the displacement target source of the target pre-compression displacement parameter, the current deformation source of the actual deformation displacement data, the correction amount source of the displacement deviation correction amount, the rotation amplitude source of the target number of rotations, the rotation direction source of the target rotation direction, the pulse count source of the drive pulse count value, and the direction control source of the level polarity state control character into the same command generation caliber, so that the subsequently formed displacement adjustment command can fully express the execution information required by the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly. Subsequently, a field correspondence verification process is performed on the drive pulse count value and the level polarity state control symbol to form a field correspondence verification result. This result indicates whether the drive pulse count value matches the level polarity state control symbol corresponding to the same displacement deviation correction amount, preventing incorrect combination of the pulse count field and direction control field in different adjustment cycles. Further, if the field correspondence verification result indicates that the drive pulse count value and the level polarity state control symbol are from the same adjustment source, the level polarity state control symbol, the drive pulse count value, the control trigger time, the antagonistic spring component identification field, and the instruction verification field are encapsulated according to the communication protocol field arrangement of the micro-screw actuator to form an initial displacement adjustment instruction. The antagonistic spring component identification field indicates the antagonistic spring component corresponding to this displacement adjustment instruction, the instruction verification field indicates the field integrity verification result of the initial displacement adjustment instruction before and after transmission, and the communication protocol field arrangement is written into the communication protocol configuration bearer record when the micro-screw actuator completes communication calibration and is read from the displacement adjustment instruction generation bearer record. Subsequently, the initial displacement adjustment command is subjected to instruction integrity verification processing; when the level polarity state control character, the drive pulse count value, the control trigger time, the antagonistic spring component identification field, and the instruction verification field in the initial displacement adjustment command are all in a readable state, the initial displacement adjustment command is established as the displacement adjustment command.The technical essence of the displacement adjustment command is to convert the difference between the target pre-compression displacement parameter and the actual deformation displacement data into control data after the micro-screw actuator can execute a direction control field and a pulse count field. The displacement adjustment command is then sent to the micro-screw actuator, enabling it to determine the target rotation direction based on the level polarity state control symbol and execute a rotational action corresponding to the target number of rotations based on the drive pulse count value. This, in turn, causes the screw nut mechanism to undergo axial linear translation, thereby changing the initial deformation scale of the antagonistic spring assembly. Through this processing, the target pre-compression displacement parameter, the actual deformation displacement data, the displacement deviation correction amount, the target number of rotations, the target rotation direction, the drive pulse count value, and the level polarity state control symbol form a continuous data conversion relationship within the same displacement adjustment command, ensuring that the equivalent physical stiffness update of the softened elastic coupling mechanism falls within the executable displacement adjustment process.

[0053] Optionally, the process of driving the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command includes: The displacement adjustment command is processed by unpacking the communication protocol to generate a set of adjustment command fields; Read the target pulse sequence parameters from the set of adjustment instruction fields; Read the target level polarity state from the set of adjustment instruction fields; The target level polarity state is applied to the drive control terminal of the micro ball screw actuator to lock the rotor rotation orientation of the servo motor inside the micro ball screw actuator. The target pulse sequence parameters are transmitted to the drive control terminal to trigger the servo motor to drive the lead screw and nut mechanism to perform axial linear translation; The initial deformation scale of the antagonistic spring assembly is changed based on the axial linear translation of the lead screw and nut mechanism.

[0054] Preferably, the specific implementation process of performing communication protocol unpacking processing on the displacement adjustment command to generate the adjustment command field set is as follows: After the displacement adjustment command has been generated by the drive pulse count value and the level polarity state control character, the communication protocol configuration carrier record corresponding to the micro screw execution component is read first, and the communication protocol field arrangement, field start position, field length, field verification, and antagonistic spring component identifier field reading are read from the communication protocol configuration carrier record. The communication protocol configuration bearer record is a configuration record formed when the micro-screw actuator completes communication calibration. This record ensures that each field in the displacement adjustment command can be parsed according to the same communication protocol field arrangement. The field start position specification defines the starting point for each field in the displacement adjustment command, the field length specification defines the reading range for each field, the field verification specification verifies the integrity of the read field content, and the antagonistic spring component identifier field reading specification locates the corresponding antagonistic spring component identifier field from the displacement adjustment command, preventing the displacement adjustment command from being directly sent to the micro-screw actuator as overall control data lacking field boundaries. Subsequently, an adjustment command unpacking bearer record is established, and the displacement adjustment command, the communication protocol configuration bearer record, the drive pulse count value, the level polarity state control character, the control trigger time, and the antagonistic spring component identifier field are written into this record. The adjustment instruction unpacking and carrying record is used to include the instruction source of the displacement adjustment instruction, the rotation amplitude source of the drive pulse count value, the rotation direction source of the level polarity state control character, the reading time point source of the control trigger moment, and the execution object source of the antagonistic spring component identification field into the same unpacking caliber. This ensures that the subsequently generated adjustment instruction field set can correspond to the antagonistic spring component whose initial deformation scale needs to be changed. Further, according to the communication protocol field arrangement caliber, the displacement adjustment instruction is sequentially processed by field boundary identification, field content truncation, and field integrity verification to parse the target pulse sequence parameter field, target level polarity state field, control trigger moment field, antagonistic spring component identification field, and instruction verification field from the displacement adjustment instruction.The target pulse sequence parameter field is derived from the drive pulse count value and is used to carry the pulse quantity information subsequently transmitted to the drive control terminal; the target level polarity state field is derived from the level polarity state control character and is used to carry the direction state information subsequently applied to the drive control terminal; the control trigger time field is used to associate the current displacement adjustment command with the control trigger time that forms the actual deformation displacement data; the antagonistic spring component identifier field is used to define the antagonistic spring component corresponding to the current displacement adjustment command; the command verification field is used to record the field integrity verification results of the displacement adjustment command before and after unpacking. Subsequently, the target pulse sequence parameter field, the target level polarity state field, the control trigger time field, the antagonistic spring component identifier field, and the command verification field are written into the adjustment command field set according to the field source and field purpose. The technical essence of the adjustment command field set is a readable control field combination formed by unpacking the displacement adjustment command. The adjustment command field set is subsequently used to read the target pulse sequence parameter and the target level polarity state, so that the micro screw actuator can obtain the rotation amplitude information and rotation direction information before performing displacement adjustment. Through this processing, the displacement adjustment command is converted from the overall control data into a set of adjustment command fields that can be read item by item, providing a field-based execution basis for subsequently driving the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly.

[0055] Preferably, the specific implementation process for reading the target pulse sequence parameters from the adjustment instruction field set is as follows: After the adjustment instruction field set has been formed by unpacking the communication protocol of the displacement adjustment instruction, a target pulse sequence parameter reading record is first established, and the adjustment instruction field set, the target pulse sequence parameter field, the drive pulse count value, the target number of rotations, the micro screw actuator, and the antagonistic spring assembly identifier field are written into the target pulse sequence parameter reading record. The target pulse sequence parameter reading record is used to include the field source of the target pulse sequence parameters, the compilation source of the drive pulse count value, the rotation amplitude source of the target number of rotations, and the execution object source of the antagonistic spring assembly identifier field into the same reading caliber, so that the target pulse sequence parameters read subsequently can be traced back to the rotation control amplitude corresponding to the displacement deviation correction amount, and can be matched with the antagonistic spring assembly whose initial deformation scale needs to be changed this time. Subsequently, the target pulse sequence parameter field is located from the set of adjustment command fields, and the pulse count field, pulse transmission rhythm field, and pulse segmentation field in the target pulse sequence parameter field are read according to the field length caliber in the communication protocol configuration bearer record. The pulse count field is derived from the drive pulse count value and is used to express the number of pulses that the servo motor needs to receive; the pulse transmission rhythm field is used to express the time arrangement caliber of the pulses sent by the drive control terminal to the servo motor, so that the servo motor can rotate according to the rhythm that matches the mechanical response of the micro lead screw actuator; the pulse segmentation field is used to express the pulse segmentation sequence corresponding to the displacement adjustment command when the drive pulse count value has undergone segmented counting processing. Further, pulse field consistency verification processing is performed on the pulse count field, the pulse transmission rhythm field, and the pulse segmentation field to confirm that the pulse count field maintains a conversion correspondence with the target number of rotations, that the pulse transmission rhythm field maintains a correspondence with the pulse drive caliber of the micro lead screw actuator, and that the pulse segmentation field maintains a correspondence with the segmented counting processing of the drive pulse count value. After completing the pulse field consistency verification process, the pulse count field, the pulse transmission rhythm field, and the pulse segmentation field are combined into the target pulse sequence parameter, and the target pulse sequence parameter is written into the target pulse sequence parameter read and bearer record. The technical essence of the target pulse sequence parameter is a pulse execution parameter read from the adjustment instruction field set, used to drive the servo motor to complete the target number of rotations. The target pulse sequence parameter is not simply the number of pulses, but rather executable pulse data that simultaneously includes the number of pulses, the pulse transmission rhythm, and the pulse segmentation relationship.The target pulse sequence parameters are subsequently transmitted to the drive control terminal, enabling the drive control terminal to trigger the servo motor to rotate according to the target pulse sequence parameters, and to convert the rotational motion of the servo motor into axial linear translation through the lead screw and nut mechanism. Thus, the target pulse sequence parameters further organize the preceding drive pulse count values ​​into pulse control data suitable for execution by the drive control terminal, allowing the target pre-compression displacement parameter to be transformed into the initial deformation scale adjustment process of the antagonistic spring assembly through the micro lead screw actuation component.

[0056] Preferably, the specific implementation process of reading the target level polarity state from the adjustment instruction field set is as follows: After the target level polarity state field has been saved in the adjustment instruction field set, a target level polarity state reading record is first established, and the adjustment instruction field set, the target level polarity state field, the level polarity state control character, the target rotation direction, the positive and negative polarity parameters, the micro-screw actuator, and the drive control terminal are written into the target level polarity state reading record. The target level polarity state reading record is used to include the field source of the target level polarity state, the compilation source of the level polarity state control character, the direction source of the target rotation direction, and the displacement difference source of the positive and negative polarity parameters into the same reading caliber, so that the target level polarity state read subsequently can correspond to the execution direction in which the antagonistic spring assembly needs to increase or decrease the pre-compression displacement. Subsequently, the target level polarity state field is located from the adjustment instruction field set, and the direction state field, holding state field, and direction verification field in the target level polarity state field are read. The direction state field originates from the level polarity state control character and is used to express the forward rotation state or reverse rotation state corresponding to the target rotation direction; the hold state field is used to express that when the displacement deviation correction amount does not have an executable displacement difference, the drive control terminal needs to maintain the current direction control state; the direction verification field is used to express the corresponding verification result between the direction state field and the target rotation direction. Further, the direction state field, the hold state field, and the direction verification field are read correspondingly with the rotation direction mapping record to confirm that the direction state field can be mapped to the compression direction field or the release direction field, and to confirm that the hold state field can be mapped to the stop hold level field; when the direction state field is mapped to the compression direction field, the direction state field is written into the compression direction level in the target level polarity state reading record; when the direction state field is mapped to the release direction field, the direction state field is written into the release direction level in the target level polarity state reading record; when the hold state field is read, the hold state field is written into the hold direction level in the target level polarity state reading record. Subsequently, the target level polarity state is formed based on the compression direction level aperture, the release direction level aperture, or the holding direction level aperture. The technical essence of the target level polarity state is a level polarity execution parameter read from the adjustment instruction field set, used to control the direction state of the drive control terminal. The target level polarity state is not an ordinary state flag, but rather is used to enable the drive control terminal to determine the rotor rotation orientation of the servo motor before receiving the target pulse sequence parameter.The target level polarity state is subsequently applied to the drive control terminal to lock the rotor rotation orientation of the servo motor inside the micro lead screw actuator, ensuring that the pulse input corresponding to the target pulse sequence parameter can drive the lead screw nut mechanism to perform axial linear translation in the correct direction. Through this process, the target level polarity state and the target pulse sequence parameter form a correspondence within the same set of adjustment command fields, preventing the servo motor from operating solely based on the number of pulses without directional constraints.

[0057] Preferably, the specific implementation process of applying the target level polarity state to the drive control terminal of the micro-screw actuator to lock the rotor rotation orientation of the servo motor inside the micro-screw actuator is as follows: After the target level polarity state has been read from the adjustment command field set, a rotor rotation orientation locking load record is first established, and the target level polarity state, the target rotation direction, the drive control terminal, the micro-screw actuator, the servo motor, the rotation direction mapping load record, and the target level polarity state reading load record are written into the rotor rotation orientation locking load record. The rotor rotation orientation locking load record is used to include the direction control source of the target level polarity state and the rotor rotation orientation of the servo motor in the same control caliber, so that the drive control terminal can complete the direction state configuration before transmitting the target pulse sequence parameters. Subsequently, the target level polarity state is written into the direction control field of the drive control terminal, and level holding verification processing is performed on the direction control field. The direction control field is used to store the current level polarity state of the drive control terminal. The level hold verification process is used to confirm that the target level polarity state remains stable during the transmission of the target pulse sequence parameters, and that the rotor rotation orientation does not change due to communication jitter or residual states from the previous adjustment cycle. Further, based on the correspondence between the target level polarity state and the rotation direction mapping record, the rotor rotation orientation corresponding to the target level polarity state is determined. When the target level polarity state corresponds to the forward rotation field, the rotor rotation orientation corresponding to the forward rotation field is written into the rotor rotation orientation lock record. When the target level polarity state corresponds to the reverse rotation field, the rotor rotation orientation corresponding to the reverse rotation field is written into the rotor rotation orientation lock record. When the target level polarity state corresponds to the stop hold level field, the stop hold state corresponding to the stop hold level field is written into the rotor rotation orientation lock record. Subsequently, a direction lock confirmation process is performed on the rotor rotation orientation to form a rotor rotation orientation lock result. The rotor rotation orientation locking result indicates that the drive control terminal has defined the rotation direction of the servo motor according to the target level polarity state. This rotor rotation orientation locking result subsequently participates in the target pulse sequence parameter transmission process, ensuring that the target pulse sequence parameters only enter the drive control terminal after the rotor rotation orientation is locked. The technical essence of the rotor rotation orientation is that the servo motor converts the pulse input into a forward or reverse rotation direction. The rotor rotation orientation directly determines whether the lead screw and nut mechanism pushes the antagonistic spring assembly to increase pre-compression displacement or drives it to decrease pre-compression displacement.By first applying the target level polarity state to the drive control terminal and forming the rotor rotation orientation lock result, the rotation action triggered by the subsequent target pulse sequence parameters can maintain a corresponding relationship with the positive and negative polarity parameters of the displacement deviation correction amount, thereby providing a directional basis for the initial deformation scale adjustment of the antagonistic spring assembly.

[0058] Preferably, the specific implementation process of transmitting the target pulse sequence parameter to the drive control terminal to trigger the servo motor to drive the lead screw and nut mechanism to perform axial linear translation is as follows: After the rotor rotation orientation locking result has been written into the rotor rotation orientation locking load record, a pulse drive transmission load record is first established, and the target pulse sequence parameter, the rotor rotation orientation locking result, the drive control terminal, the servo motor, the lead screw and nut mechanism, the micro lead screw actuator, and the target pulse sequence parameter reading load record are written into the pulse drive transmission load record. The pulse drive transmission load record is used to include the rotation amplitude source of the target pulse sequence parameter and the rotation direction source of the rotor rotation orientation locking result into the same drive caliber, so that the servo motor can receive pulse input in a state where the direction is determined. Subsequently, according to the pulse sending beat field in the target pulse sequence parameter, the number of pulses corresponding to the pulse counting field is sequentially transmitted to the drive control terminal, and a pulse receiving count record is established in the drive control terminal. The pulse reception count record is used to record the number of pulses received by the drive control terminal, the number of pulses not yet received, and the corresponding pulse transmission cycle, enabling the drive control terminal to execute pulse output according to the target pulse sequence parameters. Further, the drive control terminal converts the pulse input in the pulse reception count record into the rotational action of the servo motor based on the rotor rotation orientation locking result, and writes the rotational action of the servo motor into the rotational execution load record. The rotational execution load record is used to correlate the pulse input corresponding to the target pulse sequence parameters with the actual rotational action executed by the servo motor, so that the subsequent axial linear translation of the lead screw and nut mechanism can be traced back to the target pulse sequence parameters. Subsequently, the rotational action of the servo motor is converted to linear displacement through the lead screw and nut mechanism to form an axial linear translation, and the axial linear translation is written into the pulse drive transmission load record. The technical essence of the axial linear translation is that the lead screw and nut mechanism generates a linear displacement along the pre-compression or release direction under the rotational action of the servo motor. This axial linear translation is not an independent mechanical movement, but rather a displacement execution result formed after the direction is determined by the target voltage polarity state, the amplitude by the target pulse sequence parameters, and the result of conversion by the lead screw and nut mechanism. The axial linear translation then directly acts on the spring end position of the antagonistic spring assembly, allowing the initial deformation scale of the antagonistic spring assembly to change according to the displacement adjustment command. Through this process, the target pulse sequence parameters are converted into the rotational action of the servo motor, and further into the axial linear translation of the lead screw and nut mechanism, ensuring that the pulse control field in the displacement adjustment command falls within an executable mechanical displacement process.

[0059] Preferably, the specific implementation process of changing the initial deformation scale of the antagonistic spring assembly based on the axial linear translation of the lead screw and nut mechanism is as follows: After the lead screw and nut mechanism has undergone axial linear translation according to the target pulse sequence parameter, an initial deformation scale change load record is first established, and the axial linear translation, the lead screw and nut mechanism, the antagonistic spring assembly, the spring end position, the spring neutral reference position, the actual deformation displacement data, the target pre-compression displacement parameter, and the displacement adjustment command are written into the initial deformation scale change load record. The initial deformation scale change load record is used to include the execution source of the axial linear translation, the current deformation source of the actual deformation displacement data, and the target deformation source of the target pre-compression displacement parameter into the same deformation adjustment caliber, so that the subsequently changed initial deformation scale can correspond to the current displacement adjustment command. Subsequently, the axial linear translation of the lead screw and nut mechanism is transmitted to the spring end position of the antagonistic spring assembly. When the axial linear translation corresponds to the compression direction, the distance between the spring end position and the spring neutral reference position increases, thereby increasing the initial deformation scale of the antagonistic spring assembly. When the axial linear translation corresponds to the release direction, the distance between the spring end position and the spring neutral reference position decreases, thereby reducing the initial deformation scale of the antagonistic spring assembly. When the axial linear translation corresponds to the holding state, the spring end position remains within the position range corresponding to the actual deformation displacement data. Further, after the spring end position is displaced, the adjusted spring end position of the antagonistic spring assembly is read, and the adjusted spring end position is correlated with the spring neutral reference position to form adjusted deformation displacement data. The adjusted deformation displacement data is used to express the pre-compression displacement state of the antagonistic spring assembly after being driven by the displacement adjustment command. The adjusted deformation displacement data is subsequently compared with the target pre-compression displacement parameter for displacement arrival verification, ensuring that the initial deformation scale change load record can record whether the adjustment has reached the pre-compression displacement target corresponding to the target pre-compression displacement parameter. Subsequently, the adjusted deformation displacement data is established as the changed initial deformation scale, and the changed initial deformation scale is written into the initial deformation scale change load record. The technical essence of the initial deformation scale is the degree of pre-compression deformation already formed by the antagonistic spring assembly before the softened elastic coupling mechanism enters the stressed state. The initial deformation scale directly affects the change in elastic reaction force generated by the antagonistic spring assembly when the softened elastic coupling mechanism is stressed, thereby affecting the equivalent physical stiffness of the softened elastic coupling mechanism.After the initial deformation scale is changed, the softened elastic coupling mechanism can form an updated equivalent physical stiffness according to the new initial deformation scale, so that the equivalent physical stiffness maintains a correspondence with the stiffness adjustment requirement corresponding to the target pre-compression displacement parameter. Through this process, the target level polarity state and the target pulse sequence parameter in the displacement adjustment command are finally converted into the initial deformation scale change result of the antagonistic spring assembly, so that the stiffness adjustment requirement formed in the previous step based on the load condition level and the leg support period ground landing characteristics can be implemented in the physical deformation state of the softened elastic coupling mechanism.

[0060] Optionally, after updating the equivalent physical stiffness of the softened elastic coupling mechanism, the method further includes: The system collects real-time drag force feedback data of the human body object applied to the softened elastic coupling mechanism by the rhomboid linkage transmission component. Monitor the force value jump gradient parameter of the real-time drag force feedback data within a continuous time window; When the force value jump gradient parameter exceeds the smoothness determination threshold, the softened elastic coupling mechanism is driven to undergo a secondary passive buffer deformation. The secondary passive buffer deformation absorbs the sudden impact energy parameters generated by the wearable centaur robot's shaking when crossing obstacles, thereby blocking the rigid transmission path of the sudden impact energy parameters to the support side where the human body is located.

[0061] Preferably, the specific implementation process for collecting real-time drag force feedback data of the human body object applied to the softened elastic coupling mechanism via the rhomboid linkage is as follows: After the equivalent physical stiffness of the softened elastic coupling mechanism has been updated according to the displacement adjustment command, a real-time drag force feedback data acquisition record is first established, and the equivalent physical stiffness, the target pre-compression displacement parameter, the displacement adjustment command, the initial deformation scale of the antagonistic spring assembly, the rhomboid linkage, and the softened elastic coupling mechanism are written into the real-time drag force feedback data acquisition record. The real-time drag force feedback data acquisition record is used to correlate the elastic coupling state after the equivalent physical stiffness update with the subsequent drag force acquisition process, so that the collected real-time drag force feedback data can reflect the human body traction change borne by the softened elastic coupling mechanism under the current stiffness state corresponding to the equivalent physical stiffness, rather than being mixed with the connection state before the equivalent physical stiffness update. Subsequently, along the mechanical force transmission path between the human body, the rhomboid linkage, the softened elastic coupling mechanism, and the main skeleton of the wearable centaur robot, the connection force data formed between the connection nodes of the rhomboid linkage on the human side and the connection nodes on the main skeleton side is read, and the elastic reaction force feedback data of the antagonistic spring assembly at the current initial deformation scale is read simultaneously. The connection force data is used to express the force change formed when the human body pulls the softened elastic coupling mechanism through the rhomboid linkage, and the elastic reaction force feedback data is used to express the elastic response of the antagonistic spring assembly to the force change based on the initial deformation scale. Further, the connection force data and the elastic reaction force feedback data are processed to correspond to the sampling time, and the real-time drag force feedback data is generated according to the time sequence within the current detection period. The technical essence of the real-time drag force feedback data is the dynamic force feedback sequence applied by the human object to the softened elastic coupling mechanism via the rhomboid linkage. This real-time drag force feedback data is not a single static force value, but a data sequence including sampling time, force change direction, force change amplitude, and corresponding stiffness update state. Since the equivalent physical stiffness of the softened elastic coupling mechanism has been adjusted by the target pre-compression displacement parameter, the real-time drag force feedback data can reflect the actual traction relationship between the support side of the human object and the main skeleton of the wearable centaur robot after the equivalent physical stiffness update. The subsequent monitoring process within the continuous time window allows the force value jump gradient parameter to be formed based on continuous force changes, rather than based on a single force peak.Through this acquisition method, the real-time drag force feedback data incorporates the elastic coupling state after the equivalent physical stiffness update, the force transmission state of the rhomboid linkage, and the traction state of the support side where the human object is located into the same force feedback link, providing a source of force data for subsequent identification of the sudden impact energy parameters caused by obstacle crossing swaying.

[0062] Preferably, the specific implementation process of monitoring the force jump gradient parameter of the real-time drag force feedback data within a continuous-time sliding window is as follows: After the real-time drag force feedback data has been written into the real-time drag force feedback data acquisition bearing record, a force jump gradient parameter monitoring bearing record is first established, and the real-time drag force feedback data, the equivalent physical stiffness, the initial deformation scale, the rhomboid linkage transmission component, and the current detection period are written into the force jump gradient parameter monitoring bearing record. The force jump gradient parameter monitoring bearing record is used to correlate the force source of the real-time drag force feedback data with the current stiffness state of the softened elastic coupling mechanism corresponding to the equivalent physical stiffness, so that the subsequently formed force jump gradient parameter can represent the drag force change rate under the current stiffness state corresponding to the equivalent physical stiffness. Subsequently, the continuous-time sliding window configuration bearing record is read, and the sliding window start sampling diameter, sliding window end sampling diameter, adjacent sampling time interval diameter, and sliding window sliding diameter are read from the continuous-time sliding window configuration bearing record. The continuous-time sliding window configuration load record is a time-intercepting record configured during the initialization of the stiffness adjustment method of the softened elastic coupling mechanism. This continuous-time sliding window configuration load record is used to define the time range for extracting continuous force segments from the real-time drag force feedback data. The continuous-time sliding window is not an arbitrarily selected time period, but rather used to cover short-term force segments where the wearable Centaur robot experiences rapid changes in drag force during movement, load-bearing, and obstacle crossing. Further, according to the starting and ending sampling calibers of the sliding window, multiple drag force sample values ​​located within the same continuous-time sliding window are read from the real-time drag force feedback data. Adjacent drag force sample values ​​are then subjected to force value difference processing according to the adjacent sampling time interval caliber to form a force value change sequence within the sliding window. This force value change sequence within the sliding window is used to express the magnitude of force value change generated by the real-time drag force feedback data over time within the continuous-time sliding window. The force value change sequence within the sliding window is further gradient-processed in conjunction with the adjacent sampling time interval caliber to form the force value jump gradient parameter. The technical essence of the force value jump gradient parameter is that the real-time drag force feedback data corresponds to the intensity of the sudden force change per unit time within the continuous time window. The force value jump gradient parameter is not the drag force itself, nor its maximum value, but rather a parameter reflecting the rate of increase or decrease of the drag force within a short period. The force value jump gradient parameter is subsequently compared with a smoothing judgment threshold, enabling the softened elastic coupling mechanism to distinguish between normal walking traction changes and sudden impact changes caused by obstacle crossing. By monitoring the force value jump gradient parameter within the continuous time window, the misjudgment of slow traction changes during normal gait transitions as sudden impacts can be reduced, while ensuring that the subsequent secondary passive buffer deformation corresponds to the time range and direction of the force jump.

[0063] Preferably, when the force value jump gradient parameter exceeds the smoothing judgment threshold, the specific implementation process of driving the softened elastic coupling mechanism to undergo secondary passive buffer deformation is as follows: After the force value jump gradient parameter has been formed by the real-time drag force feedback data monitored within the continuous time sliding window, the smoothing judgment threshold configuration load record is first read, and the smoothing judgment threshold value corresponding to the current equivalent physical stiffness, the current load condition level, and the current leg support period grounding characteristics is read from the smoothing judgment threshold configuration load record. The smoothing judgment threshold configuration load record is a configuration record formed when the softened elastic coupling mechanism completes structural calibration and stiffness adjustment range calibration. The smoothing judgment threshold value is not a fixed arbitrary threshold, but a force jump judgment boundary jointly determined based on the elastic response range corresponding to the equivalent physical stiffness, the load condition level corresponding to the load state, and the human body support period force state corresponding to the leg support period grounding characteristics. The smoothing judgment threshold is used to express that, under the current stiffness state and current load state corresponding to the equivalent physical stiffness, the real-time drag force feedback data can still be regarded as the maximum gradient boundary of the gradual traction change; the force value jump gradient parameter exceeding this gradient boundary indicates that the softened elastic coupling mechanism is experiencing a short-term sudden increase or decrease in drag force. Subsequently, a secondary passive buffer deformation triggering load record is established, and the force value jump gradient parameter, the smoothing judgment threshold, the real-time drag force feedback data, the continuous time sliding window, the equivalent physical stiffness, and the initial deformation scale are written into the secondary passive buffer deformation triggering load record. The secondary passive buffer deformation triggering load record is used to include the source of the force jump, the source of the judgment boundary, and the current elastic coupling state into the same trigger caliber, so that the subsequent secondary passive buffer deformation can correspond to the force period when the force value jump gradient parameter exceeds the smoothing judgment threshold. Furthermore, the force jump gradient parameter is compared with the smoothing determination threshold; when the force jump gradient parameter does not exceed the smoothing determination threshold, the softened elastic coupling mechanism is maintained in the elastic coupling state corresponding to the equivalent physical stiffness; when the force jump gradient parameter exceeds the smoothing determination threshold, the remaining deformable range of the softened elastic coupling mechanism is released according to the initial deformation scale and the remaining deformable range of the antagonistic spring assembly, so that the softened elastic coupling mechanism undergoes the secondary passive buffer deformation.The essence of the secondary passive buffer deformation technology lies in the fact that, after the equivalent physical stiffness has been updated, the softened elastic coupling mechanism, facing the short-term elastic deformation process that occurs again due to a sudden force jump, does not regenerate the displacement adjustment command, nor does it change the target pre-compression displacement parameter again. Instead, it forms a buffer deformation based on the current initial deformation scale, utilizing the remaining deformable range of the antagonistic spring assembly and the geometric force transmission margin of the rhomboid linkage transmission component. This secondary passive buffer deformation is subsequently used to absorb the sudden impact energy parameters generated by the wearable centaur robot's obstacle-crossing swaying, ensuring that these sudden impact energy parameters are preferentially consumed during the elastic deformation process of the softened elastic coupling mechanism, rather than being directly transmitted to the support side where the human body is located along a rigid transmission path.

[0064] Preferably, the specific implementation process of absorbing the sudden impact energy parameters generated by the wearable Centaur robot's obstacle-crossing sway through the secondary passive buffer deformation, thereby blocking the rigid transmission path of the sudden impact energy parameters to the support side where the human object is located, is as follows: After the softened elastic coupling mechanism has undergone the secondary passive buffer deformation, a sudden impact energy parameter absorption and bearing record is first established, and the secondary passive buffer deformation, the force value jump gradient parameter, the real-time drag force feedback data, the rhombic linkage transmission component, the antagonistic spring assembly, the initial deformation scale, and the equivalent physical stiffness are written into the sudden impact energy parameter absorption and bearing record. The sudden impact energy parameter absorption and bearing record is used to include the deformation source of the secondary passive buffer deformation, the force jump source of the force value jump gradient parameter, and the current stiffness state of the softened elastic coupling mechanism into the same energy absorption caliber, so that the subsequently identified sudden impact energy parameters can correspond to the short-term impact process generated by the wearable Centaur robot's obstacle-crossing sway. Subsequently, based on the real-time drag force feedback data, the force value change amplitude within the continuous time window, the deformation displacement change corresponding to the secondary passive buffer deformation, and the elastic reaction force change corresponding to the equivalent physical stiffness, the sudden impact energy parameter is formed. The technical essence of the sudden impact energy parameter is the short-term impact energy characterization transmitted by the wearable centaur robot to the softened elastic coupling mechanism via the main frame and the rhomboid linkage during obstacle-crossing and swaying. The sudden impact energy parameter is not a single impact force peak, but rather impact transmission intensity data jointly expressed by rapid changes in drag force, buffer deformation displacement, and elastic reaction force. Further, after the sudden impact energy parameter is formed, based on the deformation displacement change corresponding to the secondary passive buffer deformation, the sudden impact energy parameter is converted into the additional elastic deformation energy of the antagonistic spring assembly and the force transmission direction adjustment amount of the rhomboid linkage, and the additional elastic deformation energy and the force transmission direction adjustment amount are written into the sudden impact energy parameter absorption and bearing record. The additional elastic deformation energy is used to express the portion of the impact energy absorbed by the antagonistic spring assembly during the secondary passive buffer deformation, and the force transmission direction adjustment amount is used to express the geometric decomposition effect of the rhomboid linkage transmission component on the impact transmission direction during the secondary passive buffer deformation. Subsequently, based on the additional elastic deformation energy and the force transmission direction adjustment amount, the rigid transmission path of the sudden impact energy parameter transmitted to the support side where the human body is located is reduced to form a rigid transmission path reduction result.The technical essence of the rigid transmission path is that the sudden impact energy parameter is directly transmitted to the support side of the human object along the main frame, the rhomboid linkage, and the connection point on the human side without being buffered by the softened elastic coupling mechanism. The rigid transmission path reduction result is used to express the result that the impact intensity of the sudden impact energy parameter being transmitted to the support side of the human object is reduced after being absorbed by the secondary passive buffer deformation. Thus, the secondary passive buffer deformation transfers the sudden impact energy parameter generated by obstacle crossing to the elastic buffering process jointly participated in by the antagonistic spring assembly and the rhomboid linkage, so that the softened elastic coupling mechanism can still provide short-term buffering for sudden force jumps after completing the equivalent physical stiffness update, thereby reducing the rigid traction effect of the wearable centaur robot's obstacle crossing sway on the support side of the human object.

[0065] like Figure 3 As shown, this embodiment of the present application discloses a stiffness adjustment device for a softened elastic coupling mechanism of a wearable centaur robot, comprising: The gait phase determination module is configured to determine the walking gait phase data of a human object in the current detection cycle; The counterweight data determination module is configured to determine the counterweight data of the material carried by the wearable Centaur robot in the current detection cycle; The ground landing feature extraction module is configured to extract the leg support phase ground landing features of the human object based on the walking gait phase data; The working condition level determination module is configured to determine the load working condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The pre-compression parameter determination module is configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism based on the load condition level and the leg support period ground landing characteristics. The softened elastic coupling mechanism connects the human object and the main skeleton of the wearable centaur robot through a diamond-shaped linkage transmission component. The adjustment command generation module is configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The stiffness update control module is configured to send the displacement adjustment command to the micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism.

[0066] like Figure 4 As shown, an electronic device according to an embodiment of this application includes a memory and a processor; The memory stores computer programs; When the processor executes the computer program, it implements the method described in any one of the claims of this application.

[0067] like Figure 5 As shown, this is a computer-readable storage medium according to an embodiment of the present application, wherein the computer-readable storage medium stores a computer program; When the computer program is executed by a processor, it implements the method described in any one of the claims of this application.

[0068] like Figure 6 As shown, this is a wearable centaur robot according to an embodiment of the present application, which includes a main skeleton, a softened elastic coupling mechanism, a micro screw actuator and a stiffness adjustment controller. The softened elastic coupling mechanism includes a rhomboid connecting rod transmission component and an antagonistic spring assembly. The diamond-shaped connecting rod is used to connect the human body object to the main skeleton. The stiffness adjustment controller is configured to determine the walking gait phase data of the human object in the current detection cycle; The stiffness adjustment controller is also configured to determine the weight data of the material carried by the wearable Centaur robot in the current detection cycle; The stiffness adjustment controller is also configured to extract the leg support phase ground contact characteristics of the human object based on the walking gait phase data; The stiffness adjustment controller is also configured to determine the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The stiffness adjustment controller is also configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the load condition level and the leg support period ground contact characteristics; The stiffness adjustment controller is further configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The micro-screw actuator is configured to receive the displacement adjustment command and change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command to update the equivalent physical stiffness of the softened elastic coupling mechanism.

[0069] like Figure 7 As shown, this embodiment of the present application discloses a stiffness adjustment system for a wearable centaur robot with a softened elastic coupling mechanism, which includes a gait phase sensing unit, a counterweight data sensing unit, a stiffness adjustment control unit, a micro screw actuator, and a softened elastic coupling mechanism. The softened elastic coupling mechanism includes a rhomboid connecting rod transmission component and an antagonistic spring assembly. The rhomboid linkage is used to connect the human body object to the main skeleton of the wearable centaur robot. The gait phase sensing unit is configured to determine the walking gait phase data of the human object in the current detection cycle and transmit the walking gait phase data to the stiffness adjustment control unit. The counterweight data sensing unit is configured to determine the counterweight data of the material carried by the wearable Centaur robot in the current detection cycle, and transmit the counterweight data of the material carried to the stiffness adjustment control unit. The stiffness adjustment control unit is configured to extract the leg support phase ground contact characteristics of the human object based on the walking gait phase data; The stiffness adjustment control unit is also configured to determine the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The stiffness adjustment control unit is also configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the load condition level and the leg support period ground contact characteristics; The stiffness adjustment control unit is further configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The micro-screw actuator is configured to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, so as to update the equivalent physical stiffness of the softened elastic coupling mechanism.

[0070] Figures 2-7 For an exemplary description, please refer to the above. Figure 1 This will not be elaborated upon here.

Claims

1. A method for adjusting the stiffness of a softened elastic coupling mechanism in a wearable centaur robot, characterized in that, include: Determine the walking gait phase data of the human object in the current detection cycle; Determine the weight data of the material carried by the wearable Centaur robot during the current detection cycle; Based on the walking gait phase data, extract the leg support phase ground contact characteristics of the human object; Based on the weight data of the load-bearing material, the load condition level of the wearable Centaur robot is determined; Based on the load condition level and the leg support period ground landing characteristics, the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism is determined. The softened elastic coupling mechanism connects the human body object and the main skeleton of the wearable centaur robot through a diamond-shaped linkage transmission component. Based on the target pre-compression displacement parameter, a displacement adjustment command matching the target pre-compression displacement parameter is generated; The displacement adjustment command is sent to the micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism.

2. The method for adjusting the stiffness of a softened elastic coupling mechanism according to claim 1, characterized in that, The determination of the walking gait phase data of the human object in the current detection period includes: The spatial posture of the lower limbs of the human object is collected by an inertial measurement unit deployed on the lower limbs of the human object during the current detection period. The distribution of plantar pressure of the human body during the current detection period is collected by a thin-film array sensor applied to the sole of the foot. Perform coordinate system transformation on the lower limb spatial posture to determine the human lower limb posture data in the human body base reference frame; The human lower limb pose data is appended to the plantar pressure distribution to generate a multi-source sensing fusion feature vector; Based on the multi-source sensing fusion feature vector, the walking gait phase data of the human object in the current detection cycle is determined.

3. The method for adjusting the stiffness of a softened elastic coupling mechanism according to claim 1, characterized in that, The step of extracting the leg stance phase landing features of the human object based on the walking gait phase data includes: Extract the heel strike time parameter from the walking gait phase data; The toe-off time parameter is extracted from the walking gait phase data; Calculate the duration span from heel contact to toe lift-off based on the heel contact time parameter and the toe lift-off time parameter; Based on the duration span, select motion trajectory slices from the walking gait phase data; The extreme value change point features of the trajectory are extracted from the motion trajectory slice data, and then the leg support phase landing characteristics of the human object are determined based on the extreme value change point features.

4. The method for adjusting the stiffness of a softened elastic coupling mechanism according to claim 1, characterized in that, The step of determining the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material includes: Read the baseline unloaded weight value of the wearable centaur robot when it is in an unloaded state; Based on the weight data of the load-bearing material and the baseline unloaded weight value, calculate the weight increment of the wearable Centaur robot. The weight increment value is compared with the load critical node parameters of multiple levels one by one to locate the load matching interval into which the weight increment value falls. The load condition level of the wearable centaur robot is generated based on the identifier code corresponding to the load matching interval.

5. The method for adjusting the stiffness of a softened elastic coupling mechanism according to claim 1, characterized in that, The step of determining the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism based on the load condition level and the leg support period ground landing characteristics includes: Extract the basic connection stiffness parameters corresponding to the load condition level from the multidimensional stiffness mapping association library; Extract the peak landing impact value from the leg support phase landing characteristics; Calculate the ratio between the peak impact value and the standard flat ground walking impact reference value to obtain the stiffness dynamic adjustment compensation factor; The stiffness parameters of the basic connection are numerically corrected using the stiffness dynamic adjustment compensation factor to determine the target required stiffness value. Based on the target stiffness value and the inherent elastic coefficient of the antagonistic spring assembly, the corresponding displacement conversion value is calculated, and then the target pre-compression displacement parameter of the antagonistic spring assembly is determined based on the displacement conversion value.

6. A stiffness adjustment device for a softened elastic coupling mechanism of a wearable centaur robot, characterized in that, include: The gait phase determination module is configured to determine the walking gait phase data of a human object in the current detection cycle; The counterweight data determination module is configured to determine the counterweight data of the material carried by the wearable Centaur robot in the current detection cycle; The ground landing feature extraction module is configured to extract the leg support phase ground landing features of the human object based on the walking gait phase data; The working condition level determination module is configured to determine the load working condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The pre-compression parameter determination module is configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly in the softened elastic coupling mechanism based on the load condition level and the leg support period ground landing characteristics. The softened elastic coupling mechanism connects the human object and the main skeleton of the wearable centaur robot through a diamond-shaped linkage transmission component. The adjustment command generation module is configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The stiffness update control module is configured to send the displacement adjustment command to the micro-screw actuator to drive the micro-screw actuator to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, thereby updating the equivalent physical stiffness of the softened elastic coupling mechanism.

7. An electronic device, characterized in that, Including memory and processor; The memory stores computer programs; When the processor executes the computer program, it implements the method described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program; When the computer program is executed by a processor, it implements the method described in any one of claims 1-5.

9. A wearable centaur robot, characterized in that, It includes the main frame, the softened elastic coupling mechanism, the micro lead screw actuator, and the stiffness adjustment controller; The softened elastic coupling mechanism includes a rhomboid connecting rod transmission component and an antagonistic spring assembly. The diamond-shaped connecting rod is used to connect the human body object to the main skeleton. The stiffness adjustment controller is configured to determine the walking gait phase data of the human object in the current detection cycle; The stiffness adjustment controller is also configured to determine the weight data of the material carried by the wearable Centaur robot in the current detection cycle; The stiffness adjustment controller is also configured to extract the leg support phase ground contact characteristics of the human object based on the walking gait phase data; The stiffness adjustment controller is also configured to determine the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The stiffness adjustment controller is also configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the load condition level and the leg support period ground contact characteristics; The stiffness adjustment controller is further configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The micro-screw actuator is configured to receive the displacement adjustment command and change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command to update the equivalent physical stiffness of the softened elastic coupling mechanism.

10. A stiffness adjustment system for a softened elastic coupling mechanism of a wearable centaur robot, characterized in that, It includes a gait phase sensing unit, a counterweight data sensing unit, a stiffness adjustment and control unit, a micro lead screw actuator, and a softened elastic coupling mechanism; The softened elastic coupling mechanism includes a rhomboid connecting rod transmission component and an antagonistic spring assembly. The rhomboid linkage is used to connect the human body object to the main skeleton of the wearable centaur robot. The gait phase sensing unit is configured to determine the walking gait phase data of the human object in the current detection cycle and transmit the walking gait phase data to the stiffness adjustment control unit. The counterweight data sensing unit is configured to determine the counterweight data of the material carried by the wearable Centaur robot in the current detection cycle, and transmit the counterweight data of the material carried to the stiffness adjustment control unit. The stiffness adjustment control unit is configured to extract the leg support phase ground contact characteristics of the human object based on the walking gait phase data; The stiffness adjustment control unit is also configured to determine the load condition level of the wearable Centaur robot based on the weight data of the load-bearing material. The stiffness adjustment control unit is also configured to determine the target pre-compression displacement parameter of the antagonistic spring assembly based on the load condition level and the leg support period ground contact characteristics; The stiffness adjustment control unit is further configured to generate a displacement adjustment command that matches the target pre-compression displacement parameter based on the target pre-compression displacement parameter. The micro-screw actuator is configured to change the initial deformation scale of the antagonistic spring assembly according to the displacement adjustment command, so as to update the equivalent physical stiffness of the softened elastic coupling mechanism.

Images