Multi-mode motion switching control method and device of intelligent power assisting exoskeleton for mine

Abstract

This disclosure proposes a multimodal motion switching control method and device for a mining intelligent assistive exoskeleton, relating to the field of mining safety equipment technology. The method includes: collecting multimodal sensing data from the wearer, extracting temporal features and inputting them into a lightweight classifier to identify motion patterns and output confidence scores; when the confidence score continuously exceeds a threshold, a mode switching intention is determined; the control subroutine corresponding to the current motion mode is invoked to obtain impedance parameters and the desired trajectory; based on the impedance parameters, the desired trajectory, and the current trajectory, interpolation fusion is performed using a hybrid transition strategy to generate smooth transition joint control commands to control the exoskeleton. This disclosure enables real-time and accurate identification of motion intentions in confined underground environments, smooth and compliant switching between multiple modes, effectively reducing switching impact, ensuring motion continuity and human-machine collaboration safety, and improving the adaptability and overall operational safety of the exoskeleton in confined environments.

Description

Technical Field

[0001] This disclosure relates to the field of mining safety equipment technology, and in particular to a multimodal motion switching control method and device for a mining intelligent assistive exoskeleton. Background Technology

[0002] When working in confined spaces such as underground mines, workers often need to use non-standard gaits such as crouching and lateral movement to pass through low tunnels or gaps between equipment. Research on assistive exoskeletons in related technologies has largely focused on assisting walking on flat ground or rehabilitation training, with relatively simple movement patterns and control strategies mostly designed for stable environments. There is a lack of dedicated multimodal motion definitions and rapid switching schemes for complex and narrow underground scenarios. Pattern recognition is not real-time enough, and mechanical impacts are easily generated during mode switching, affecting the continuity of movements and the safety of human-machine collaboration. Furthermore, maintaining balance and achieving rapid instability protection under confined postures are also key technical challenges faced by existing exoskeletons in this application scenario. Summary of the Invention

[0003] This disclosure aims to at least partially address one of the technical problems in the related art.

[0004] Therefore, the first aspect of this disclosure proposes a multimodal motion switching control method for a mining intelligent assistive exoskeleton, comprising the following steps:

[0005] Real-time acquisition of multimodal sensor data from wearers of intelligent power-assisted exoskeletons used in mining; The multimodal sensing data is preprocessed and features are extracted to construct a temporal feature vector; The time-series feature vector is input into a trained lightweight classifier to identify the wearer's current movement pattern and the identification confidence level corresponding to the current movement pattern; In response to the satisfaction of the mode switching trigger condition, the control subroutine corresponding to the current motion mode is invoked to obtain the time-varying impedance stiffness parameter, time-varying impedance damping parameter and target desired trajectory corresponding to the current motion mode. The mode switching trigger condition includes the recognition confidence exceeding a preset threshold. Based on the time-varying impedance stiffness parameter, the time-varying impedance damping parameter, the target desired trajectory, and the current trajectory, interpolation fusion is performed through a hybrid transition strategy to generate joint control commands. The joints of the mining intelligent power exoskeleton are controlled according to the joint control commands.

[0006] In some embodiments of this disclosure, the multimodal sensing data includes at least two of the following: surface electromyography signals, plantar pressure distribution signals, joint angle signals, and inertial measurement unit signals.

[0007] In some embodiments of this disclosure, the mode switching triggering conditions further include at least one of the following: receiving a one-click switching command issued by the wearer through a physical or virtual button on the mining intelligent assistive exoskeleton; or receiving a remote switching command from an external control system.

[0008] In some embodiments of this disclosure, when the one-click switching instruction or the remote switching instruction is received, before executing the step of calling the control subroutine corresponding to the current motion mode, a posture verification step is further included: acquiring current environmental perception data and the wearer's posture data; determining whether the environmental perception data and the posture data meet the safety access conditions of the current motion mode; if the environmental perception data and the posture data meet the safety access conditions, then switching is allowed; if the environmental perception data and the posture data do not meet the safety access conditions, then switching is refused and a prompt is issued to the wearer.

[0009] In some embodiments of this disclosure, the safety access conditions include at least one of the following: the wearer's head height is lower than the height threshold corresponding to the current movement mode; the distance between the wearer and surrounding obstacles is greater than the safety distance threshold; and the center of pressure on the sole of the foot is within the support surface and the volatility is lower than the stability threshold.

[0010] In some embodiments of this disclosure, the lightweight classifier is trained by: collecting multimodal sensor data samples from multiple wearers under different motion modes and load conditions in a simulated downhole working environment; labeling the data samples to obtain mode labels; processing the data samples and labels using data augmentation methods to construct a training set; and using the training set to train the lightweight classifier.

[0011] In some embodiments of this disclosure, the joint control commands are generated based on the time-varying impedance stiffness parameter, the time-varying impedance damping parameter, the target desired trajectory, and the current trajectory through interpolation fusion using a hybrid transition strategy, using the following formula:

[0012] in, The joint control command at time t. It is a smooth interpolation function from 0 to 1. This refers to the joint control parameters before mode switching. This refers to the joint control quantity corresponding to the current motion mode. Let be the desired joint angle at time t. Let be the actual joint angle at time t. Let be the expected joint angular velocity at time t. The actual joint angular velocity at time t. Let be the time-varying impedance stiffness parameter at time t. Let be the time-varying impedance damping parameter at time t.

[0013] In some embodiments of this disclosure, the method further includes: acquiring plantar pressure distribution signals and inertial data collected by the inertial measurement unit in the mining intelligent assistive exoskeleton in real time; determining whether the wearer is in an unstable state based on the plantar pressure distribution signals and the inertial data; and triggering a priority safety protection strategy when the wearer is determined to be in an unstable state. The priority safety protection strategy includes: reducing joint stiffness and applying a supporting torque to the force-bearing end, locking key joints within a preset time range, and triggering an audio-visual vibration prompt.

[0014] A second aspect of this disclosure provides a multimodal motion switching control device for a mining intelligent assistive exoskeleton, comprising: The data acquisition module is used to collect multimodal sensor data from wearers of the intelligent power exoskeleton for mining in real time. The feature extraction module is used to preprocess and extract features from the multimodal sensing data to construct a temporal feature vector; The recognition module is used to input the temporal feature vector into a trained lightweight classifier to recognize the wearer's current movement pattern and the recognition confidence level corresponding to the current movement pattern. The acquisition module is used to call the control subroutine corresponding to the current motion mode in response to the satisfaction of the mode switching trigger condition, so as to obtain the time-varying impedance stiffness parameter, time-varying impedance damping parameter and target expected trajectory corresponding to the current motion mode. The mode switching trigger condition includes the recognition confidence exceeding a preset threshold. The generation module is used to generate joint control commands by interpolating and fusing the time-varying impedance stiffness parameter, the time-varying impedance damping parameter, the target desired trajectory and the current trajectory through a hybrid transition strategy. The control module is used to control the joints of the mining intelligent power exoskeleton according to the joint control commands.

[0015] A third aspect of this disclosure provides an electronic device, including: a processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method described in the first aspect above.

[0016] The multimodal motion switching control method for the intelligent power-assisted exoskeleton for mining disclosed herein, based on multimodal sensor data and a lightweight classifier, can accurately identify the wearer's complex movement intentions in the confined space of an underground mine in real time. It effectively filters out momentary misjudgments through confidence threshold judgment, ensuring the reliability of mode switching decisions. By calling a predefined subroutine matching the target mode, it quickly responds and obtains precise motion control parameters. A hybrid transition strategy that integrates target and current parameters is adopted to generate joint control commands, achieving smooth and compliant switching between different motion modes. This significantly reduces mechanical shock and abruptness during switching, ensuring the wearer's motion continuity and comfort.

[0017] Additional aspects and advantages of this disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this disclosure. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this disclosure will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, in which: Figure 1 A flowchart illustrating a multimodal motion switching control method for a mining intelligent assistive exoskeleton provided in this embodiment of the present disclosure; Figure 2 A reference schematic diagram showing the total transition time corresponding to different motion modes provided in the embodiments of this disclosure; Figure 3 A reference schematic diagram showing the impedance parameter range corresponding to different motion modes provided in the embodiments of this disclosure; Figure 4 This is a schematic diagram of a multimodal motion switching control device for a mining intelligent assistive exoskeleton provided in an embodiment of this disclosure. Detailed Implementation

[0019] Embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0020] Specifically, the following describes a multimodal motion switching control method and apparatus for a mining intelligent power-assisted exoskeleton according to embodiments of the present disclosure, with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart illustrating a multimodal motion switching control method for a mining intelligent assistive exoskeleton provided in an embodiment of this disclosure. Figure 1 As shown, the multimodal motion switching control method of this intelligent assistive exoskeleton for mining can include the following steps: Step 101: Real-time acquisition of multimodal sensor data from wearers of the mining intelligent assistive exoskeleton.

[0022] In some embodiments of this disclosure, multimodal sensing data may include at least two of the following: surface electromyography signals, plantar pressure distribution signals, joint angle signals, and inertial measurement unit signals.

[0023] As an example, surface electromyography (sEMG) signals from the bilateral hips, quadriceps, and hamstrings can be acquired at a sampling rate of 1000 Hz; plantar pressure signals distributed across multiple areas of the soles of both feet can be acquired, with 4–8 pressure sensing units per foot at a sampling rate of 200 Hz; angle encoder signals from the hip, knee, and ankle joints can be acquired with a resolution ≥0.1° and a sampling rate of 200 Hz; and inertial measurement units (IMUs) located in the chest, pelvis, and legs can be acquired at 200 Hz each for chest / pelvis / legs for attitude / angular velocity detection.

[0024] In some embodiments of this disclosure, the mining intelligent power exoskeleton may employ a lightweight, structured aluminum alloy / carbon fiber composite frame, with adjustable-length linkages at the joints, and driveable joint modules at the hip and knee, symmetrically arranged on both sides. Mechanical limiting and quick-release mechanisms are provided at the joints. Each main joint uses a brushless DC motor with a series elastic element (SEA) or a force / torque sensor and a planetary gear reducer to achieve measurable torque and a certain degree of compliance. Real-time CAN communication can be used between the main controller and the drive; the sensor can communicate with the main controller via wired / low-power wireless; the main controller adopts an ARM Cortex-A main controller + MCU slave architecture.

[0025] Step 102: Preprocess and extract features from the multimodal sensing data to construct a temporal feature vector.

[0026] In one implementation, preprocessing of multimodal sensing data may include: rectifying, enveloping, and normalizing sEMG; estimating the partitioned torque / center pressure point (CoP) for pressure; performing low-pass filtering and attitude estimation for angle / IMU; and constructing time-domain / frequency-domain / time-series feature vectors within a short time window (100 ms).

[0027] In one implementation, the input features used as a lightweight classifier may include: sEMG envelope mean, RMS, ZC (zero crossing rate), MF (intermediate frequency), etc.; plantar pressure distribution and CoP location; hip / knee / ankle angle and angular velocity; IMU tilt angle and angular velocity; short-window (200 ms) temporal feature vectors superimposed with the past 3 windows as input.

[0028] Step 103: Input the temporal feature vector into the trained lightweight classifier to identify the wearer's current motion mode and the corresponding recognition confidence level.

[0029] In some embodiments of this disclosure, multiple motion modes can be preset, including at least one of forward walking, crouching walking, lateral movement, low-position crawling, standard straight walking, and stationary walking. Furthermore, motion subroutines and support strategies for constrained postures corresponding to each motion mode are predefined.

[0030] Optionally, the lightweight classifier can be a lightweight convolutional neural network-long short-term memory network combined model (CNN-LSTM) or a random forest model. The temporal feature vector is input into the trained lightweight classifier to identify the wearer's current motion pattern and the corresponding recognition confidence level. Furthermore, to eliminate jitter in the recognition results caused by momentary misclassification, the system performs temporal smoothing processing on the recognition results of multiple consecutive frames (e.g., using a sliding window majority voting method), outputting a temporally smoothed result. This temporally smoothed result serves as the final determination of the current motion pattern, used for subsequent switching logic judgments, achieving low-latency, high-accuracy pattern recognition.

[0031] In some embodiments of this disclosure, the lightweight classifier is obtained through the following steps: S11, in a simulated downhole working environment, collects multimodal sensor data samples from multiple wearers under different movement modes and load conditions, such as normal / carrying tools / carrying weight, etc. S12, label the data samples to obtain pattern labels; S13 uses data augmentation methods to process data samples and labels, and constructs a training set to improve robustness; S14, train the lightweight classifier using the training set.

[0032] Optionally, in some embodiments of this disclosure, a lightweight CNN-LSTM model or a random forest classifier supplemented by a sliding window voting post-processing method can be used to reduce instantaneous misjudgments in real-time recognition. The model can be deployed in ONNX format and run on the main control unit with an acceleration library to meet the design goal of recognition latency <150ms.

[0033] Step 104: In response to the mode switching trigger condition being met, the control subroutine corresponding to the current motion mode is called to obtain the time-varying impedance stiffness parameter, time-varying impedance damping parameter and target desired trajectory corresponding to the current motion mode. The mode switching trigger condition includes identifying confidence exceeding a preset threshold.

[0034] In one implementation, to improve the accuracy of intent recognition, the current motion mode determined based on the time-series smoothing results is different from the motion mode at the previous moment, and the recognition confidence exceeds a preset threshold and continues for a period of time, then the current mode switching trigger condition is determined to be met.

[0035] In some embodiments of this disclosure, in addition to determining the wearer's intention to switch modes by identifying confidence levels, the mode switching triggering condition may also include at least one of the following: It receives a one-click switching command from the wearer via physical or virtual buttons on the mining smart assistive exoskeleton; for example, the user can trigger it by long-pressing (>300 ms) or double-clicking (quick switching) (user can customize); Receive remote switching commands from external control systems.

[0036] The mode switching steps 104-106 can be triggered by three methods: determining the wearer's intention to switch modes by identifying confidence level, one-click switching command, and remote switching command.

[0037] When a one-click switching command or a remote switching command is received, the wearer's current posture can be verified before executing the step of calling the control subroutine corresponding to the current motion mode: S21, acquire current environmental perception data and wearer's posture data; S22, determine whether the environmental perception data and attitude data meet the safety access conditions of the current motion mode; S23, If the environmental perception data and attitude data meet the safety access conditions, then switching is allowed; S24. If the environmental perception data and posture data do not meet the safety access conditions, the switch will be refused and a prompt will be issued to the wearer.

[0038] In some embodiments of this disclosure, security access conditions may include at least one of the following: The wearer's head height is below the height threshold corresponding to the current movement mode; The distance between the wearer and surrounding obstacles is greater than the safe distance threshold; The center of pressure on the sole of the foot is within the support surface and the volatility is below the stability threshold.

[0039] In one example, taking squatting as an example, the security access conditions can be as follows:

[0040] In some embodiments of this disclosure, different motion modes may correspond to different total transition times (i.e., the time required to completely switch from an old motion mode to a new motion mode) and impedance parameter ranges. The total transition time for switching motion modes can be set to different values ​​according to different motion mode switching types to ensure a safe and smooth transition. The impedance parameter range is used to reflect the support requirements of this mode. After identifying the wearer's current motion mode, the time-varying impedance stiffness parameter and time-varying impedance damping parameter corresponding to different moments during the motion mode switching process can be determined according to the target impedance parameter range corresponding to the current motion mode and the total transition time, so that the impedance parameters smoothly transition with time during the mode switching process. For example, the time-varying impedance stiffness parameter... Can be followed The transition from lower stiffness to target stiffness allows for greater control over the wearer during the initial switching phase, while also mitigating sudden impacts.

[0041] As an example, the impedance parameter variation strategy used to determine the time-varying impedance stiffness parameter and the time-varying impedance damping parameter can be as follows:

[0042]

[0043]

[0044]

[0045] in, For time-varying impedance stiffness parameters, This is the minimum impedance stiffness parameter. Here, t is the maximum impedance stiffness parameter, t is the elapsed time after the start of the motion mode transition, and T is the total transition time corresponding to the current motion mode, where 0 ≤ t ≤ T. These are time-varying impedance damping parameters. The minimum impedance damping parameter, This represents the maximum impedance damping parameter. , , , This refers to the impedance parameter range corresponding to the current motion mode.

[0046] Figure 2 This is a reference schematic diagram showing the total transition time corresponding to different motion modes, provided in an embodiment of this disclosure. Figure 3 This is a reference schematic diagram showing the impedance parameter range corresponding to different motion modes provided in the embodiments of this disclosure.

[0047] Step 105: Based on the time-varying impedance stiffness parameter, time-varying impedance damping parameter, target desired trajectory and current trajectory, interpolation fusion is performed through a hybrid transition strategy to generate joint control commands.

[0048] In some embodiments of this disclosure, based on time-varying impedance stiffness parameters, time-varying impedance damping parameters, the target desired trajectory, and the current trajectory, interpolation fusion is performed using a hybrid transition strategy to generate joint control commands based on the following formula:

[0049] in, The joint control command at time t. This is a smooth interpolation function from 0 to 1 to achieve a smooth transition between the old and new control parameters. This refers to the joint control parameters before the mode switch (i.e., under the old motion mode). This refers to the joint control quantity corresponding to the current motion mode. Let be the desired joint angle at time t. Let be the actual joint angle at time t. Let be the expected joint angular velocity at time t. The actual joint angular velocity at time t. Let be the time-varying impedance stiffness parameter at time t. Let be the time-varying impedance damping parameters at time t. The desired joint angle and desired joint angular velocity are the target desired trajectory corresponding to the current motion mode, while the actual joint angle and actual joint angular velocity are the current trajectory of the joint.

[0050] Optionally, the transition function It can employ cubic polynomials or cosine curves, with the entire switching transition time adjustable within the range of 0.1 to 0.5 seconds. The hybrid transition strategy reduces instantaneous joint impact and obstruction to the wearer's movement intentions, ensuring a stable and smooth switching.

[0051] Step 106: Perform joint control on the mining intelligent power-assisted exoskeleton according to the joint control commands.

[0052] After receiving the joint control command, it is distributed to the torque / speed loops of each joint in the mining intelligent power-assisted exoskeleton through the low-level controller.

[0053] In some embodiments of this disclosure, during the joint control of the exoskeleton according to joint control commands, real-time instability assessment can also be performed. If instability is detected, an emergency protection process is initiated to ensure personnel safety. By real-time monitoring of key stability parameters and triggering active protection in the event of instability, the overall human-machine safety and system robustness are effectively improved in confined, unstructured downhole environments.

[0054] In one implementation, the system can acquire plantar pressure distribution signals and inertial data collected by the inertial measurement unit in the mining intelligent power-assisted exoskeleton in real time. Based on the plantar pressure distribution signals and inertial data, it can determine whether the wearer is in an unstable state. If the wearer is determined to be in an unstable state, a priority safety protection strategy is triggered. This priority safety protection strategy includes: reducing joint stiffness and applying supporting torque to the force-bearing end, locking key joints within a preset time range, and triggering audible and visual vibration alerts. Furthermore, it can also disconnect the power assist, allow manual disengagement, or trigger a ground rescue signal.

[0055] Optionally, the instability state may include at least one of the following conditions: The center of pressure on the sole of the foot continuously extends beyond the preset support surface boundary; The inertial measurement unit detected a torso tilt angle that exceeded the safe angle threshold; The plantar pressure on one side drops sharply beyond a set ratio within a short period of time; The structural framework of this disclosure can be: High-level: Pattern recognizer + pattern manager (decision making and transition planning); Middle layer: Trajectory generation / impedance parameter generator (based on target pattern generation) With impedance parameters K(t), B(t)); Lower layer: Joint closed-loop control (torque / position controller) and safety monitoring.

[0056] Taking the squatting movement mode as an example, the mode can be defined to lower the pelvic height to a preset low position, with a larger relative knee / hip angle; the system should maintain support or partial support to reduce muscle burden. The mid-level trajectory ensures gait stability while lowering the height, and uses CoP monitoring to avoid slippage caused by backward / forward leaning.

[0057] By implementing the embodiments of this disclosure, based on multimodal sensing data and a lightweight classifier, the complex movement intentions of the wearer in the confined space of an underground well can be identified in real time and accurately. A confidence threshold is used to effectively filter out momentary misjudgments, ensuring the reliability of mode switching decisions. By calling a predefined subroutine matching the target mode, precise motion control parameters are quickly obtained and responded to. A hybrid transition strategy that fuses target and current parameters is used to generate joint control commands, achieving smooth and compliant switching between different movement modes. This significantly reduces mechanical shock and abruptness during switching, ensuring the wearer's motion continuity and comfort.

[0058] Figure 2 This is a schematic diagram of a multimodal motion switching control device for a mining intelligent assistive exoskeleton provided in an embodiment of this disclosure. Figure 2As shown, the multimodal motion switching control device of the intelligent assistive exoskeleton for mining can include: a data acquisition module 201, a feature extraction module 202, a recognition module 203, an acquisition module 204, a generation module 205, and a control module 206.

[0059] The acquisition module 201 is used to acquire multimodal sensor data of the wearer of the mining intelligent assistive exoskeleton in real time; Feature extraction module 202 is used to preprocess and extract features from multimodal sensing data to construct temporal feature vectors; The recognition module 203 is used to input the temporal feature vector into a trained lightweight classifier to recognize the wearer's current motion mode and the recognition confidence level corresponding to the current motion mode; The acquisition module 204 is used to call the control subroutine corresponding to the current motion mode in response to the satisfaction of the mode switching trigger condition, so as to obtain the time-varying impedance stiffness parameter, time-varying impedance damping parameter and target expected trajectory corresponding to the current motion mode. The mode switching trigger condition includes the identification confidence level exceeding a preset threshold. The generation module 205 is used to generate joint control commands by interpolating and fusing time-varying impedance stiffness parameters, time-varying impedance damping parameters, target desired trajectory and current trajectory through a hybrid transition strategy. The control module 206 is used to control the joints of the mining intelligent power-assisted exoskeleton according to the joint control commands.

[0060] In some embodiments of this disclosure, the multimodal sensing data includes at least two of the following: surface electromyography signals, plantar pressure distribution signals, joint angle signals, and inertial measurement unit signals.

[0061] In some embodiments of this disclosure, the mode switching triggering conditions further include at least one of the following: receiving a one-click switching command issued by the wearer through a physical button or virtual button on the mining intelligent assistive exoskeleton; or receiving a remote switching command from an external control system.

[0062] In some embodiments of this disclosure, such as Figure 2 Based on the illustrated embodiment, the multimodal motion switching control device for a mining intelligent power-assisted exoskeleton may include a posture verification module; wherein, the posture verification module is used to: upon receiving a one-click switching command or a remote switching command, acquire current environmental perception data and the wearer's posture data; determine whether the environmental perception data and posture data meet the safety access conditions of the current motion mode; if the environmental perception data and posture data meet the safety access conditions, allow switching; if the environmental perception data and posture data do not meet the safety access conditions, refuse switching and issue a prompt to the wearer.

[0063] In some embodiments of this disclosure, the safety access conditions include at least one of the following: the wearer's head height is lower than the height threshold corresponding to the current movement mode; the distance between the wearer and surrounding obstacles is greater than the safety distance threshold; and the center of pressure on the sole of the foot is within the support surface and the volatility is lower than the stability threshold.

[0064] In some embodiments of this disclosure, the lightweight classifier is trained by: collecting multimodal sensor data samples from multiple wearers under different motion modes and load conditions in a simulated downhole working environment; labeling the data samples to obtain mode labels; using data augmentation methods to process the data samples and labels to construct a training set; and using the training set to train the lightweight classifier.

[0065] In some embodiments of this disclosure, based on time-varying impedance stiffness parameters, time-varying impedance damping parameters, the target desired trajectory, and the current trajectory, interpolation fusion is performed using a hybrid transition strategy to generate joint control commands based on the following formula:

[0066] in, The joint control command at time t. It is a smooth interpolation function from 0 to 1. This refers to the joint control parameters before mode switching. This represents the joint control parameters corresponding to the current motion mode. Let be the desired joint angle at time t. Let be the actual joint angle at time t. Let be the expected joint angular velocity at time t. The actual joint angular velocity at time t. Let be the time-varying impedance stiffness parameter at time t. Let be the time-varying impedance damping parameter at time t.

[0067] In some embodiments of this disclosure, such as Figure 2 Based on the illustrated embodiment, the multimodal motion switching control device for the mining intelligent power-assisted exoskeleton may include an instability detection module; wherein, the instability detection module is used to: acquire plantar pressure distribution signals and inertial data collected by the inertial measurement unit in the mining intelligent power-assisted exoskeleton in real time; determine whether the wearer is in an unstable state based on the plantar pressure distribution signals and inertial data; and trigger a priority safety protection strategy when it is determined that the wearer is in an unstable state; the priority safety protection strategy includes: reducing joint stiffness and applying supporting torque to the force-bearing end, locking key joints within a preset time range and triggering audio-visual vibration prompts.

[0068] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0069] To implement the above embodiments, this disclosure also proposes an electronic device, including: a processor and a memory communicatively connected to the processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method provided in the foregoing embodiments. In some embodiments, the electronic device may be a mining intelligent assistive exoskeleton.

[0070] To implement the above embodiments, this disclosure also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the foregoing embodiments.

[0071] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0072] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0073] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of this disclosure pertain.

[0074] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0075] It should be understood that various parts of this disclosure can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0076] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0077] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0078] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present disclosure. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present disclosure.

Claims

1. A multimodal motion switching control method for a mining intelligent assistive exoskeleton, characterized in that, Includes the following steps: Real-time acquisition of multimodal sensor data from wearers of intelligent power-assisted exoskeletons used in mining; The multimodal sensing data is preprocessed and features are extracted to construct a temporal feature vector; The time-series feature vector is input into a trained lightweight classifier to identify the wearer's current movement pattern and the identification confidence level corresponding to the current movement pattern; In response to the satisfaction of the mode switching trigger condition, the control subroutine corresponding to the current motion mode is invoked to obtain the time-varying impedance stiffness parameter, time-varying impedance damping parameter and target desired trajectory corresponding to the current motion mode. The mode switching trigger condition includes the recognition confidence exceeding a preset threshold. Based on the time-varying impedance stiffness parameter, the time-varying impedance damping parameter, the target desired trajectory, and the current trajectory, interpolation fusion is performed through a hybrid transition strategy to generate joint control commands. The joints of the mining intelligent power exoskeleton are controlled according to the joint control commands.

2. The method according to claim 1, characterized in that, The multimodal sensing data includes at least two of the following: surface electromyography signals, plantar pressure distribution signals, joint angle signals, and inertial measurement unit signals.

3. The method according to claim 1, characterized in that, The mode switching trigger condition also includes at least one of the following: Receives a one-click switching command from the wearer via physical or virtual buttons on the mining intelligent power exoskeleton; Receive remote switching commands from external control systems.

4. The method according to claim 3, characterized in that, When the one-click switching command or the remote switching command is received, before executing the step of calling the control subroutine corresponding to the current motion mode, an attitude verification step is also included: Acquire current environmental perception data and the wearer's posture data; Determine whether the environmental perception data and the posture data meet the safety access conditions of the current motion mode; If the environmental perception data and the attitude data meet the security access conditions, then switching is permitted; If the environmental perception data and the posture data do not meet the security access conditions, the switch will be refused and a prompt will be issued to the wearer.

5. The method according to claim 4, characterized in that, The security access conditions include at least one of the following: The wearer's head height is lower than the height threshold corresponding to the current movement mode; The distance between the wearer and surrounding obstacles is greater than the safe distance threshold; The center of pressure on the sole of the foot is within the support surface and the volatility is below the stability threshold.

6. The method according to claim 1, characterized in that, The lightweight classifier is trained in the following way: In a simulated downhole working environment, multimodal sensor data samples were collected from multiple wearers under different motion modes and load conditions. The data samples are labeled to obtain pattern labels; The data samples and labels are processed using data augmentation methods to construct a training set; The lightweight classifier is trained using the training set.

7. The method according to claim 1, characterized in that, Based on the time-varying impedance stiffness parameter, the time-varying impedance damping parameter, the target desired trajectory, and the current trajectory, interpolation fusion is performed through a hybrid transition strategy to generate joint control commands based on the following formula: in, The joint control command at time t. It is a smooth interpolation function from 0 to 1. This refers to the joint control parameters before mode switching. This refers to the joint control quantity corresponding to the current motion mode. Let be the desired joint angle at time t. Let be the actual joint angle at time t. Let be the expected joint angular velocity at time t. The actual joint angular velocity at time t. Let be the time-varying impedance stiffness parameter at time t. Let be the time-varying impedance damping parameter at time t.

8. The method according to any one of claims 1-7, characterized in that, Also includes: Real-time acquisition of plantar pressure distribution signals and inertial data collected by the inertial measurement unit in the mining intelligent power exoskeleton; The wearer is determined to be in an unstable state based on the plantar pressure distribution signal and the inertial data. When the wearer is determined to be in an unstable state, a priority safety protection strategy is triggered. The priority safety protection strategy includes: reducing joint stiffness, applying supporting torque to the force-bearing end, locking key joints within a preset time range, and triggering audio-visual vibration prompts.

9. A multimodal motion switching control device for a mining intelligent assistive exoskeleton, characterized in that, include: The data acquisition module is used to collect multimodal sensor data from wearers of the intelligent power exoskeleton for mining in real time. The feature extraction module is used to preprocess and extract features from the multimodal sensing data to construct a temporal feature vector; The recognition module is used to input the temporal feature vector into a trained lightweight classifier to recognize the wearer's current movement pattern and the recognition confidence level corresponding to the current movement pattern. The acquisition module is used to call the control subroutine corresponding to the current motion mode in response to the satisfaction of the mode switching trigger condition, so as to obtain the time-varying impedance stiffness parameter, time-varying impedance damping parameter and target expected trajectory corresponding to the current motion mode. The mode switching trigger condition includes the recognition confidence exceeding a preset threshold. The generation module is used to generate joint control commands by interpolating and fusing the time-varying impedance stiffness parameter, the time-varying impedance damping parameter, the target desired trajectory and the current trajectory through a hybrid transition strategy. The control module is used to control the joints of the mining intelligent power exoskeleton according to the joint control commands.

10. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-8.

Images