Control method and system for lower extremity exoskeleton assistance device for tower climbing

Abstract

A control method and system of a lower limb exoskeleton assisting device for tower climbing, the method comprising: electromyographic signal processing and muscle activation index calculation; double-channel LSTM neural network for online prediction of motion intention; tower cross arm structure edge identification and safe climbing path generation; asymmetric impedance function design and joint stiffness damping parameter optimization; metabolic consumption evaluation model and minimum energy consumption assisting curve solution; three-level control cycle setting and hardware acceleration for real-time calculation. Based on the above method, the modules of the exoskeleton assisting device work cooperatively; through accurate data collection by sensors, flexible adjustment of the power and execution modules according to instructions, efficient operation of the control and operation modules, and stable power supply of the energy module, the safety and efficiency of tower climbing operations are significantly improved, and the burden on workers is reduced.

Description

Technical Field

[0001] This invention relates to the field of exoskeleton assistive device control technology, specifically a control method and system for a lower limb exoskeleton assistive device used for climbing iron towers. Background Technology

[0002] With the increasing demand for industrial and power facility maintenance, the safety and efficiency of tower climbing operations are receiving growing attention. As a crucial piece of equipment for enhancing climbing capabilities, the technological development of lower limb exoskeletons is of paramount importance.

[0003] Traditional lower limb exoskeletons often rely on data from a single type of sensor, such as surface electromyography (EMG) sensors or plantar pressure sensors. EMG signals alone are susceptible to external electromagnetic interference and cannot comprehensively reflect movement status; plantar pressure data alone cannot accurately capture subtle muscle changes, resulting in low accuracy in recognizing movement intentions. This makes it difficult for exoskeletons to accurately match user movements during climbing operations, leading to clumsy climbing motions, low efficiency, and severely impacting work progress.

[0004] Existing exoskeletons lack comprehensive perception and effective response mechanisms for complex tower environments. Most devices are not equipped with high-precision environmental sensing devices such as lidar, making it difficult to accurately identify critical environmental information such as the edges of the tower's crossarm structure and to plan safe climbing paths in advance. In terms of joint control, a fixed control mode is used, which cannot adjust joint stiffness and damping parameters in real time according to different forces and movement stages during climbing. When encountering emergencies or complex climbing conditions, user safety cannot be guaranteed in a timely manner, increasing operational risks.

[0005] Traditional exoskeletons lack effective energy consumption optimization algorithms and cannot dynamically adjust the assist output based on the user's real-time movement status and operational needs. During prolonged tower climbing operations, excessive energy consumption leads to insufficient battery life, and frequent charging or battery replacements severely impact the continuity and efficiency of the operation. Furthermore, traditional devices lack reasonable strategies for allocating battery power, further exacerbating the energy consumption problem.

[0006] Given the many shortcomings of traditional lower limb exoskeletons in terms of motion intention recognition, safety assurance, and energy management, there is an urgent need for an innovative lower limb exoskeleton assistive device and algorithm to improve the safety, efficiency, and endurance of tower climbing operations and meet actual operational needs. Summary of the Invention

[0007] In view of the above problems, the present invention is proposed to provide a control method and system for a lower limb exoskeleton assistive device for tower climbing that overcomes or at least partially solves the above problems.

[0008] To address the aforementioned technical problems, the embodiments of this application disclose the following technical solutions:

[0009] In a first aspect, embodiments of the present invention disclose a control method for a lower limb exoskeleton assistive device for climbing iron towers, comprising:

[0010] The user's electromyography (EMG) signals and plantar pressure distribution data are collected in real time using surface electromyography (EMG) sensors and plantar pressure sensors.

[0011] The electromyographic signal is decomposed to calculate the muscle activation index; a dual-channel LSTM neural network is constructed to realize online prediction of the user's movement intention.

[0012] Collect tower point cloud data, identify the edge of the tower crossarm structure based on the tower point cloud data, generate a safe climbing path through cubic spline interpolation, calculate the minimum distance threshold between the user's feet and the workspace and obstacles, and establish a gradient sensitivity model to assist in motion safety decision-making.

[0013] An asymmetric impedance function is designed, and a model predictive control algorithm is used to generate the desired joint trajectory in the upper layer of the gradient sensitivity model. In the lower layer of the gradient sensitivity model, the joint stiffness and damping parameters are optimized in real time by an improved particle swarm optimization algorithm, and a nonlinear disturbance observer is introduced to compensate for the hysteresis effect of the magnetorheological actuator.

[0014] A metabolic consumption assessment model was established, and a rolling time-domain optimization algorithm was used to solve the minimum energy consumption assist curve. A safety threshold constraint was set to limit joint power overshoot.

[0015] The system is configured with a three-stage control cycle, including actuator current loop control in the first stage, impedance parameter adjustment in the second stage, and path planning update in the third stage. Real-time calculations are achieved through FPGA and GPU hardware acceleration modules.

[0016] Furthermore, the electromyographic signal is decomposed to calculate the muscle activation index. Specifically, this involves: acquiring electromyographic signals of the main muscle groups of the lower limbs during climbing using a surface electromyography sensor; decomposing the electromyographic signal x(t) to obtain the intrinsic mode function (IMF). i (t), i = 1, 2, 3, ..., n, select the first 6 eigenmode functions; adopt the improved Teager energy operator TKO(x[n]) = x 2 [n]-x[n-1]x[n+1], calculate the energy of each intrinsic mode, and sum them by weight to obtain the muscle activation index, where x[n] is the electromyographic signal sampling sequence.

[0017] Furthermore, a dual-channel LSTM neural network is constructed to realize online prediction of the user's movement intention. The specific method includes: constructing a dual-channel LSTM neural network, the first channel processing muscle activation indicators to predict the direction of the user's lower limb joint movement, the second channel parsing plantar pressure distribution data, determining the climbing stage through the plantar pressure distribution data, and dynamically weighting multi-sensor data through an attention mechanism to realize online prediction of the user's movement intention.

[0018] Furthermore, the multi-sensor data is dynamically weighted through an attention mechanism. The specific method includes: using an attention mechanism, where the output feature of the first channel is F1 and the output feature of the second channel is F2, and the weights α1 and α2 are calculated by the attention model, α1 + α2 = 1, to obtain the fused feature F = α1F1 + α2F2.

[0019] Furthermore, point cloud data of the tower is collected using a lidar mounted on the exoskeleton. By improving the RANSAC algorithm, a linear model is assumed by randomly sampling from the point cloud data. The distance from other points to the model is calculated to determine the interior points, and the edge of the tower crossarm structure is found iteratively. Using the starting point and the target climbing point as endpoints, a safe climbing path is generated by cubic spline interpolation. Based on the exoskeleton structure and kinematic parameters, the minimum distance threshold between the reachable workspace and obstacles at the feet is calculated, and a gradient sensitivity model is established to assist in safety decision-making.

[0020] Furthermore, the expression for the asymmetric impedance function is as follows: K d B d For asymmetric stiffness and damping coefficient, K d B d Adjustments are made according to different movement directions and climbing stages. The model predictive control algorithm is based on the exoskeleton dynamics model and the current state to predict the system response and solve the optimization problem to generate the desired joint trajectory. The improved particle swarm optimization algorithm uses the joint stiffness damping parameters as the particle position to iteratively optimize, and introduces a nonlinear disturbance observer to compensate for the hysteresis effect of the magnetorheological actuator.

[0021] Furthermore, the metabolic consumption assessment model is expressed as P. m =f(v,a,F) m A), where v is the user's velocity, a is the acceleration, and F is the acceleration. m F to assist exoskeleton m A represents the muscle activation index, and P represents the muscle activation index. m To address energy consumption, a rolling time-domain optimization algorithm is employed to predict the future system state starting from the current moment, solve for the exoskeleton assistance curve that minimizes metabolic consumption, and simultaneously set a safety threshold to constrain joint power.

[0022] Furthermore, the first-stage actuator current loop control is a 1kHz actuator current loop control, which precisely regulates the exoskeleton output force by controlling the magnetorheological actuator current at a 1kHz frequency; the second-stage impedance parameter adjustment is a 100Hz impedance parameter adjustment, which adjusts the joint stiffness and damping parameters based on the motion state and intention; the third-stage path planning update is a 10Hz path planning update, which updates the safe climbing path based on LiDAR data and user status; the sensor data acquisition logic is implemented using an FPGA, and large-scale matrix operations are calculated using a GPU to process the LSTM neural network.

[0023] Secondly, embodiments of the present invention disclose a control system for a lower limb exoskeleton assistive device for tower climbing, comprising: an electromyography signal and plantar pressure distribution data acquisition unit, an online prediction unit for movement intention, a tower crossarm structure edge and safe climbing path generation unit, an asymmetric impedance function design and joint stiffness damping parameter optimization unit, a metabolic consumption assessment model and minimum energy consumption assist curve solving unit, and a three-level control cycle setting and hardware acceleration unit; wherein:

[0024] The electromyography (EMG) signal and plantar pressure distribution data acquisition unit is used to acquire the user's EMG signal and plantar pressure distribution data in real time through surface EMG sensors and plantar pressure sensors.

[0025] An online prediction unit for movement intention is used to decompose electromyographic signals and calculate muscle activation indices; a dual-channel LSTM neural network is constructed to realize online prediction of user movement intention;

[0026] The tower crossarm structure edge and safe climbing path generation unit is used to decompose electromyographic signals and calculate muscle activation index; a dual-channel LSTM neural network is constructed to realize online prediction of user movement intention through the dual-channel LSTM neural network;

[0027] An asymmetric impedance function design and joint stiffness damping parameter optimization unit is used to design the asymmetric impedance function. The model predictive control algorithm is used to generate the desired joint trajectory in the upper layer of the gradient sensitivity model. The joint stiffness damping parameters are optimized in real time in the lower layer of the gradient sensitivity model through an improved particle swarm optimization algorithm. A nonlinear disturbance observer is introduced to compensate for the hysteresis effect of the magnetorheological actuator.

[0028] The metabolic consumption assessment model and the minimum energy consumption assist curve solving unit are used to establish the metabolic consumption assessment model, use the rolling time domain optimization algorithm to solve the minimum energy consumption assist curve, and set a safety threshold constraint to limit the joint power overshoot.

[0029] The three-level control cycle setting and hardware acceleration unit is used to set the three-level control cycle, including the first-level actuator current loop control, the second-level impedance parameter adjustment, and the third-level path planning update, and realizes real-time calculation through FPGA and GPU hardware acceleration modules.

[0030] Thirdly, embodiments of the present invention disclose an electronic device, comprising:

[0031] One or more processors;

[0032] Memory, used to store one or more programs;

[0033] When the one or more programs are executed by the one or more processors, the one or more processors implement the control method of the lower limb exoskeleton assistive device.

[0034] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0035] This invention discloses a control method and system for a lower limb exoskeleton assistive device for tower climbing. By integrating surface electromyography signals, plantar pressure data, and lidar point cloud data, and employing an adaptive noise-complete empirical mode decomposition algorithm, an improved Teager energy operator, and a dual-channel LSTM neural network with an attention mechanism, it comprehensively and accurately perceives the user's movement intention and environmental information. This overcomes the limitations of single data processing, making movement intention recognition more accurate, assisting in smoother climbing movements, improving efficiency, and increasing comfort. The dynamic adaptive control algorithm employs an asymmetric impedance function. The upper-level MPC generates the desired joint trajectory and adjusts the strategy in real time based on actual conditions. The lower-level improved particle swarm optimization algorithm optimizes joint stiffness and damping parameters, achieving deep synergy with hardware power output. This allows for flexible adaptation to complex climbing conditions and ensures safety. Hardware acceleration and real-time optimization algorithms utilize FPGA and GPU hardware acceleration modules, setting a three-level control cycle and using the RHO algorithm to solve for the minimum energy consumption assist curve, improving system real-time performance and energy management efficiency, and extending battery life to meet the needs of long-term operation.

[0036] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0037] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0038] Figure 1 This is a flowchart of a control method for a lower limb exoskeleton assistive device for climbing iron towers, as described in Embodiment 1 of the present invention.

[0039] Figure 2 This is a structural diagram of the control system of a lower limb exoskeleton assistive device for climbing iron towers, as shown in Embodiment 3 of the present invention.

[0040] Figure 3 This is a schematic diagram of the structure of an electronic device in Embodiment 4 of the present invention. Detailed Implementation

[0041] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0042] To address the problems existing in the prior art, embodiments of the present invention provide a control method and system for a lower limb exoskeleton assistive device for climbing iron towers.

[0043] Example 1

[0044] This invention discloses a control method for a lower limb exoskeleton assistive device for climbing iron towers, such as... Figure 1 ,include:

[0045] S100. Real-time acquisition of user's electromyography signals and plantar pressure distribution data via surface electromyography sensor and plantar pressure sensor;

[0046] Surface electromyography (SEMG) sensors and plantar pressure sensors provide crucial data support for motion analysis, rehabilitation assessment, and human-computer interaction by acquiring biomechanical signals in real time. SEMG sensors record potential changes during muscle contraction using electrodes attached to the skin, reflecting neuromuscular activity. Plantar pressure sensors measure pressure values ​​in different areas of the foot using piezoelectric films or capacitive sensor arrays, generating dynamic pressure distribution maps.

[0047] S200. Decompose the electromyographic signal and calculate the muscle activation index; construct a dual-channel LSTM neural network and use the dual-channel LSTM neural network to realize online prediction of the user's movement intention.

[0048] In this embodiment, the electromyography (EMG) signal is decomposed to calculate the muscle activation index. Specifically, the method includes: collecting EMG signals of the main muscle groups of the lower limbs during climbing using a surface EMG sensor; decomposing the EMG signal x(t) to obtain the intrinsic mode function (IMF). i (t), i = 1, 2, 3, ..., n, select the first 6 eigenmode functions; adopt the improved Teager energy operator TKO(x[n]) = x 2[n]-x[n-1]x[n+1], calculate the energy of each intrinsic mode, and sum them by weight to obtain the muscle activation index, where x[n] is the electromyographic signal sampling sequence. It is understood that this invention collects the user's electromyographic signals and plantar pressure distribution data in real time, and uses a dual-channel LSTM neural network to predict the user's movement intention. Specifically, in the process of climbing the iron tower, the user's climbing operation mainly includes the climbing start-up stage, the climbing process stage, and the climbing rest and adjustment stage.

[0049] In this embodiment, multi-sensor data is dynamically weighted through an attention mechanism. The specific method includes: using an attention mechanism, wherein the output feature of the first channel is F1, the output feature of the second channel is F2, and the weights α1 and α2 are calculated by the attention model, α1+α2=1, to obtain the fused feature F=α1F1+α2F2.

[0050] S300. Collect tower point cloud data, identify the edge of the tower crossarm structure based on the tower point cloud data, generate a safe climbing path through cubic spline interpolation, calculate the minimum distance threshold between the user's feet and the workspace and obstacles, and establish a gradient sensitivity model to assist in motion safety decision-making.

[0051] In S300 of this embodiment, a lidar mounted on the exoskeleton is used to collect point cloud data of the iron tower. By improving the RANSAC algorithm, a straight line model is assumed by randomly sampling in the point cloud data, and the distance from other points to the model is calculated to determine the interior points. The edge of the iron tower crossarm structure is found iteratively. With the starting point and the target climbing point as the endpoints, a safe climbing path is generated by cubic spline interpolation. Based on the exoskeleton structure and kinematic parameters, the minimum distance threshold between the reachable workspace and obstacles at the feet is calculated, and a gradient sensitivity model is established to assist in safety decision-making.

[0052] Specifically, the improved RANSAC algorithm is used for line model detection. Clustering algorithms (such as DBSCAN or K-means) are used to cluster point cloud data to reduce the impact of noise and outliers. Sample points are randomly selected within each cluster for model estimation to increase the probability of finding inliers. The least squares method is used to fit a line to the selected sample points. The distance from other points to the fitted line is calculated, and inliers are determined based on a set threshold. Iterative optimization involves repeating the above process, selecting the model with the most inliers as the optimal model. During iteration, the distance threshold is dynamically adjusted based on the current proportion of inliers to improve the algorithm's robustness and efficiency.

[0053] In this embodiment, cubic spline interpolation is used to generate a safe climbing path. The specific method includes: determining the endpoints, using the starting point and the target climbing point as the endpoints of the cubic spline interpolation; performing interpolation calculations, performing cubic spline interpolation between the endpoints to generate a smooth climbing path; and optimizing the path, optimizing the path based on the kinematic parameters of the exoskeleton to ensure its accessibility and safety.

[0054] By improving the RANSAC algorithm, straight edges of tower crossarm structures can be detected more efficiently and robustly. Combining safe climbing paths generated by cubic spline interpolation with the calculation of the minimum distance threshold between the exoskeleton's foot-reachable workspace and obstacles ensures safe movement of the exoskeleton in complex environments. This method not only improves the efficiency and safety of path planning but also provides reliable technical support for intelligent navigation of exoskeletons.

[0055] S400. Design an asymmetric impedance function, use a model predictive control algorithm to generate the desired joint trajectory in the upper layer of the gradient sensitivity model, optimize the joint stiffness and damping parameters in real time in the lower layer of the gradient sensitivity model by an improved particle swarm optimization algorithm, and introduce a nonlinear disturbance observer to compensate for the hysteresis effect of the magnetorheological actuator.

[0056] In S400 of this embodiment, the expression for the asymmetric impedance function is: K d B d For asymmetric stiffness and damping coefficient, K d B d Adjustments are made according to different movement directions and climbing stages. The model predictive control algorithm is based on the exoskeleton dynamics model and the current state to predict the system response and solve the optimization problem to generate the desired joint trajectory. The improved particle swarm optimization algorithm uses the joint stiffness damping parameters as the particle position to iteratively optimize, and introduces a nonlinear disturbance observer to compensate for the hysteresis effect of the magnetorheological actuator.

[0057] In this embodiment, the model predictive control generates the desired joint trajectory, and the specific method includes:

[0058] Construct a state-space model; establish kinematic and dynamic models of the joints and represent them in state-space form. Use linear or nonlinear models, selecting the appropriate model form based on the system characteristics.

[0059] The model is used for prediction and control; at each time step, the system state for a future period is predicted. The optimal control input is solved using an optimization algorithm (such as quadratic programming) to minimize the error between the predicted state and the desired state.

[0060] The desired trajectory is generated using the MPC algorithm, ensuring that the trajectory meets dynamic constraints and performance requirements.

[0061] In this embodiment, the improved particle swarm optimization algorithm optimizes joint stiffness and damping parameters. The specific method includes:

[0062] The particle swarm optimization (PSO) algorithm is improved by adjusting the adaptive inertia weights. Adaptive inertia weights are introduced and dynamically adjusted based on the convergence of the particle swarm, thereby enhancing the algorithm's global and local search capabilities. The particle update strategy is also improved by introducing crossover and mutation operations from chaotic mapping or genetic algorithms to enhance the algorithm's exploration capabilities.

[0063] The joint stiffness and damping parameters are optimized in real time. An improved particle swarm optimization algorithm is used to optimize the joint stiffness and damping parameters in each control cycle. Optimization objectives include minimizing tracking error, energy consumption, or other performance metrics.

[0064] By designing an asymmetric impedance function, combining it with MPC to generate the desired joint trajectory, utilizing an improved PSO algorithm to optimize joint parameters in real time, and introducing NDO to compensate for hysteresis effects, the dynamic performance and robustness of the system can be significantly improved. This method not only optimizes the control strategy but also enhances the system's adaptability and stability in complex environments.

[0065] S500. Establish a metabolic consumption assessment model, use a rolling time-domain optimization algorithm to solve the minimum energy consumption assist curve, and set a safety threshold constraint to limit joint power overshoot;

[0066] In S500 of this embodiment, the metabolic consumption assessment model expression is P. m =f(v,a,F) m A), where v is the user's velocity, a is the acceleration, and F is the acceleration. m F to assist exoskeleton m A represents the muscle activation index, and P represents the muscle activation index. m To address energy consumption, a rolling time-domain optimization algorithm is employed to predict the future system state starting from the current moment, solve for the exoskeleton assistance curve that minimizes metabolic consumption, and simultaneously set a safety threshold to constrain joint power.

[0067] By establishing a metabolic consumption assessment model, combining it with a rolling time-domain optimization algorithm to solve for the minimum energy consumption assist curve, and setting a safety threshold constraint to limit joint power overshoot, efficient and safe motion control can be achieved. This method not only optimizes energy consumption but also improves the stability and reliability of the system, making it suitable for various application scenarios such as exoskeleton robots.

[0068] The S600 is configured with a three-level control cycle, including the first level of actuator current loop control, the second level of impedance parameter adjustment, and the third level of path planning update, and real-time calculation is achieved through FPGA and GPU hardware acceleration modules.

[0069] In S600 of this embodiment, the first-stage actuator current loop control is a 1kHz actuator current loop control, which precisely regulates the exoskeleton output force by controlling the magnetorheological actuator current at a 1kHz frequency; the second-stage impedance parameter adjustment is a 100Hz impedance parameter adjustment, which adjusts the joint stiffness and damping parameters according to the motion state and intention; the third-stage path planning update is a 10Hz path planning update, which updates the safe climbing path according to LiDAR data and user status; the sensor data acquisition logic operation is implemented using FPGA, and large-scale matrix operations are calculated using LSTM neural network processing using GPU.

[0070] By setting a three-level control cycle and utilizing FPGA and GPU hardware acceleration modules for real-time computation, the system's response speed and control accuracy can be significantly improved. This design not only optimizes the performance of the control system but also enhances its reliability and stability, making it suitable for applications requiring high precision and real-time response.

[0071] This embodiment discloses a control method for a lower limb exoskeleton assistive device for tower climbing. By integrating surface electromyography signals, plantar pressure data, and lidar point cloud data, it employs an adaptive noise-complete ensemble empirical mode decomposition algorithm, an improved Teager energy operator, and a dual-channel LSTM neural network with an attention mechanism to comprehensively and accurately perceive the user's movement intention and environmental information. This overcomes the limitations of single data processing, making movement intention recognition more accurate, assisting in smoother climbing movements, improving efficiency, and increasing comfort. The dynamic adaptive control algorithm employs an asymmetric impedance function. The upper-level MPC generates the desired joint trajectory and adjusts the strategy in real time based on actual conditions. The lower-level improved particle swarm optimization algorithm optimizes joint stiffness and damping parameters, achieving deep collaboration with hardware power output. This allows for flexible adaptation to complex climbing conditions and ensures safety. Hardware acceleration and real-time optimization algorithms utilize FPGA and GPU hardware acceleration modules, setting a three-level control cycle and using the RHO algorithm to solve for the minimum energy consumption assist curve. This improves system real-time performance and energy management efficiency, extending battery life to meet the needs of long-term operation.

[0072] Example 2

[0073] To better understand the control method of the lower limb exoskeleton assistive device for climbing iron towers disclosed in Embodiment 1, this embodiment applies the method disclosed in Embodiment 1 to a specific iron tower maintenance and repair scenario.

[0074] First, let's introduce the application scenario and user overview. Suppose a power company needs to perform maintenance and repair work on an old iron tower. The tower is 50 meters high and has a complex structure. A worker is responsible for this climbing operation, wearing the lower limb exoskeleton assistive device of this invention.

[0075] Next, the donning and initial setup of the exoskeleton device for the operator will be introduced. The donning process is as follows: The operator stands next to a specially designed donning aid and first aligns the leg portion of the exoskeleton device with their lower limbs. By adjusting the straps, ensure that the hip, knee, and ankle joints fit snugly and comfortably. The straps are elastic and adjustable to accommodate different body types. Next, the lumbar support component is secured, ensuring a stable fit and providing overall support. The entire donning process takes approximately 3 minutes.

[0076] Initial parameter settings: Turn on the device power and perform initial settings via the handheld control terminal. Based on the operator's body parameters (assuming a height of 175cm and a weight of 75kg), set the basic dimensions and adaptation parameters of the exoskeleton. Simultaneously, calibrate the surface electromyography (EMG) sensor and plantar pressure sensor to ensure accurate data acquisition. For example, the EMG sensor needs to be tightly attached to the surface of key muscle groups such as the quadriceps and hamstrings, and signal transmission is enhanced by applying a special conductive gel.

[0077] Secondly, the climbing operation process of the workers is described using the method disclosed in Example 1; the entire climbing operation process is divided into the starting stage, the climbing process stage, and the rest and adjustment stage; specifically:

[0078] Initial Stage: The worker approaches the tower and grips the climbing ladder. At this moment, surface electromyography (EMG) sensors collect initial contraction signals from the quadriceps and calf muscles. These signals are decomposed using the CEEMDAN algorithm to extract key IMF features, and combined with TKO (Total Kinematic Response) to calculate muscle activation indices. Simultaneously, plantar pressure sensors detect the pressure distribution on both feet. Channel 1 of the dual-channel LSTM neural network processes the EMG features to predict the worker's intention to climb, while channel 2 analyzes the plantar pressure center trajectory to determine the initial stage. An attention mechanism integrates this information, and the device determines that the worker is preparing to begin climbing. The upper-layer MPC algorithm generates the initial desired joint trajectory based on a preset safe climbing strategy, while the lower-layer improved particle swarm optimization algorithm rapidly optimizes joint stiffness and damping parameters, allowing the exoskeleton to gently assist the worker in lifting their legs and taking the first step.

[0079] During the climb: When the worker reaches approximately 10 meters, they encounter an area with a large gap between the crossarms. LiDAR scans the surrounding environment in real time, and point cloud data is used to quickly identify the edges of the tower's crossarm structure using an improved RANSAC algorithm. Based on this, the device uses cubic spline interpolation to generate a safe climbing path. This safe climbing path consists of the foot pressure points of each step the worker takes while climbing the tower; several consecutive foot pressure points form the safe climbing path. During this process, the worker's body posture changes, and surface electromyography signals and plantar pressure data also change accordingly. A multimodal data fusion algorithm continuously and accurately senses these changes, and a dynamic adaptive control algorithm adjusts joint stiffness and damping parameters in a timely manner. For example, when the worker needs to turn sideways to bypass an obstacle, the hip and knee joints of the exoskeleton dynamically adjust their stiffness and damping according to algorithmic instructions, providing appropriate assistance and support to ensure smooth and safe movement. Meanwhile, the hardware acceleration module ensures rapid algorithm computation, the 1kHz actuator current loop control enables the magnetorheological actuator to respond quickly, the 100Hz impedance parameter adjustment ensures comfortable joint movement, and the 10Hz path planning update fine-tunes the climbing path according to real-time environmental changes.

[0080] Rest and Adjustment Phase: Upon reaching 25 meters, the workers take a short rest. At this point, the metabolic consumption assessment model estimates the current energy consumption based on the worker's motion status (speed, acceleration, etc.), the exoskeleton's assist output, and muscle activation indicators. The rolling time-domain optimization algorithm adjusts the subsequent assist curve based on this assessment result to achieve minimal energy consumption assist. Simultaneously, the battery management system monitors battery power in real time and allocates power rationally to ensure sufficient power for subsequent climbs. For example, if a battery is detected to be depleting rapidly, the system automatically adjusts the current allocation, prioritizing the use of batteries with sufficient power to extend the overall range.

[0081] After the workers climb the tower, the climbing operation is complete, and they begin the descent. Once the workers have finished their maintenance work at the top of the tower, they begin their descent. The device determines the descent intention based on surface electromyography signals and plantar pressure data, and dynamically adjusts the control strategy. For example, the joint stiffness and damping parameters are adjusted to a mode more suitable for descent, providing stable support and cushioning. During the descent, lidar continuously monitors the surrounding environment to ensure safety.

[0082] Upon returning to the ground, the workers unfastened the exoskeleton straps, completing the unloading process. The device was then placed in a dedicated maintenance area for inspection by technicians. The inspection included checking for damaged sensors, sufficient battery power, and wear on the joint components. For example, the signal acquisition accuracy of the surface electromyography (EMG) sensors was checked using specialized testing equipment; if a decrease in accuracy was detected, the sensors were replaced or calibrated. The battery was charged, and maintenance was performed according to usage to ensure the device would function properly for the next use.

[0083] Example 3

[0084] Based on the same inventive concept, this disclosure also provides a control system for a lower limb exoskeleton assistive device for climbing iron towers, such as... Figure 2 It includes: an electromyography signal and plantar pressure distribution data acquisition unit, an online prediction unit for movement intention, a unit for generating the edge of the tower crossarm structure and a safe climbing path, a unit for designing asymmetric impedance functions and optimizing joint stiffness and damping parameters, a unit for evaluating metabolic consumption and solving for the minimum energy consumption assist curve, and a unit for setting a three-level control cycle and hardware acceleration; among which:

[0085] The electromyography (EMG) signal and plantar pressure distribution data acquisition unit is used to acquire the user's EMG signal and plantar pressure distribution data in real time through surface EMG sensors and plantar pressure sensors.

[0086] An online prediction unit for movement intention is used to decompose electromyographic signals and calculate muscle activation indices; a dual-channel LSTM neural network is constructed to realize online prediction of user movement intention;

[0087] The tower crossarm structure edge and safe climbing path generation unit is used to decompose electromyographic signals and calculate muscle activation index; a dual-channel LSTM neural network is constructed to realize online prediction of user movement intention through the dual-channel LSTM neural network;

[0088] An asymmetric impedance function design and joint stiffness damping parameter optimization unit is used to design the asymmetric impedance function. The model predictive control algorithm is used to generate the desired joint trajectory in the upper layer of the gradient sensitivity model. The joint stiffness damping parameters are optimized in real time in the lower layer of the gradient sensitivity model through an improved particle swarm optimization algorithm. A nonlinear disturbance observer is introduced to compensate for the hysteresis effect of the magnetorheological actuator.

[0089] The metabolic consumption assessment model and the minimum energy consumption assist curve solving unit are used to establish the metabolic consumption assessment model, use the rolling time domain optimization algorithm to solve the minimum energy consumption assist curve, and set a safety threshold constraint to limit the joint power overshoot.

[0090] The three-level control cycle setting and hardware acceleration unit is used to set the three-level control cycle, including the first-level actuator current loop control, the second-level impedance parameter adjustment, and the third-level path planning update, and realizes real-time calculation through FPGA and GPU hardware acceleration modules.

[0091] Example 4

[0092] Based on the same inventive concept, this disclosure also provides an electronic device. Figure 3This is a schematic diagram of the structure of an electronic device according to an embodiment of this disclosure. Figure 3 As shown, this disclosure provides an electronic device including: one or more processors 101, a memory 102, and one or more I / O interfaces 103. The memory 102 stores one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement any of the optimization methods described in the above embodiments; the one or more I / O interfaces 103 are connected between the processor and the memory, configured to enable information interaction between the processor and the memory.

[0093] The processor 101 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 102 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory (FLASH); the I / O interface (read / write interface) 103 is connected between the processor 101 and the memory 102, and can realize information interaction between the processor 101 and the memory 102, including but not limited to a data bus (Bus).

[0094] In some embodiments, the processor 101, memory 102, and I / O interface 103 are interconnected via bus 104, and thus connected to other components of the computing device.

[0095] In some embodiments, the one or more processors 101 include a field-programmable gate array.

[0096] According to embodiments of this disclosure, a computer-readable medium is also provided. This computer-readable medium stores a computer program, which, when executed by a processor, implements the steps of any of the optimized methods described in the above embodiments.

[0097] It should be understood that the specific order or hierarchy of steps in the disclosed process is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process may be rearranged without departing from the scope of this disclosure. The appended method claims provide elements of various steps in an exemplary order and are not intended to limit the scope to the specific order or hierarchy described.

[0098] In the detailed description above, various features are combined together in a single embodiment to simplify this disclosure. This approach to disclosure should not be construed as reflecting an intention that embodiments of the claimed subject matter require more features than are explicitly stated in each claim. Rather, as reflected in the appended claims, the invention is presented with fewer features than all of the features in a single disclosed embodiment. Therefore, the appended claims are hereby explicitly incorporated into the detailed description, with each claim representing a separate preferred embodiment of the invention.

[0099] Those skilled in the art will also understand that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments herein can be implemented as electronic hardware, computer software, or a combination thereof. To clearly illustrate the interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps described above are generally described in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art can implement the described functionality in alternative ways for each specific application; however, such implementation decisions should not be construed as departing from the scope of this disclosure.

[0100] The steps of the methods or algorithms described in conjunction with the embodiments herein can be directly embodied in hardware, software modules executed by a processor, or a combination thereof. The software modules can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium well known in the art. An exemplary storage medium is connected to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. The ASIC can reside in a user terminal. Alternatively, the processor and storage medium can exist as discrete components in the user terminal.

[0101] For software implementation, the techniques described in this application can be implemented using modules (e.g., procedures, functions, etc.) that perform the functions described in this application. This software code can be stored in memory units and executed by a processor. The memory units can be implemented within the processor or outside the processor; in the latter case, they are communicatively coupled to the processor via various means, as is well known in the art.

[0102] The foregoing description includes examples of one or more embodiments. It is certainly impossible to describe all possible combinations of components or methods in order to describe the above embodiments, but those skilled in the art will recognize that further combinations and arrangements of the various embodiments are possible. Therefore, the embodiments described herein are intended to cover all such changes, modifications, and variations that fall within the scope of the appended claims. Furthermore, the term "comprising" as used in the specification or claims is interpreted in a manner similar to the term "including," as interpreted when used as a conjunction in the claims. Additionally, the use of any term "or" in the specification of the claims is intended to mean "non-exclusive or."

Claims

1. A control method for a lower limb exoskeleton assistive device for climbing iron towers, characterized in that, include: The user's electromyography (EMG) signals and plantar pressure distribution data are collected in real time using surface electromyography (EMG) sensors and plantar pressure sensors. The electromyographic signal is decomposed to calculate the muscle activation index; a dual-channel LSTM neural network is constructed to realize online prediction of the user's movement intention. Collect tower point cloud data, identify the edge of the tower crossarm structure based on the tower point cloud data, generate a safe climbing path through cubic spline interpolation, calculate the minimum distance threshold between the user's feet and the workspace and obstacles, and establish a gradient sensitivity model to assist in motion safety decision-making. An asymmetric impedance function is designed. A model predictive control algorithm is used to generate the desired joint trajectory at the upper layer of the gradient sensitivity model. At the lower layer of the gradient sensitivity model, an improved particle swarm optimization algorithm is used to optimize the joint stiffness and damping parameters in real time. A nonlinear perturbation observer is introduced to compensate for the hysteresis effect of the magnetorheological actuator. The expression for the asymmetric impedance function is: + , , For asymmetric stiffness and damping coefficient, , Adjusting according to different movement directions and climbing stages, the model predictive control algorithm is based on the exoskeleton dynamics model and the current state predicts the system response to solve the optimization problem and generate the desired joint trajectory; the improved particle swarm algorithm uses the joint stiffness damping parameters as the particle position for iterative optimization, and introduces a nonlinear disturbance observer to compensate for the hysteresis effect of the magnetorheological actuator. A metabolic consumption assessment model is established, and a rolling time-domain optimization algorithm is used to solve for the minimum energy consumption assist curve. A safety threshold constraint is set to limit joint power overshoot. The expression of the metabolic consumption assessment model is as follows: , For user movement speed , For acceleration , Assisting exoskeletons A represents the muscle activation index. To minimize energy consumption, a rolling time-domain optimization algorithm is used to predict the future system state starting from the current moment, solve for the exoskeleton assistance curve that minimizes metabolic consumption, and set a safety threshold to constrain joint power. The system is configured with a three-stage control cycle, including actuator current loop control in the first stage, impedance parameter adjustment in the second stage, and path planning update in the third stage. Real-time calculations are achieved through FPGA and GPU hardware acceleration modules.

2. The control method for the lower limb exoskeleton assistive device as described in claim 1, characterized in that, The electromyography (EMG) signals are decomposed to calculate muscle activation indices. Specific methods include: acquiring EMG signals from the main lower limb muscle groups during climbing using a surface EMG sensor; and analyzing the EMG signals... Decomposition yields the intrinsic mode functions. , The first six intrinsic mode functions are selected; an improved Teager energy operator is used. The intrinsic modal energies are calculated, and the weighted summation yields the muscle activation index, where... This is a sequence of electromyographic (EMG) signals.

3. The control method for the lower limb exoskeleton assistive device as described in claim 1, characterized in that, A dual-channel LSTM neural network is constructed to predict the user's movement intention online. The specific method includes: constructing a dual-channel LSTM neural network, where the first channel processes muscle activation indicators to predict the direction of the user's lower limb joint movement, and the second channel analyzes plantar pressure distribution data to determine the climbing stage. The multi-sensor data is dynamically weighted through an attention mechanism to achieve online prediction of the user's movement intention.

4. The control method for the lower limb exoskeleton assistive device as described in claim 3, characterized in that, Dynamically weighting multi-sensor data through an attention mechanism, specifically including: employing an attention mechanism, where the first channel outputs features as follows: The output characteristics of the second channel are: The weights are calculated using the attention model. and , , obtain fusion features .

5. The control method for the lower limb exoskeleton assistive device as described in claim 1, characterized in that, The tower point cloud data is collected by using a lidar installed on the exoskeleton. By improving the RANSAC algorithm, a straight line model is assumed by randomly sampling in the point cloud data. The distance from other points to the model is calculated to determine the interior points. The edge of the tower crossarm structure is found iteratively. Using the starting point and the target climbing point as endpoints, a safe climbing path is generated by cubic spline interpolation. Based on the exoskeleton structure and kinematic parameters, the minimum distance threshold between the reachable workspace and obstacles at the foot is calculated, and a gradient sensitivity model is established to assist in safety decision-making.

6. The control method for the lower limb exoskeleton assistive device as described in claim 1, characterized in that, The first-stage actuator current loop control is a 1kHz actuator current loop control, which precisely regulates the exoskeleton output force by controlling the magnetorheological actuator current at a 1kHz frequency; the second-stage impedance parameter adjustment is a 100Hz impedance parameter adjustment, which adjusts the joint stiffness and damping parameters based on the motion state and intention; the third-stage path planning update is a 10Hz path planning update, which updates the safe climbing path based on LiDAR data and user status; the sensor data acquisition logic is implemented using an FPGA, and large-scale matrix operations are calculated using a GPU to process the LSTM neural network.

7. A control system for a lower limb exoskeleton assistive device for tower climbing, employing the control method of any one of the lower limb exoskeleton assistive devices according to claims 1-6, characterized in that, include: The system includes: an electromyography (EMG) signal and plantar pressure distribution data acquisition unit; an online prediction unit for movement intention; a unit for generating the edge and safe climbing path of the tower crossarm structure; a unit for designing asymmetric impedance functions and optimizing joint stiffness and damping parameters; a unit for evaluating metabolic consumption and solving for the minimum energy consumption assist curve; and a unit for setting the three-level control cycle and hardware acceleration. Among these components: The electromyography (EMG) signal and plantar pressure distribution data acquisition unit is used to acquire the user's EMG signal and plantar pressure distribution data in real time through surface EMG sensors and plantar pressure sensors. An online prediction unit for movement intention is used to decompose electromyographic signals and calculate muscle activation indices; a dual-channel LSTM neural network is constructed to realize online prediction of user movement intention; The tower crossarm structure edge and safe climbing path generation unit is used to collect tower point cloud data, identify the tower crossarm structure edge based on the tower point cloud data, generate a safe climbing path through cubic spline interpolation, calculate the minimum distance threshold between the user's feet and the workspace and obstacles, and establish a gradient sensitivity model to assist in motion safety decision-making. An asymmetric impedance function design and joint stiffness damping parameter optimization unit is used to design the asymmetric impedance function. The model predictive control algorithm is used to generate the desired joint trajectory in the upper layer of the gradient sensitivity model. The joint stiffness damping parameters are optimized in real time in the lower layer of the gradient sensitivity model through an improved particle swarm optimization algorithm. A nonlinear disturbance observer is introduced to compensate for the hysteresis effect of the magnetorheological actuator. The metabolic consumption assessment model and the minimum energy consumption assist curve solving unit are used to establish the metabolic consumption assessment model, use the rolling time domain optimization algorithm to solve the minimum energy consumption assist curve, and set a safety threshold constraint to limit the joint power overshoot. The three-level control cycle setting and hardware acceleration unit is used to set the three-level control cycle, including the first-level actuator current loop control, the second-level impedance parameter adjustment, and the third-level path planning update, and realizes real-time calculation through FPGA and GPU hardware acceleration modules.

8. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the control method of any of the lower limb exoskeleton assistive devices according to claims 1-6.

Images