Hip exoskeleton control system and method integrating terrain perception and variable impedance drive

Abstract

This invention provides a hip exoskeleton control system and method integrating terrain perception and variable impedance drive. The system includes a hybrid drive unit and a multi-source sensing module. The hybrid drive unit includes a servo motor, a reducer, a series elastic actuator, and a magnetorheological damping module. The output of the servo motor is connected to the input of the reducer, the output of the reducer is connected to the input of the series elastic actuator, and the output of the series elastic actuator is connected to one end of the thigh link. The stator of the magnetorheological damping module is fixed to the lumbar base, and the rotor of the magnetorheological damping module is coaxially connected to the output of the series elastic actuator. The multi-source sensing module includes a flexible pressure sensor array, an inertial measurement unit, and a joint angle encoder. The flexible pressure sensor array is laid in the forefoot and heel areas of the sole of the foot. The inertial measurement unit is fixed to the outside of the thigh link, and the joint angle encoder is integrated into the servo motor.

Description

Technical Field

[0001] This invention relates to the field of exoskeleton robots and human-computer interaction control technology, specifically to a hip joint exoskeleton control system and method based on multi-sensor fusion of proprioception and variable impedance drive. Background Technology

[0002] Lower limb exoskeleton robots, as wearable devices that can enhance human function or assist in rehabilitation, have broad application prospects in fields such as medical rehabilitation, material handling, and individual combat. The hip joint, as the core hub connecting the trunk and lower limbs, undertakes the crucial tasks of driving thigh swing and maintaining upper body balance during walking.

[0003] Existing hip exoskeleton drive technologies are mainly divided into two categories: rigid drive and elastic drive. Rigid drive systems have a fast response but lack compliance. Impact loads from the ground are directly transmitted to the human skeleton, and can easily cause "human-machine conflict" or even injury when human-machine movements are not coordinated. To solve this problem, series elastic actuators (SEAs) have been introduced into exoskeleton design, which achieve force compliance and energy storage by connecting elastic elements in series between the motor and the load.

[0004] Regarding hybrid drive technology, existing technologies, such as Chinese Patent Application No. CN201811199841.4, disclose a hip joint rehabilitation exoskeleton based on a multi-functional actuator. This exoskeleton integrates a servo motor, a planetary reducer, and a magnetorheological damper to achieve drive, braking, and hybrid braking functions, providing precise rehabilitation assist torque. However, such existing technologies primarily focus on torque assistance during rehabilitation training and lack the ability to adaptively adjust to complex outdoor terrain (such as slope variations).

[0005] On the one hand, traditional SEA or simple hybrid actuators have relatively fixed physical parameters and cannot dynamically adjust joint impedance according to terrain changes (such as high stiffness support for uphill and high damping buffer for downhill); and when subjected to high-frequency impact or rapid movement, the elastic element is prone to underdamped oscillation, affecting system stability.

[0006] On the other hand, in terms of environmental perception, existing adaptive exoskeletons mostly rely on visual sensors (such as depth cameras and LiDAR) to identify terrain slope and obstacles. However, visual solutions have drawbacks such as high computational power consumption, susceptibility to interference from light and occlusion, high risk of privacy leaks, and high hardware costs, making it difficult to widely adopt them in everyday life scenarios.

[0007] Therefore, how to accurately estimate the terrain using only the robot's own proprioceptive information without relying on external visual sensors, and dynamically adjust the stiffness and damping of the joints accordingly to achieve coordinated execution of driving torque and physical damping, is a problem that current exoskeleton technology urgently needs to solve in its transition from "rehabilitation training" to "all-terrain assistance". Summary of the Invention

[0008] To address the shortcomings of existing technologies, the present invention aims to provide a hip exoskeleton control system and method that integrates terrain perception and variable impedance actuation.

[0009] According to one aspect of the present invention, a hip exoskeleton control system integrating terrain perception and variable impedance drive includes a hybrid drive unit and a multi-source sensing module. The hybrid drive unit includes a servo motor, a reducer, a series elastic actuator, and a magnetorheological damping module. The output end of the servo motor is connected to the input end of the reducer, the output end of the reducer is connected to the input end of the series elastic actuator, the output end of the series elastic actuator is connected to one end of the thigh link, the stator of the magnetorheological damping module is fixed to the waist base, and the rotor of the magnetorheological damping module is coaxially connected to the output end of the series elastic actuator. The multi-source sensing module includes a flexible pressure sensor array, an inertial measurement unit, and a joint angle encoder. The flexible pressure sensor array is laid on the forefoot and heel areas of the sole of the foot, the inertial measurement unit is fixed to the outside of the thigh linkage, and the joint angle encoder is integrated on the servo motor.

[0010] According to another aspect of the present invention, a method for controlling a hip exoskeleton that integrates terrain perception and variable impedance actuation includes: Step S1: Obtain the pitch angle of the thigh, the relative rotation angle and angular velocity of the hip joint, and the plantar pressure signal; Step S2: Based on the pitch angle of the thigh, the relative rotation angle and angular velocity of the hip joint, and the plantar pressure signal, obtain the current ground slope and gait phase signal; Step S3: Based on the current ground slope angle, determine the current terrain conditions, and combine the gait phase signal to calculate the target stiffness coefficient and target auxiliary torque corresponding to both hip joints; Step S4: Combining the calculation results and the coordinated adjustment strategy, generate control commands to control the hybrid drive unit to perform adaptive impedance switching based on the terrain slope. In this process, the servo motor outputs low-frequency auxiliary torque through a series elastic driver, and the magnetorheological damping module changes the joint damping characteristics by adjusting the excitation current to match the target stiffness.

[0011] Preferably, step S2 includes: Step S2.1: The plantar pressure signal is used as an external disturbance term and introduced into a preset nonlinear oscillation model. The wearer's walking frequency is locked through the parameter convergence process of the model, and a continuous phase signal representing the percentage progress of gait is output. Step S2.2: During the support phase, based on the thigh pitch angle and the relative rotation angle of the hip joint, calculate the tilt angle of the lower leg relative to the gravity vector; combine the foot contact state determined based on the plantar pressure signal, use the Kalman filter algorithm to filter out motion noise, and estimate the current ground slope value in real time.

[0012] Preferably, step S2.1 further includes: when the flexible pressure sensing array detects the rising edge signal of the heel contact pressure, forcibly resetting the phase variable.

[0013] Preferably, the coordinated regulation strategy in step S4 specifically includes: When the slope angle When in the support phase, it is determined to be an uphill working condition. The magnetorheological damping module is controlled to increase the excitation current, and the servo motor is controlled to output a positive auxiliary torque. When the slope angle When the condition is determined to be downhill, the magnetorheological damping module is controlled to continuously output high damping torque, and the servo motor is controlled to reduce or stop actively outputting auxiliary torque. When in the oscillating phase, the excitation current of the magnetorheological damping module is controlled to be zero or maintained at the lowest threshold.

[0014] Among them, based on the following mapping relationship between stiffness coefficient and excitation current, according to the target stiffness coefficient Calculate the excitation current required for the magnetorheological damping module:

[0015] in, This represents the target excitation current of the i-th side magnetorheological damping module in the k-th control cycle. This indicates that the result within the parentheses is restricted to a range. Saturation function within, Indicates the maximum permissible excitation current. , , This represents the equivalent stiffness-current mapping coefficient obtained through offline calibration.

[0016] Preferably, the coordinated regulation strategy in step S4 further includes: The auxiliary torque signal is decomposed into low-frequency and high-frequency components; Among them, the low-frequency component is sent to the servo motor as an auxiliary power command, which drives the elastic element of the series elastic actuator to deform and provide biomechanical assistance. The high-frequency component is sent to the magnetorheological damping module as an oscillation suppression command. When the elastic element of the series elastic actuator is detected to resonate or is subjected to high-frequency impact from the road surface, the damping of the magnetorheological damping module is dynamically adjusted.

[0017] Preferably, it also includes a safety protection step: when the plantar pressure signal disappears instantly and the joint angular velocity exceeds a preset safety threshold, it is determined that there is a risk of stepping into a hole or falling, and the bilateral MR modules are controlled to enter the magnetic saturation state to lock the hip joint.

[0018] Preferably, in step S3, determining the current terrain conditions includes:

[0019] in, This indicates the current terrain and working conditions. The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; Indicates uphill working conditions. Indicates flat ground conditions. This indicates a downhill working condition.

[0020] Preferably, in step S3, calculating the target stiffness coefficient includes: The continuous gait phases corresponding to the left and right hip joints were normalized separately. The specific normalization process is as follows:

[0021] In the formula, For the first Lateral continuous gait phase recording, Indicates left or right side. ; Based on plantar pressure signals, the plantar contact status is obtained. Combined with the plantar contact status, the weights of the hip joint support phases on both sides are calculated. The expression for the hip joint support phase weights is as follows:

[0022] in, This is a foot contact point. Indicates the first The side of the foot is in a supporting position with its foot touching the ground. Indicates the first The side of the foot is in a swinging, off-the-ground position; This is the off-ground phase threshold; Calculate the target stiffness coefficient:

[0023] in, This represents the target stiffness coefficient of the i-th hip joint in the k-th control cycle. This represents a saturation function, used to restrict the result within the parentheses to a specific interval. Inside, This indicates the lower limit of the target stiffness. Indicates the upper limit of the target stiffness; Indicates the reference stiffness of the oscillating phase; This indicates the additional stiffness of the supporting phase; This represents the stiffness gain coefficient for uphill conditions. This represents the stiffness gain coefficient under downhill conditions. Indicates the weight of the i-th hip joint support phase; The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; , which represents the positive value truncation operator.

[0024] Preferably, in step S3, calculating the target auxiliary torque includes:

[0025] in:

[0026]

[0027] In the formula, This represents the target auxiliary torque of the i-th hip joint in the k-th control cycle; { The parentheses `}` represent a saturation function, used to restrict the result within the parentheses to a specific interval. Inside; Indicates the lower limit of the target auxiliary torque; Indicates the upper limit of the target auxiliary torque; This represents the reference auxiliary torque under flat ground conditions. This represents the torque gain coefficient for uphill operation. This indicates the torque adjustment coefficient for downhill conditions.

[0028] Compared with the prior art, the present invention has the following beneficial effects: This invention abandons the reliance on visual sensors and innovatively uses the fusion of IMU and joint kinematic data to achieve terrain slope estimation, which effectively reduces system cost and computing power requirements, and is not affected by environmental factors such as light and occlusion, making it highly adaptable to the environment. This invention adopts a hybrid drive architecture integrating SEA and MR modules. It utilizes the millisecond-level response characteristics of the MR module to achieve dynamic adjustment of joint stiffness, solving the pain point that traditional SEA cannot cope with the requirements of downhill braking and uphill strong support. This invention proposes a frequency domain division of labor control strategy, which uses the MR module to suppress the high-frequency resonance of the SEA spring, significantly improving the stability and safety of human-computer interaction. This invention combines the AO algorithm with IMU data to achieve precise synchronization of bilateral independent phases, which can adapt to the special needs of hemiplegic patients with asymmetrical gait. Attached Figure Description

[0029] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 A schematic diagram of the overall wearable structure of the hip joint exoskeleton control system provided by the present invention; Figure 2 A schematic diagram of the internal mechanical structure of the hybrid drive unit provided by this invention; Figure 3 The electrical schematic diagram of the control system hardware provided for this invention; Figure 4 A schematic diagram illustrating the principle of the ontology-aware terrain estimation and variable impedance control strategy provided by this invention; In the figure, 1. Waist base; 2. Hybrid drive unit; 3. Inertial measurement unit; 4. Thigh linkage; 5. Flexible pressure sensor array. Detailed Implementation

[0030] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0031] Example 1: This embodiment provides a hip exoskeleton control system that integrates terrain perception and variable impedance drive, including a hybrid drive unit 2 and a multi-source sensing module. The hybrid drive unit 2 includes a servo motor, a reducer, a series elastic actuator, and a magnetorheological damping module. The output end of the servo motor is connected to the input end of the reducer, the output end of the reducer is connected to the input end of the series elastic actuator, and the output end of the series elastic actuator is connected to one end of the thigh link 4. The stator of the magnetorheological damping module is fixed to the waist base 1, and the rotor of the magnetorheological damping module is coaxially connected to the output end of the series elastic actuator. The multi-source sensing module includes a flexible pressure sensor array 5, an inertial measurement unit 3, and a joint angle encoder. The flexible pressure sensor array is laid on the forefoot and heel areas of the sole of the foot, the inertial measurement unit 3 is fixed to the outside of the thigh link, and the joint angle encoder is integrated on the servo motor.

[0032] Example 2: A method for controlling a hip exoskeleton that integrates terrain perception and variable impedance actuation, applicable to the system of Embodiment 1, includes: Step S1: Obtain the pitch angle of the thigh, the relative rotation angle and angular velocity of the hip joint, and the plantar pressure signal; Step S2: Based on the pitch angle of the thigh, the relative rotation angle and angular velocity of the hip joint, and the plantar pressure signal, obtain the current ground slope and gait phase signal.

[0033] In this embodiment, step S2 includes: Step S2.1: The plantar pressure signal is used as an external disturbance term and introduced into a preset nonlinear oscillation model. The wearer's walking frequency is locked through the parameter convergence process of the model, and a continuous phase signal representing the percentage progress of gait is output.

[0034] Step S2.2: During the support phase, based on the thigh pitch angle and the relative rotation angle of the hip joint, calculate the tilt angle of the lower leg relative to the gravity vector; combine the foot contact state determined based on the plantar pressure signal, use the Kalman filter algorithm to filter out motion noise, and estimate the current ground slope value in real time.

[0035] Based on the above scheme, in order to address the inherent defects of adaptive oscillators (AO) and Kalman filters in wearable exoskeletons or bionic prostheses, a deeply decoupled and dynamically compensated motion intention recognition and terrain estimation system was constructed.

[0036] Regarding gait phase tracking, to address the frequency convergence lag issue of the AO algorithm during sudden gait frequency changes (such as sudden acceleration or deceleration by the user), the system does not directly adopt its purely algorithm-driven closed-loop feedback. Instead, it introduces a hardware-level phase reset mechanism based on multi-source ontological perception. By capturing key physical signals of heel contact in real time through plantar pressure sensors, when a new gait cycle is detected, the system forcibly resets the phase deviation of the AO model. This transforms the software-level algorithmic oscillation convergence into physically triggered instantaneous synchronization, fundamentally eliminating the "human-machine confrontation" and fall risk caused by phase misalignment.

[0037] In the environmental terrain perception dimension, to address the distortion issue of Kalman filtering when handling severe nonlinear acceleration and gait impact noise during the swaying phase, the system adopts a phase-separated observation strategy based on plantar pressure. This strategy restricts the terrain estimation logic to the "support phase" window when the foot is fully in contact with the ground, using the relatively stable kinematic data during this period as the observation input for the Kalman filter, while automatically blocking state updates during the noisy swaying phase. This time-domain filtering mechanism effectively shields dynamic interference, ensuring high accuracy and robustness of slope estimation, avoiding abnormal joint force or protective locking triggered by controller misjudgment of terrain, and guaranteeing the wearer's walking safety and dynamic adaptability in complex road conditions.

[0038] In this embodiment, step S2.1 further includes: when the flexible pressure sensing array detects the rising edge signal of the heel contact pressure, forcibly resetting the phase variable.

[0039] Step S3: Based on the current ground slope angle, determine the current terrain conditions, and combine the gait phase signal to calculate the target stiffness coefficient and target auxiliary torque corresponding to both hip joints; Step S4: Combining the calculation results and the coordinated adjustment strategy, generate control commands to control the hybrid drive unit to perform adaptive impedance switching based on the terrain slope. In this process, the servo motor outputs low-frequency auxiliary torque through a series elastic driver, and the magnetorheological damping module changes the joint damping characteristics by adjusting the excitation current to match the target stiffness.

[0040] In this embodiment, the coordinated adjustment strategy in step S4 specifically includes: When the slope angle When in the support phase, it is determined to be an uphill working condition. The magnetorheological damping module is controlled to increase the excitation current, and the servo motor is controlled to output a positive auxiliary torque. When the slope angle When the condition is determined to be downhill, the magnetorheological damping module is controlled to continuously output high damping torque, and the servo motor is controlled to reduce or stop actively outputting auxiliary torque. When in the oscillating phase, the excitation current of the magnetorheological damping module is controlled to be zero or maintained at the lowest threshold.

[0041] Among them, based on the following mapping relationship between stiffness coefficient and excitation current, according to the target stiffness coefficient Calculate the excitation current required for the magnetorheological damping module:

[0042] in, This represents the target excitation current of the i-th side magnetorheological damping module in the k-th control cycle. This indicates that the result within the parentheses is restricted to a range. Saturation function within, Indicates the maximum permissible excitation current. , , This represents the equivalent stiffness-current mapping coefficient obtained through offline calibration.

[0043] In this embodiment, the coordinated adjustment strategy in step S4 further includes: The auxiliary torque signal is decomposed into low-frequency and high-frequency components; Among them, the low-frequency component is sent to the servo motor as an auxiliary power command, which drives the elastic element of the series elastic actuator to deform and provide biomechanical assistance. The high-frequency component is sent to the magnetorheological damping module as an oscillation suppression command. When the elastic element of the series elastic actuator is detected to resonate or is subjected to high-frequency impact from the road surface, the damping of the magnetorheological damping module is dynamically adjusted.

[0044] In this embodiment, a safety protection step is also included: when the plantar pressure signal disappears instantly and the joint angular velocity exceeds a preset safety threshold, it is determined that there is a risk of stepping into a hole or falling, and the bilateral MR modules are controlled to enter the magnetic saturation state to lock the hip joint.

[0045] In this embodiment, step S3, determining the current terrain conditions, includes:

[0046] in, This indicates the current terrain and working conditions. The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; Indicates uphill working conditions. Indicates flat ground conditions. This indicates a downhill working condition.

[0047] In this embodiment, step S3, calculating the target stiffness coefficient, includes: The continuous gait phases corresponding to the left and right hip joints were normalized separately. The specific normalization process is as follows:

[0048] In the formula, For the first Lateral continuous gait phase recording, Indicates left or right side. ; Based on plantar pressure signals, the plantar contact status is obtained. Combined with the plantar contact status, the weights of the hip joint support phases on both sides are calculated. The expression for the hip joint support phase weights is as follows:

[0049] in, This is a foot contact point. Indicates the first The side of the foot is in a supporting position with its foot touching the ground. Indicates the first The side of the foot is in a swinging, off-the-ground position; This is the off-ground phase threshold; Calculate the target stiffness coefficient:

[0050] in, This represents the target stiffness coefficient of the i-th hip joint in the k-th control cycle. This represents a saturation function, used to restrict the result within the parentheses to a specific interval. Inside, This indicates the lower limit of the target stiffness. Indicates the upper limit of the target stiffness; Indicates the reference stiffness of the oscillating phase; This indicates the additional stiffness of the supporting phase; This represents the stiffness gain coefficient for uphill conditions. This represents the stiffness gain coefficient under downhill conditions. Indicates the weight of the i-th hip joint support phase; The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; , which represents the positive value truncation operator.

[0051] In this embodiment, step S3, calculating the target auxiliary torque, includes:

[0052] in:

[0053]

[0054] In the formula, This represents the target auxiliary torque of the i-th hip joint in the k-th control cycle; { The parentheses `}` represent a saturation function, used to restrict the result within the parentheses to a specific interval. Inside; Indicates the lower limit of the target auxiliary torque; Indicates the upper limit of the target auxiliary torque; This represents the reference auxiliary torque under flat ground conditions. This represents the torque gain coefficient for uphill operation. This indicates the torque adjustment coefficient for downhill conditions.

[0055] Based on the above scheme, the calculation of the target stiffness coefficient and the target auxiliary moment shows that, within the uphill support phase, the target stiffness coefficient... and target auxiliary torque The target stiffness coefficient increases synchronously; in flat terrain, both remain at a baseline level; in downhill terrain, the target stiffness coefficient maintains the equivalent joint impedance required for support, while the target auxiliary torque decreases to zero or a slightly negative value; in the oscillating phase, due to... The target auxiliary torque automatically and smoothly decreases to zero, and the target stiffness coefficient falls back to a lower swing reference stiffness. This enables transparent following during the swing period.

[0056] Example 3: like Figure 1 and Figure 2 As shown, this embodiment provides a hip exoskeleton control system that integrates proprioception and variable impedance actuation. In terms of physical structure, it mainly includes: wearable components, hybrid actuation unit, multi-source sensing module and central controller.

[0057] Hybrid Drive Unit: Located at both hip joints. It employs a "rigid-flexible" topology: the output shaft of the servo motor is connected to a reducer, and the reducer output is connected in series with an elastic element (such as a planar spiral spring or torsion spring) to the thigh linkage, forming a series elastic drive (SEA) path for outputting low-frequency, high-torque assistance; the stator of the magnetorheological (MR) damping module is fixed to the lumbar base, and the rotor is coaxially connected to the thigh linkage, forming a parallel damping path. By adjusting the current in the MR module coils, the shear yield stress of the magnetorheological fluid can be changed, thereby generating a variable passive damping torque at the joint.

[0058] Multi-source perception module: Used to construct a non-visual ontology perception environment. Includes: Inertial Measurement Unit (IMU): Rigidly fixed to the outside of the thigh linkage, used to collect the absolute attitude angle (pitch angle) and angular velocity of the thigh in the inertial frame.

[0059] Flexible foot sensor array: laid on the forefoot and heel areas of the foot to detect the pressure distribution on the foot and generate ground contact / removal event signals.

[0060] Joint angle encoder: Integrated at the servo motor end and the joint output end respectively, used to measure relative rotation angle and deformation of SEA elastic element.

[0061] like Figure 3 and Figure 4As shown, the core control process of this system includes four stages: information acquisition, ontology perception and calculation, strategy decision-making and collaborative execution, with a control cycle of milliseconds.

[0062] The following are instructions on how to use this system: Step S1: Multi-source information collection.

[0063] After the system is powered on, the central controller reads the raw data streams from each sensor in real time via the bus. Specifically, this includes: acquiring plantar pressure signals from the flexible plantar sensor array, acquiring absolute thigh tilt data from the IMU, and acquiring joint relative angle and angular velocity data from the joint angle encoder.

[0064] Step S2: Blind source terrain and phase calculation (processing layer).

[0065] The controller internally runs two parallel observer algorithms to solve for gait and environmental state without relying on external vision sensors: Gait phase synchronization: The adaptive oscillator (AO) algorithm is used to introduce the acquired plantar pressure signal as an external perturbation term into the nonlinear oscillation model. This model can automatically converge and lock the wearer's walking frequency, outputting a phase signal that varies continuously between 0 and 2π, accurately representing the percentage progress of whether the wearer is in the support phase or the swing phase.

[0066] Propriometry-based terrain estimation: Data fusion is performed using state observers (such as Kalman filters). By establishing a geometric model of the lower limb kinematic chain, the absolute inclination angle of the thigh acquired by the IMU and the relative angle of the hip joint acquired by the encoder are differentially and fused to filter out dynamic swaying noise during human walking and estimate the slope of the ground in real time.

[0067] It should be noted that both the Adaptive Oscillator (AO) algorithm and the Kalman filter have a serious drawback: the AO algorithm suffers from an unavoidable frequency convergence time lag when the wearer's gait frequency changes abruptly, while the traditional Kalman filter is highly susceptible to high-frequency impact noise from human walking and severe nonlinear acceleration of the sway phase when estimating terrain throughout the entire time period. Directly using these two algorithms can lead to a misalignment between the gait phase and the actual state, causing strong "human-machine conflict" and resulting in wearer falls, as well as frequent misjudgments of terrain by the controller (such as misjudging flat ground as a steep slope), leading to abnormal joint force or even lock-up, among other serious consequences. To address this, this invention introduces a hardware-level phase reset mechanism based on multi-source ontology perception, using the heel contact pressure signal to forcibly reset the AO model phase, achieving zero-delay synchronization. Furthermore, it proposes a phase-separated observation strategy based on plantar pressure, using the Kalman filter to estimate the slope only during the "support phase" window when the foot is stably in contact with the ground, completely shielding the dynamic interference of the sway phase, thus enabling its application in this field.

[0068] Step S3: Variable impedance strategy decision (decision level).

[0069] Based on the calculated terrain slope and gait phase, and combined with a frequency domain division of labor strategy, the controller generates low-level control commands: Operating condition determination and impedance mapping: Uphill condition (when the estimated slope is positive and exceeds the threshold): The system determines that the user needs assistance and support. The controller sets a high target stiffness command and generates a positive feedforward torque command.

[0070] Downhill condition (when the estimated gradient is negative and exceeds the threshold): It is determined that the user needs braking and energy absorption. The controller sets a high target damping command and sets the motor's active output torque to zero or slightly negative in order to dissipate gravitational potential energy using passive damping.

[0071] Flat terrain (when the estimated slope is within the threshold range): The system determines that the user needs low-resistance walking. The controller is set to zero-damping and low-stiffness commands to achieve transparent following.

[0072] Frequency domain splitting processing: The controller monitors the joint angular velocity in real time and decomposes the total torque demand into low-frequency and high-frequency components: Low-frequency components (<3Hz): represent the intention of normal walking motion and are assigned to the servo motor for execution.

[0073] High-frequency components (>5Hz): represent road impacts or resonance of elastic elements, assigned to the MR module for execution.

[0074] Based on the above scheme, step S3 includes: Step S3.1: Terrain Condition Determination: Record the current ground slope angle obtained in real-time estimation in step S2 as... Where k is the current control cycle; set the uphill judgment threshold. Downhill detection threshold The current terrain condition is indicated. satisfy:

[0075] in, Indicates uphill working conditions. Indicates flat ground conditions. This indicates a downhill working condition.

[0076] Step S3.2: Phase normalization and support phase weight construction.

[0077] Calculate separately for the left and right hip joints, let Indicates left or right. The output of step S2 is the first... The lateral continuous gait phase is denoted as ,in Normalize it to:

[0078] Plantar contact markers obtained by combining plantar pressure signals Construct the first Weight of the support phase of the lateral hip joint :

[0079] in, Indicates the first The side of the foot is in a supporting position with its foot touching the ground. Indicates the first The side of the foot is in a swinging, off-the-ground position; This is the off-ground phase threshold.

[0080] Step S3.3: Calculation of target stiffness coefficient.

[0081] No. Target stiffness coefficient of the lateral hip joint Calculate using the following formula:

[0082] This represents the target stiffness coefficient of the i-th hip joint in the k-th control cycle. This represents a saturation function, used to restrict the result within the parentheses to a specific interval. Inside, This indicates the lower limit of the target stiffness. Indicates the upper limit of the target stiffness; Indicates the reference stiffness of the oscillating phase; This indicates the additional stiffness of the supporting phase; This represents the stiffness gain coefficient for uphill conditions. This represents the stiffness gain coefficient under downhill conditions. Indicates the weight of the i-th hip joint support phase; The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; , which represents the positive value truncation operator.

[0083] Step S3.4: Calculation of target auxiliary torque.

[0084] No. Target auxiliary torque of the lateral hip joint Calculate using the following formula:

[0085] in:

[0086]

[0087] In the formula, This represents the target auxiliary torque of the i-th hip joint in the k-th control cycle; { The parentheses `}` represent a saturation function, used to restrict the result within the parentheses to a specific interval. Inside; Indicates the lower limit of the target auxiliary torque; Indicates the upper limit of the target auxiliary torque; This represents the reference auxiliary torque under flat ground conditions. This represents the torque gain coefficient for uphill operation. This indicates the torque adjustment coefficient for downhill conditions.

[0088] From steps S3.3 and S3.4, it can be seen that within the uphill support phase, the target stiffness coefficient... and target auxiliary torque The target stiffness coefficient increases synchronously; in flat terrain, both remain at a baseline level; in downhill terrain, the target stiffness coefficient maintains the equivalent joint impedance required for support, while the target auxiliary torque decreases to zero or a slightly negative value; in the oscillating phase, due to... The target auxiliary torque automatically and smoothly decreases to zero, and the target stiffness coefficient falls back to a lower swing reference stiffness. This enables transparent following during the swing period.

[0089] Step S4: Collaborative Driven Execution (Execution Layer).

[0090] The underlying driver receives the instructions generated in step S3 and drives the hardware to perform actions: Servo motor operation: Responding to low-frequency torque and stiffness commands, it outputs stable biomechanical assistance (such as uphill thrust) to the human body by compressing the SEA elastic element.

[0091] MR module actions: responding to damping commands and high-frequency vibration suppression commands.

[0092] In downhill mode, the current controller injects a continuous high current into the coil, causing the magnetorheological fluid to "harden" instantly, providing a flexible braking torque. When vibration is detected, a pulsed current is superimposed to suppress the aftershocks of the SEA spring; During the swing phase, the current is cut off, allowing the joint to enter a zero-resistance state and reducing the metabolic consumption of the human body.

[0093] Fault protection mechanism.

[0094] The system is equipped with safety redundancy logic. When the sensing module data is abnormal (such as IMU data packet loss or the plantar pressure being zero for a long time), the controller immediately triggers the fail-safe mode: it forces the MR module to output a constant medium damping current, locks the position of the servo motor or puts it in zero torque mode, and uses the inherent viscosity of the magnetorheological fluid to prevent the wearer from falling due to joint instability.

[0095] Example 4: This embodiment presents a hip exoskeleton control method that integrates proprioception and variable impedance actuation, comprising the following steps: Step S1: Data Acquisition: Real-time acquisition of plantar pressure data from the flexible sensor array on both sides of the wearer's feet, attitude angle data from the inertial measurement unit (IMU), and angle and angular velocity data from the joint angle encoder; Step S2: Volumetric terrain and phase calculation: The plantar pressure data is processed using an adaptive oscillator (AO) to converge and output independent gait phase signals of both lower limbs in real time. Based on the IMU attitude angle data and joint kinematic data, a state observer is constructed to estimate the ground slope angle in real time without the need for external visual information. Step S3: Target Impedance and Torque Generation: Based on the ground slope angle, determine the current terrain conditions, and combine the gait phase signal to calculate the target stiffness coefficient and target auxiliary torque of the bilateral joints through a preset mapping model; Step S4: Collaborative Drive Execution: The controller collaboratively adjusts the hybrid drive unit according to the calculation results. The servo motor outputs low-frequency auxiliary torque through the series elastic actuator (SEA), and the magnetorheological (MR) damping module changes the joint damping characteristics by adjusting the excitation current to match the target stiffness.

[0096] In this embodiment, the ground slope angle estimation in step S2 specifically includes: During the support phase, the absolute tilt angle of the thigh acquired by the IMU and the relative angle of the hip joint acquired by the encoder are kinematically fused to calculate the tilt angle of the lower leg relative to the gravity vector. By combining the foot contact state determined by the flexible foot sensor array, the Kalman filter algorithm is used to filter out motion noise and estimate the current ground slope value in real time.

[0097] The coordinated adjustment strategy in step S4 specifically includes adaptive impedance switching based on terrain slope: When the slope angle When in the support phase, it is determined to be an uphill working condition. The MR module is controlled to increase the excitation current to improve the joint support stiffness and prevent the center of gravity from tilting backward. At the same time, the servo motor outputs positive assist torque. When the slope angle When the condition is determined to be downhill, the MR module is controlled to continuously output high damping torque for flexible braking and energy dissipation, and the servo motor reduces or stops active power output. When in the swing phase, regardless of the slope, the excitation current of the MR module is controlled to be zero or maintained at the lowest threshold to achieve zero-impedance transparent following of the joint.

[0098] Among them, based on the following mapping relationship between stiffness coefficient and excitation current, according to the target stiffness coefficient Calculate the excitation current required for the magnetorheological damping module:

[0099] in, This represents the target excitation current of the i-th side magnetorheological damping module in the k-th control cycle. This indicates that the result within the parentheses is restricted to a range. Saturation function within, Indicates the maximum permissible excitation current. , , This represents the equivalent stiffness-current mapping coefficient obtained through offline calibration.

[0100] In this embodiment, step S4 further includes a frequency domain division of labor control strategy: The controller decomposes the target torque signal into low-frequency and high-frequency components; The low-frequency component is sent to the servo motor as an auxiliary assist command, driving the elastic element of the SEA to deform in order to provide biomechanical assistance; High-frequency components are sent to the MR module as oscillation suppression commands. When the SEA elastic element is detected to resonate or be subjected to high-frequency impact from the road surface, the damping of the MR module is dynamically adjusted to dissipate vibration energy and improve the stability of human-machine interaction.

[0101] In this embodiment, the adaptive oscillator (AO) model includes a phase reset mechanism: when the flexible foot sensor array detects the rising edge signal of the heel contact pressure, the phase variable of the AO is forcibly reset to eliminate frequency convergence lag during variable speed walking.

[0102] In this embodiment, the method also includes safety protection logic: when the system detects that the plantar pressure signal disappears instantly and the joint angular velocity exceeds the preset safety threshold, it determines that there is a risk of stepping into a hole or falling, and immediately controls the bilateral MR modules to enter the magnetic saturation state to lock the hip joint to provide rigid support.

[0103] The present invention also provides a hip exoskeleton control system that integrates proprioception and variable impedance actuation. The hip exoskeleton control system that integrates proprioception and variable impedance actuation can be implemented by executing the process steps of the hip exoskeleton control method that integrates proprioception and variable impedance actuation. That is, those skilled in the art can understand the hip exoskeleton control method that integrates proprioception and variable impedance actuation as a preferred embodiment of the hip exoskeleton control system that integrates proprioception and variable impedance actuation.

[0104] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0105] In the description of this application, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0106] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A hip exoskeleton control system integrating terrain perception and variable impedance actuation, characterized in that, The system includes a hybrid drive unit (2) and a multi-source sensing module. The hybrid drive unit (2) includes a servo motor, a reducer, a series elastic actuator, and a magnetorheological damping module. The output end of the servo motor is connected to the input end of the reducer, the output end of the reducer is connected to the input end of the series elastic actuator, the output end of the series elastic actuator is connected to one end of the thigh connecting rod, the stator of the magnetorheological damping module is fixed to the waist base, and the rotor of the magnetorheological damping module is coaxially connected to the output end of the series elastic actuator. The multi-source sensing module includes a flexible pressure sensor array (5), an inertial measurement unit (3), and a joint angle encoder. The flexible pressure sensor array is laid on the forefoot and heel areas of the sole of the foot. The inertial measurement unit (3) is fixed to the outside of the thigh link (4). The joint angle encoder is integrated on the servo motor.

2. A method for controlling a hip exoskeleton that integrates terrain perception and variable impedance actuation, characterized in that, Using the system of claim 1, the method includes: Step S1: Obtain the pitch angle of the thigh, the relative rotation angle and angular velocity of the hip joint, and the plantar pressure signal; Step S2: Based on the pitch angle of the thigh, the relative rotation angle and angular velocity of the hip joint, and the plantar pressure signal, obtain the current ground slope and gait phase signal; Step S3: Based on the current ground slope angle, determine the current terrain conditions, and combine the gait phase signal to calculate the target stiffness coefficient and target auxiliary torque corresponding to both hip joints; Step S4: Combining the calculation results and the coordinated adjustment strategy, generate control commands to control the hybrid drive unit to perform adaptive impedance switching based on the terrain slope. In this process, the servo motor outputs low-frequency auxiliary torque through a series elastic driver, and the magnetorheological damping module changes the joint damping characteristics by adjusting the excitation current to match the target stiffness coefficient.

3. The method according to claim 2, characterized in that, Step S2 includes: Step S2.1: The plantar pressure signal is used as an external disturbance term and introduced into a preset nonlinear oscillation model. The wearer's walking frequency is locked through the parameter convergence process of the model, and a continuous phase signal representing the percentage progress of gait is output. Step S2.2: During the support phase, based on the thigh pitch angle and the relative rotation angle of the hip joint, calculate the tilt angle of the lower leg relative to the gravity vector; combine the foot contact state determined based on the plantar pressure signal, use the Kalman filter algorithm to filter out motion noise, and estimate the current ground slope value in real time.

4. The method according to claim 3, characterized in that, Step S2.1 further includes: when the flexible pressure sensing array detects the rising edge signal of the heel contact pressure, the phase variable is forcibly reset.

5. The method according to claim 2, characterized in that, The coordinated regulation strategy in step S4 specifically includes: When the slope angle When in the support phase, it is determined to be an uphill working condition. The magnetorheological damping module is controlled to increase the excitation current, increase the stiffness coefficient, and control the servo motor to output a positive auxiliary torque. When the slope angle When the condition is determined to be a downhill condition, the magnetorheological damping module is controlled to continuously output high damping torque to maintain a high stiffness coefficient, and the servo motor is controlled to reduce or stop actively outputting auxiliary torque. When in the swing phase, the excitation current of the magnetorheological damping module is controlled to be zero or maintained at the lowest threshold. Among them, based on the following mapping relationship between stiffness coefficient and excitation current, according to the target stiffness coefficient Calculate the excitation current required for the magnetorheological damping module: in, This represents the target excitation current of the i-th side magnetorheological damping module in the k-th control cycle. This indicates that the result within the parentheses is restricted to a range. Saturation function within, Indicates the maximum permissible excitation current. , , This represents the equivalent stiffness-current mapping coefficient obtained through offline calibration.

6. The method according to claim 5, characterized in that, The coordinated regulation strategy in step S4 also includes: The auxiliary torque signal is decomposed into low-frequency and high-frequency components; Among them, the low-frequency component is sent to the servo motor as an auxiliary power command, which drives the elastic element of the series elastic actuator to deform and provide biomechanical assistance. The high-frequency component is sent to the magnetorheological damping module as an oscillation suppression command. When the elastic element of the series elastic actuator is detected to resonate or is subjected to high-frequency impact from the road surface, the damping of the magnetorheological damping module is dynamically adjusted.

7. The method according to claim 2, characterized in that, It also includes safety protection steps: when the plantar pressure signal disappears instantly and the joint angular velocity exceeds the preset safety threshold, it is determined that there is a risk of stepping into a hole or falling, and the bilateral MR modules are controlled to enter the magnetic saturation state to lock the hip joint.

8. The method according to claim 2, characterized in that, In step S3, determining the current terrain conditions includes: in, This indicates the current terrain and working conditions. The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; Indicates uphill working conditions. Indicates flat ground conditions. This indicates a downhill working condition.

9. The method according to claim 2, characterized in that, In step S3, calculating the target stiffness coefficient includes: The continuous gait phases corresponding to the left and right hip joints were normalized separately. The specific normalization process is as follows: In the formula, For the first Side continuous gait phase recording, Indicates left or right side. ; Based on plantar pressure signals, the plantar contact status is obtained. Combined with the plantar contact status, the weights of the hip joint support phases on both sides are calculated. The expression for the hip joint support phase weights is as follows: in, The foot touches the ground. Indicates the first The side of the foot is in a supporting position with its foot touching the ground. Indicates the first The side of the foot is in a swinging, off-ground position. This is the off-ground phase threshold; Calculate the target stiffness coefficient: in, This represents the target stiffness coefficient of the i-th hip joint in the k-th control cycle. This represents a saturation function, used to restrict the result within the parentheses to a specific interval. Inside, This indicates the lower limit of the target stiffness. Indicates the upper limit of the target stiffness; Indicates the reference stiffness of the oscillating phase; This indicates the additional stiffness of the supporting phase; This represents the stiffness gain coefficient for uphill conditions. This represents the stiffness gain coefficient under downhill conditions. Indicates the weight of the i-th hip joint support phase; The current ground slope angle is k; the current control cycle is k. The threshold for determining an uphill slope; The threshold for determining a downhill slope; , which represents the positive value truncation operator.

10. The method according to claim 9, characterized in that, In step S3, calculating the target auxiliary torque includes: in: In the formula, This represents the target auxiliary torque of the i-th hip joint in the k-th control cycle; { The parentheses `}` represent a saturation function, used to restrict the result within the parentheses to a specific interval. Inside; Indicates the lower limit of the target auxiliary torque; Indicates the upper limit of the target auxiliary torque; This represents the reference auxiliary torque under flat ground conditions. This represents the torque gain coefficient for uphill operation. This indicates the torque adjustment coefficient for downhill conditions.

Images