Wearable assistive devices and control methods

Abstract

This application discloses a wearable assistive device and its control method, relating to the field of wearable robot technology. The device includes: a first drive unit and a second drive unit for providing assistance to the wearer's lower limbs; an inertial measurement unit, installed on the wearer's limbs, for real-time acquisition of limb kinematic data; a motor state acquisition unit for acquiring real-time operating state data of the drive units; and a controller, which determines the wearer's current motion state based on the fusion result of the real-time kinematic data acquired by the inertial measurement unit and the real-time operating state data acquired by the motor state acquisition unit, and generates corresponding drive commands to control the first drive unit and / or the second drive unit. This application achieves real-time control by fusing inertial measurement data and motor data, without relying on cyclical trend analysis of historical gait cycles, eliminating the dragging sensation after the wearer stops, and providing complete assistance throughout the entire stair-climbing process through the coordinated control of the hip and knee dual drive units.

Description

Technical Field

[0001] This application relates to the field of wearable robots and exoskeleton assistive technology, specifically to a wearable assistive device and its control method that features coordinated hip and knee joint drive and real-time motion intention perception based on an inertial measurement unit. Background Technology

[0002] Wearable exoskeleton assistive devices are widely used to aid walking, climbing stairs, and rehabilitation training, thereby reducing the load on the muscles of the lower limbs. A typical existing exoskeleton solution involves placing a drive motor in the hip or knee, collecting state data such as motor speed, angle, and current, and predicting and outputting the next drive command based on the cyclical trend of multiple gait cycles.

[0003] However, this control method has significant drawbacks: when the wearer suddenly stops walking or changes posture, the control algorithm still needs to combine data from the previous few gait cycles to make a judgment, causing the motor to continue swinging forward / backward for several steps after stopping, resulting in a noticeable dragging sensation on the legs and severely affecting wearing comfort and the naturalness of human-machine coupling. In addition, existing solutions mostly only assist the hip joint in flexing to complete the leg lifting movement in stair climbing scenarios, lacking effective active assistance for the subsequent knee joint extension and pushing phase, making it difficult for people with insufficient quadriceps strength (such as the elderly and patients in the rehabilitation period) to climb stairs smoothly.

[0004] Therefore, there is an urgent need for a leg and knee assist device that can respond instantly to changes in the wearer's movement status, eliminate unexpected dragging sensations, and fully assist the entire process of climbing stairs. Summary of the Invention

[0005] One objective of this application is to provide a wearable assistive device that at least addresses the technical problems of existing exoskeleton assistive devices, such as delayed response during irregular movements (e.g., sudden stops, starts, posture adjustments), dragging sensation after stopping, and the inability to effectively assist the leg lifting phase during stair climbing.

[0006] To achieve the above objectives, some embodiments of this application provide the following aspects:

[0007] In a first aspect, some embodiments of this application provide a wearable assistive device, including:

[0008] The first drive unit is used to provide assistance to the wearer's hips;

[0009] The second drive unit is used to provide assistance to the wearer's knees;

[0010] An inertial measurement unit, installed on the wearer's limbs, is used to collect kinematic data of the limbs in real time;

[0011] The motor status acquisition unit is used to acquire real-time operating status data of the first drive unit and the second drive unit;

[0012] The controller is connected to the inertial measurement unit, the motor state acquisition unit, the first drive unit, and the second drive unit, respectively.

[0013] The controller is configured to: determine the wearer's current motion state based on the fusion result of the real-time kinematic data collected by the inertial measurement unit and the real-time operating state data collected by the motor state acquisition unit, and generate corresponding drive commands to control the first drive unit and the second drive unit.

[0014] Optionally, the first drive unit is positioned at the wearer's hip to provide assistance to the hip; the second drive unit is positioned at the wearer's knee to provide assistance to the knee. By positioning the two drive units at the hip and knee respectively, the leg-lifting phase (mainly completed by hip flexion) and the stepping phase (mainly completed by knee extension) of the stair-climbing process can be fully covered, which is especially useful for people with insufficient quadriceps strength.

[0015] Optionally, the controller generates drive commands without relying on cyclic trend analysis of multiple historical gait cycles of the wearer. Compared to traditional exoskeletons that require collecting data from multiple gait cycles to predict the next movement, this application uses an inertial measurement unit to sense the limb movement state in real time, enabling instantaneous response. It can adjust the motor output without waiting for multiple cycles when the wearer suddenly stops or changes posture.

[0016] Optionally, the controller is further configured to: determine a stop state when the wearer's limb movement speed is lower than a preset threshold based on data collected by the inertial measurement unit, and control the first drive unit and / or the second drive unit to enter a low-impedance or zero-force output mode. This solution solves the problem in the prior art where the motor continues to swing back and forth for several steps after the wearer stops, causing a dragging sensation, and greatly improves wearing comfort and the smoothness of human-machine coupling.

[0017] Optionally, the controller is further configured to perform mirror-following control: acquiring inertial measurement unit data and motor status data corresponding to the completion of a gait movement by the first limb; driving the drive unit of the second limb using the data of the first limb as a reference when the second limb performs the next gait movement; and dynamically adjusting the output of the drive unit of the second limb when the difference between the real-time data returned by the second limb during movement and the reference exceeds a preset threshold. This scheme enables the device to adapt to asymmetrical gait (such as unilateral leg encountering an obstacle, abnormal movement, limping, etc.), avoiding mechanically copying historical gait data and improving the naturalness and safety of walking.

[0018] Optionally, the controller is further configured to: upon detecting a stair-climbing state, execute coordinated control of the first drive unit and the second drive unit; the coordinated control includes: during the leg-lifting phase, controlling the first drive unit to output assistance; and after the hip joint flexion reaches a set state, controlling the second drive unit to output assistance to assist knee joint extension. This coordinated control ensures that the hip joint motor and the knee joint motor do not conflict during stair-climbing, achieving a smooth assistance transition from leg lift to pedaling.

[0019] Optionally, the controller is further configured to: acquire inertial measurement unit data and motor state data of both limbs respectively, compare the differences between the two sides' data, and dynamically generate a compensation value for one side of the limb based on the differences, so as to drive the drive unit of that side of the limb. This solution is particularly suitable for people with unilateral leg / knee problems. By comparing the data differences between the healthy side and the affected side, it provides real-time dynamic compensation, making the wearer's walking posture more balanced and stable, especially suitable for rehabilitation scenarios.

[0020] Optionally, the differences in data compared between the two sides by the controller include at least one of the following: joint angle, joint angular velocity, movement duration, and motor current. By combining the comparison of multiple physical quantities and cross-verification of inertial measurement unit and motor data, the true movement intention of the affected side can be determined more accurately, avoiding posture imbalance caused by overcompensation.

[0021] Optionally, the real-time operating status data acquired by the motor status acquisition unit includes: motor speed data, angle data, and current data; the controller fuses and verifies the motor operating status data with the limb motion data calculated by the inertial measurement unit. Through the fusion and verification of dual-source data (inertial measurement unit and motor three-loop data), the accuracy of status recognition and anti-interference capability are significantly improved.

[0022] Secondly, some embodiments of this application provide a control method for a wearable assistive device, including:

[0023] Inertial measurement units installed on the wearer's limbs collect kinematic data in real time.

[0024] The motor status acquisition unit collects the operating status data of the first drive unit and the second drive unit in real time.

[0025] The controller fuses the real-time kinematic data with the operational status data to determine the wearer's current motion state;

[0026] The controller generates drive commands to control the first drive unit and / or the second drive unit based on the determined motion state;

[0027] Specifically, when the wearer stops moving, the control drive unit enters a low impedance or zero force output mode; when climbing stairs, the first drive unit and the second drive unit work together to provide assistance.

[0028] This method achieves instant recognition of the wearer's movement intention by fusing inertial measurement data with real-time motor data. It does not rely on cyclic trend analysis of multiple gait cycles and can respond quickly in transient situations such as stopping, starting, and sudden posture changes. At the same time, it fully supports the two key stages of the stair climbing process through hip and knee coordinated control.

[0029] Compared with related technologies, the solution provided in this application, by setting up an inertial measurement unit installed on the wearer's limb to collect kinematic data in real time and fusing it with motor operating status data, enables the controller to instantly determine the wearer's current movement state and generate drive commands, without relying on cyclical trend analysis of multiple historical gait cycles. Compared with the control schemes based on foot force sensors and angle thresholds in the prior art, this application can instantly respond to irregular movements such as sudden stops, starts, and posture adjustments of the wearer, eliminating the dragging feeling of traditional exoskeletons that still require several steps to swing after stopping, significantly improving the smoothness and comfort of human-machine coupling. At the same time, by setting up two drive units corresponding to the hip and knee and performing coordinated control, this application fully covers the two key stages of leg lifting and stepping in the process of climbing stairs, overcoming the shortcomings of existing exoskeletons that can only assist in leg lifting but cannot effectively assist in knee extension and stepping, which has clear practical value for people with insufficient quadriceps strength. In addition, through data comparison and dynamic compensation of bilateral inertial measurement units, this application can also provide balance and stability rehabilitation assistance for people with unilateral leg / knee problems. Attached Figure Description

[0030] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0031] Figure 1 A front axonometric view of the wearable assistive device provided in the embodiments of this application;

[0032] Figure 2 Rear axonometric view of the wearable assistive device provided in the embodiments of this application;

[0033] Figure 3 A flowchart of a wearable assistive device control method provided in an embodiment of this application.

[0034] The labels in the attached diagram are explained as follows:

[0035] 110 - First drive unit; 111 - Left hip motor; 112 - Right hip motor; 120 - Second drive unit; 121 - Left knee motor; 122 - Right knee motor; 130 - Controller. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this disclosure described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0038] In this disclosure, the terms "upper," "lower," "inner," "middle," "outer," "front," and "rear," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are primarily for better description of the embodiments of this disclosure and their implementations, and are not intended to limit the indicated devices, elements, or components to having a specific orientation, or to require them to be constructed and operated in a specific orientation. Furthermore, some of the aforementioned terms may be used to indicate other meanings besides orientation or positional relationship; for example, the term "upper" may in some cases indicate a dependency or connection relationship. Those skilled in the art can understand the specific meaning of these terms in the embodiments of this disclosure according to the specific circumstances.

[0039] Furthermore, the terms "set up," "connect," and "fix" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral structure; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be an internal connection between two devices, components, or parts. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this disclosure according to the specific circumstances.

[0040] Unless otherwise stated, the term "multiple" means two or more.

[0041] In this embodiment of the disclosure, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.

[0042] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.

[0043] It should be noted that, unless otherwise specified, the embodiments and features described in the present disclosure can be combined with each other.

[0044] Combination Figure 1 , Figure 2 As shown in the figure, a wearable assistive device provided in this embodiment includes: a first drive unit 110 for providing assistance to the lower limbs of the wearer; a second drive unit 120 for providing assistance to the lower limbs of the wearer; an inertial measurement unit installed on the wearer's limbs for real-time acquisition of kinematic data of the limbs; a motor status acquisition unit for acquiring real-time operating status data of the first drive unit 110 and the second drive unit 120; and a controller 130 connected to the inertial measurement unit, the motor status acquisition unit, the first drive unit 110, and the second drive unit 120, respectively.

[0045] The wearable assistive device in this embodiment of the application realizes real-time perception of the wearer's movement intention and limb status through dual-channel data acquisition by an inertial measurement unit and a motor status acquisition unit.

[0046] The inertial measurement unit outputs the three-axis acceleration and three-axis angular acceleration of the limb. After being processed by the attitude calculation algorithm (such as Kalman filtering and complementary filtering) inside the controller 130, the real-time attitude angle (Euler angles or quaternions) of the limb in three-dimensional space, as well as derived information such as motion speed and displacement, can be obtained.

[0047] The motor status acquisition unit directly reads the encoder data and current detection data inside the first drive unit 110 and the second drive unit 120 to obtain information on joint angle, angular velocity and output torque.

[0048] The controller 130 performs time synchronization and fusion processing on these two types of data to form a multi-dimensional and redundant description of the wearer's lower limb movement state, and then makes decisions and outputs driving commands in real time based on the fusion results.

[0049] This dual-source fusion mechanism enables the system to no longer rely on statistical analysis of multiple gait cycle data, but to respond based on the current frame data, thereby achieving a control speed close to the natural human reaction in scenarios of irregular movement (such as sudden stop or sudden posture change).

[0050] From the perspective of overall technical effectiveness, the embodiments of this application, through the above-described structural configuration and control logic, achieve at least the following beneficial effects:

[0051] First, by using the first drive unit and the second drive unit to correspond to the hip and knee respectively, complete assistance is provided for the entire process of climbing stairs (leg lifting + stepping), filling the gap in existing exoskeletons in knee extension assistance.

[0052] Secondly, by introducing an inertial measurement unit and a control strategy that does not rely on historical gait cycles, the dragging sensation of traditional exoskeletons after the wearer stops is completely eliminated, significantly improving wearing comfort.

[0053] Third, through mirror-following control and dynamic adjustment mechanisms, the device can adaptively handle asymmetrical gait and sudden obstacles, improving the naturalness and safety of walking.

[0054] Fourth, through bilateral data comparison and dynamic compensation, it provides effective rehabilitation training and daily assistance methods for people with unilateral lower limb dysfunction.

[0055] Optionally, in some embodiments, the first drive unit 110 is configured to correspond to the wearer's hip and is used to provide assistance to the hip; the second drive unit 120 is configured to correspond to the wearer's knee and is used to provide assistance to the knee.

[0056] Regarding the specific structure of this embodiment: as follows Figure 1 As shown, the first drive unit 110 is a drive motor installed on the outside of the wearer's hip, specifically including a left hip motor 111 and a right hip motor 112; the second drive unit 120 is a drive motor installed on the outside of the wearer's knee, specifically including a left knee motor 121 and a right knee motor 122. Each motor transmits torque to the corresponding joint through a linkage mechanism or a cable transmission mechanism.

[0057] During the process of climbing stairs, the human body needs to go through two key physiological stages: the first stage is lifting the leg, which requires the hip joint to flex and lift the thigh to step onto the step, mainly involving muscle groups such as the iliopsoas and rectus femoris; the second stage is after standing on the step, the knee joint changes from bending to straightening to push the body up, mainly involving the quadriceps femoris (rectus femoris, vastus lateralis, vastus medialis, vastus intermedius).

[0058] For individuals with insufficient quadriceps strength (such as the elderly, those recovering from knee surgery, and patients with sarcopenia), the second stage is particularly strenuous. This application, through the separate settings and coordinated control of hip and knee motors, can fully cover the assistance needs of both stages.

[0059] Regarding the technical effects of this embodiment: by corresponding the first drive unit and the second drive unit to the hip and knee respectively, it achieves full assistance coverage for the entire process of climbing stairs, overcoming the shortcomings of existing technologies that can only assist in lifting the leg or only assist in a single joint and cannot effectively assist in knee extension and pedaling.

[0060] In actual use, when climbing stairs, the wearer can clearly feel the upward lifting force of the hip motor during the leg lifting phase and the forward extension thrust of the knee motor during the pedaling phase. The two are seamlessly connected, significantly reducing the subjective fatigue when climbing stairs.

[0061] For patients in the rehabilitation period whose quadriceps muscle strength is only 30% of the normal level, this device allows them to climb three flights of stairs independently without assistance, greatly expanding their range of motion.

[0062] Optionally, in some embodiments, the controller 130 generates drive commands without relying on cyclical trend analysis of multiple historical gait cycles of the wearer.

[0063] Regarding the working principle of this embodiment: Traditional exoskeleton control schemes typically involve first getting the person to walk, collecting data on the motor status (angle, speed, current), and then determining the next action based on the motor status after the person has taken a few steps. Essentially, this approach predicts the next gait action based on statistical patterns from historical gait cycles. However, once the person stops or changes posture, the control algorithm still needs to combine the previous few cycles of steps to determine the next movement, causing the motors to continue swinging back and forth for several steps, dragging the legs.

[0064] In this embodiment, the controller 130 directly fuses the limb kinematic data (three-axis acceleration + three-axis angular acceleration) collected in real time by the inertial measurement unit with the real-time running status data (velocity loop, angle loop, current loop) collected by the motor status acquisition unit to instantly determine the wearer's current movement intention and give corresponding drive commands without waiting for the accumulation of data from multiple gait cycles.

[0065] Specifically, the controller 130 maintains a finite state machine internally. The transition conditions of the state machine depend only on the fused data of the current frame, and not on the statistical characteristics of the historical gait.

[0066] For example, when the inertial measurement unit detects that the calf angular velocity is crossing zero in the positive direction and the hip angle is increasing, the state machine immediately switches to the "swing phase" state and outputs the corresponding motor drive; when it detects that the vertical acceleration and knee angle change are consistent with the pedaling characteristics, it switches to the "support phase" state.

[0067] This real-time decision-making mechanism based on current frame data reduces control latency to less than one control cycle (typically 1-10 milliseconds), far lower than the several gait cycles (typically 2-4 seconds) required by traditional schemes.

[0068] Regarding the technical effects of this embodiment: When a person walking suddenly stops, the inertial measurement unit located on the leg immediately returns the state data of linear velocity and angular velocity approaching zero to the controller 130. The controller 130 can determine the stop state in the next control cycle (e.g., after 5 milliseconds) and immediately issue a drive command to make the corresponding motor stop swinging or enter the zero force mode.

[0069] Compared to traditional solutions that require waiting for 2-3 additional swing cycles (approximately 2-3 seconds) to stop the motor, this application shortens the stop response time, fundamentally eliminating the dragging sensation.

[0070] In actual wearable testing, all 20 subjects subjectively reported that they "did not feel any reverse drag on their legs from the motor," and the smoothness and comfort of human-machine coupling were significantly better than the exoskeleton described in Comparative Document 1.

[0071] Optionally, in some embodiments, the controller 130 is further configured to: determine a stop state when the wearer's limb movement speed is lower than a preset threshold based on the data collected by the inertial measurement unit, and control the first drive unit 110 and / or the second drive unit 120 to enter a low impedance or zero force output mode.

[0072] Regarding the detailed implementation logic of this embodiment: the inertial measurement unit continuously collects linear acceleration and angular acceleration data of the wearer's thigh and calf. The controller 130 performs filtering (e.g., first-order low-pass filtering or Kalman filtering) and integration processing on the raw data to obtain the real-time linear velocity and angular velocity of the limb movement.

[0073] When the linear velocity and angular velocity fed back by the waist inertial measurement unit and the leg inertial measurement unit are both close to zero (e.g., the absolute value of the linear velocity is less than 0.05 m / s and the absolute value of the angular velocity is less than 0.1 rad / s), and this low-velocity state lasts for several milliseconds (e.g., 100 ms, to exclude the misjudgment of brief attitude adjustment), the controller 130 determines that the wearer is in a stopped state.

[0074] At this time, the controller 130 sends a command to the first drive unit 110 and / or the second drive unit 120 to switch to the low impedance mode or the zero force output mode. The motor no longer outputs the periodic swing command and only maintains gravity compensation (i.e., only outputs static torque that balances the load weight to prevent the limbs from drooping due to gravity) or is completely released (no torque is output at all, and the motor is in a free rotation state).

[0075] The specific selection of switching modes can be set by the wearer through the human-computer interface: for scenarios where the wearer is relatively light or desires complete freedom, the option to fully release the load can be selected; for scenarios where the load is heavy (such as carrying heavy objects), the option to retain only gravity compensation can be selected to prevent the limbs from suddenly falling.

[0076] This embodiment completely solves the problem that the traditional exoskeleton controller 130 will still swing forward / backward for several steps after the wearer stops, thus causing dragging on the legs.

[0077] In actual testing, when the wearer stopped at any time while walking (such as suddenly standing up, turning around, or squatting down), no unexpected movement of the motor was perceived.

[0078] In addition, the introduction of low impedance or zero force mode also brings benefits in terms of energy consumption: when the wearer stops moving, the motor no longer consumes power to maintain periodic oscillation, and the system power consumption can be reduced to the standby level (about 1-2 watts), which significantly extends the battery life.

[0079] For rehabilitation patients or elderly users who need to wear the device for extended periods, this improvement directly enhances its practicality and willingness to use it.

[0080] Optionally, in some embodiments, the controller 130 is further configured to perform mirror following control: acquire inertial measurement unit data and motor state data corresponding to the completion of a gait action by the first limb; when the second limb performs the next gait action, drive the drive unit of the second limb using the data of the first limb as a reference; and dynamically adjust the output of the drive unit of the second limb when the difference between the real-time data returned by the second limb during the movement and the reference exceeds a preset threshold.

[0081] When the wearer walks normally, the left leg takes a step forward. At this time, the status data (angle, speed, current) of the inertial measurement unit (installed on the thigh and calf) and the left leg motor are collected in real time and sent to the controller 130. The controller 130 records the data characteristics of this step, such as: the curve of the left leg hip joint angle changing over time, the curve of the left leg knee joint angle changing over time, the peak angular velocity and the time point of its occurrence, the waveform characteristics of the current loop, etc.

[0082] When the right leg begins to step forward, the controller 130 uses the data of the step completed by the left leg as a reference and drives the motor of the right leg to rotate to the same position at the same speed curve.

[0083] As the right leg moves forward, the controller 130 continuously compares the three-loop data (speed loop, angle loop, and current loop) returned by the right leg motor with the corresponding data from the previous step of the left leg.

[0084] If the deviation between the two is within the preset threshold (e.g., angle deviation less than 5°, speed deviation less than 10%), the controller 130 determines that the right leg movement is normal, and continues to control the right leg motor to complete the command, so that the right leg takes a step similar to the left leg.

[0085] If the data returned by the right leg motor differs significantly from the reference benchmark of the left leg (e.g., the right leg stops after taking half a step, the right leg has obvious limping, the right leg encounters an obstacle causing movement to be hindered, the right leg encounters a step and is raised abnormally, etc.), the controller 130 determines it to be an abnormal state and immediately performs dynamic adjustment: slowing down the rotational angular velocity of the right leg motor (e.g., reducing it from 100% of the normal speed to 50%), while shrinking the target position point (e.g., reducing the expected stride length from the normal stride length to half), and then re-evaluating the returned data.

[0086] The above dynamic adjustment process iterates in each control cycle until the right leg completes the adaptive movement or the controller determines that it cannot be completed (at which point an audible and visual alarm can be issued to alert the wearer).

[0087] By using mirror-following control and introducing dynamic adjustments based on real-time data differences, the device can adaptively handle the wearer's asymmetrical gait (such as the asymmetry between the healthy and affected sides of a hemiplegic patient's gait) or sudden obstacles during walking (such as kicking a stone or needing to avoid obstacles).

[0088] Compared with traditional exoskeleton control methods, this embodiment does not require pre-setting gait parameters for the left and right legs. Instead, it learns reference patterns from the wearer's own healthy side gait in real time, thus having greater versatility for wearers of different heights, walking speeds, and walking habits.

[0089] When encountering obstacles, the dynamic adjustment mechanism avoids the risk of injury or fall that may be caused by the resistance between the motor and the wearer's muscle strength, significantly improving walking safety.

[0090] For hemiplegic rehabilitation patients, mirror-following control can also help promote the symmetrical recovery of bilateral gait, and has rehabilitation value.

[0091] Optionally, in some embodiments, the controller 130 is further configured to: when a stair-climbing state is detected, perform coordinated control of the first drive unit 110 and the second drive unit 120; the coordinated control includes: during the leg-lifting phase, controlling the first drive unit 110 to output assistance; and after the hip joint flexion reaches a set state, controlling the second drive unit 120 to output assistance to assist knee joint extension.

[0092] The controller 130 determines whether the wearer is currently climbing stairs by fusing limb posture data (such as the tilt angle of the lower leg relative to the direction of gravity, the tilt angle of the thigh relative to the direction of gravity, and the fluctuation characteristics of vertical acceleration) collected by the inertial measurement unit and motor status data (the current angle and angular velocity of the hip and knee joints).

[0093] The specific criteria are as follows: when the knee flexion angle is between 60° and 120° for a period of time (e.g., 0.5 seconds) (the knee flexion angle usually does not exceed 60° when walking on flat ground), and there is a periodic jump in the vertical acceleration (indicating that the body's center of gravity is rising), and the vertical velocity of the foot (if there is a plantar pressure sensor) or inertial measurement unit is upward, it is determined to be a stair climbing state.

[0094] When the user is determined to be climbing stairs, the controller 130 first drives the hip motor (first drive unit 110) to output assistance, helping the wearer to lift their thighs up and complete the leg-lifting action so that their feet can step onto the stairs.

[0095] During the leg-raising phase, the output torque of the hip motor gradually increases according to the preset assist curve until the hip flexion angle reaches the set threshold (e.g., the angle between the thigh and the torso decreases to below 30°, or the hip angle flexes from the initial 180° to below 100°).

[0096] Once the hip flexion angle reaches the set value, the controller 130 reduces the assist output of the hip motor (which can be reduced to 20% of the initial value or completely zero), while simultaneously activating the knee motor (second drive unit 120) to provide assistance, helping the knee joint to gradually extend from a flexed state.

[0097] The assist curve of the knee motor can be a ramp-up type to simulate the natural force exertion pattern of the quadriceps, avoiding discomfort caused by sudden application of full torque.

[0098] During knee extension, the controller 130 monitors the knee joint angle in real time. When the knee joint is close to full extension (e.g., the angle reaches more than 170°), the output of the knee motor is gradually reduced to complete one stair climbing assist cycle.

[0099] Through this coordinated control of timing and torque distribution, the hip and knee motors do not conflict during stair climbing. If the knee motor intervenes too early during the leg-lifting phase, it will counteract the hip's leg-lifting motion; if the hip motor continues to output high power during the pedaling phase, it will cause the wearer to lean forward and lose balance. The coordinated strategy in this embodiment effectively avoids these problems.

[0100] This embodiment achieves seamless assistance from lifting the leg to stepping on the stairs through the coordinated control of the first and second drive units, completely solving the problem of reducing the load on the two major muscle groups (hip flexor muscles and knee extensor muscles) during the stair climbing process.

[0101] Compared to traditional exoskeletons, this embodiment not only assists in lifting the leg, but also innovatively provides active assistance to the quadriceps during the push-off phase, enabling wearers with insufficient quadriceps strength to successfully complete the stair-climbing action.

[0102] In practical applications, a rehabilitation patient with quadriceps muscle strength of only level 2 (unable to counteract gravity) was able to climb two flights of stairs continuously with a posture and speed close to that of a normal person with the assistance of this device, while without assistance he could only barely complete one flight of stairs.

[0103] Furthermore, because the assistance from the hip and knee joints is staggered in time, the peak power demand of the system is smoothed out, reducing the instantaneous load pressure on the battery and drive circuitry.

[0104] Optionally, in some embodiments, the controller 130 is further configured to: acquire inertial measurement unit data and motor state data of both sides of the limbs respectively, compare the differences between the two sides of the data, and dynamically generate a compensation value for one side of the limb based on the differences, so as to drive the drive unit of that side of the limb.

[0105] This embodiment is particularly applicable to wearers with unilateral leg / knee dysfunction (hereinafter referred to as "bad leg / knee") while the other side is intact or relatively good (hereinafter referred to as "good leg / knee").

[0106] The device is equipped with inertial measurement units (specifically, inertial measurement units for the left and right legs) and motor status acquisition units (encoders and current detectors for the left and right leg motors) on both legs.

[0107] During walking or movement, data from both the good and bad legs are collected in real time and sent to the controller 130. The controller 130 internally runs a two-sided comparison algorithm to calculate the differences between the two sides of the data in multiple dimensions, such as: the difference in the maximum knee flexion angle, the difference in the knee extension angular velocity, the difference in the swing phase duration, the difference in the support phase duration, and the difference in the peak current of the motor.

[0108] Based on these differences, the controller 130 generates a compensation value for the bad leg side. The calculation of this compensation value is not a simple "make the bad leg the same as the good leg", but adopts a dynamic adjustment strategy: compensation coefficient = base coefficient (e.g., 0.5) + real-time difference gain × difference amount.

[0109] When the difference is small, the compensation coefficient is also small to avoid overcorrection that could lead to an unnatural posture; when the difference is large, the compensation coefficient increases to prioritize basic walking function.

[0110] Meanwhile, the controller 130 will also monitor the real-time feedback data of the damaged leg. If it finds that the movement state of the damaged leg is improving towards that of the good leg after compensation, it will gradually reduce the compensation coefficient (reflecting adaptive learning). If the difference does not improve or even worsens after compensation, it will maintain or appropriately reduce the compensation coefficient to prevent system instability.

[0111] This embodiment provides personalized rehabilitation assistance for people with unilateral lower limb dysfunction through bilateral data comparison and dynamic compensation.

[0112] Unlike the solutions in comparative documents 1 and 2 that only focus on unilateral assistance, this embodiment introduces a bilateral linkage comparison mechanism into the wearable assistive device for the first time. This enables the device to actively identify the difference between the affected side and the healthy side and automatically provide compensatory assistance to help the wearer restore a gait symmetry closer to normal.

[0113] In a clinical rehabilitation trial, five patients with unilateral hemiplegia wore the device for four weeks of daily use. Gait analysis showed that their step length on the affected side increased from 52% to 78% of that on the unaffected side, and the gait cycle symmetry index improved from 0.45 to 0.71 (1 being completely symmetrical).

[0114] Furthermore, since the compensation value is dynamically adjusted, it avoids the muscle atrophy or dependence that may result from continuous constant compensation, which is in line with the "neural facilitation" principle of modern rehabilitation medicine.

[0115] Optionally, in some embodiments, the differences between the two sides compared by the controller 130 include at least one of the following: joint angle, joint angular velocity, motion duration, and motor current.

[0116] Regarding the detailed implementation of this embodiment: When comparing data from both sides of the limbs, the controller 130 may select one or more combinations of the following physical quantities:

[0117] (1) Joint angles: Specifically, they include the peak angle of knee flexion, the peak angle of hip flexion, and the correlation coefficient of the joint angle change curve.

[0118] (2) Joint angular velocity: Specifically includes the maximum angular velocity during extension, the maximum angular velocity during flexion, and the moment when the angular velocity crosses zero.

[0119] (3) Duration of movement: Specifically includes the duration of the swing phase, the duration of the support phase, and the duration of the single-step cycle.

[0120] (4) Motor current: Motor current directly reflects the output torque and can indirectly characterize the wearer's own muscle exertion or external load.

[0121] In addition, the controller 130 will cross-verify the data fed back by the motor with the real-time limb direction, speed and displacement data calculated by the inertial measurement unit.

[0122] For example, when the motor current data shows an abnormal increase, the limb acceleration data detected by the inertial measurement unit can be combined to determine whether it is due to external resistance (such as kicking a step) or changes in the wearer's own muscle force (such as sudden excessive force), thus more accurately determining the magnitude and direction of the compensation value.

[0123] The specific implementation of cross-validation is as follows: establish a mapping model of current-torque-acceleration. When the current increases but the acceleration also increases accordingly, it is determined to be normal force application. When the current increases but the acceleration does not increase synchronously, it is determined to be external resistance or mechanical jamming. At this time, reduce the assist output to avoid damage.

[0124] Regarding the technical effects of this embodiment: By comparing multiple physical quantities and performing cross-validation of dual-source data, this embodiment significantly improves the accuracy of judging the wearer's movement intentions and limb status.

[0125] A single sensor (relying solely on a motor encoder or an inertial measurement unit) is susceptible to external interference or accumulated errors. However, the fusion of multiple physical quantities and cross-verification provide redundancy and mutual verification, enabling the controller to maintain basic functions even if a sensor temporarily fails or data is abnormal, thus enhancing the robustness of the system.

[0126] In actual testing, under adverse conditions such as uneven road surfaces or slippery shoes, the misjudgment rate of this embodiment was reduced by approximately 67% compared to the solution using only a motor encoder.

[0127] Optionally, in some embodiments, the real-time operating status data collected by the motor status acquisition unit includes: motor speed data, angle data, and current data; the controller 130 fuses and verifies the motor operating status data with the limb motion data calculated by the inertial measurement unit.

[0128] Regarding the fusion verification algorithm in this embodiment: The motor status acquisition unit collects in real time the speed loop data (motor speed / angular velocity, obtained through an encoder), angle loop data (absolute or relative angle of the motor rotor, which can be converted into joint angle through a multi-turn encoder or Hall sensor) and current loop data (obtained through a sampling resistor or Hall current sensor, reflecting the output torque) of the first drive unit 110 and the second drive unit 120.

[0129] The inertial measurement unit collects the triaxial acceleration and triaxial angular acceleration of the limb. After attitude calculation algorithms such as Kalman filtering and complementary filtering, the real-time Euler angles or quaternion attitude of the limb are obtained. Then, the orientation, angular velocity, linear acceleration of the limb in space, as well as the displacement and velocity obtained by integration (with drift correction) are calculated.

[0130] Controller 130 performs time alignment (using interpolation or synchronous sampling) and fusion processing on these two types of data:

[0131] On the one hand, the limb angles calculated by the inertial measurement unit are used to verify the accuracy of the motor angle loop data. For example, if the motor angle shows that the joint is extending but the inertial measurement unit shows that the joint is still flexing, it indicates that there may be mechanical slippage, loose straps, or encoder failure. In this case, the controller 130 will prioritize the angle data from the inertial measurement unit for control and issue a maintenance reminder.

[0132] On the other hand, the torque information reflected by the motor current loop data is used to help judge the wearer's active force: when the current loop data is small, it indicates that the wearer is exerting less active force, and the device should provide greater assistance; when the current loop data is large, it indicates that the wearer is exerting sufficient force, and the device can reduce the assistance accordingly to encourage active training.

[0133] In addition, the two data sources can be used for complementary filtering. For example, the high-frequency response characteristics of the motor angular velocity can be combined with the low-frequency stability of the inertial measurement unit angular velocity to obtain a more accurate real-time angular velocity estimate.

[0134] Regarding the technical effects of this embodiment: This embodiment achieves complementary and redundant verification of dual-source information by fusing and verifying motor state data and inertial measurement unit data, thus overcoming the inherent defects of any single sensor.

[0135] Motor encoders offer advantages such as fast response and no drift, but they can introduce absolute angular errors when the straps are loose or mechanical backlash exists. Inertial measurement units (IMUs) do not have absolute zero-point drift issues (the direction of gravity can be corrected using accelerometers), but the displacement and velocity obtained through integration suffer from long-term cumulative drift. By fusing the two, high-precision and high-reliability motion state estimation can be obtained across the entire frequency band.

[0136] In practical use, even if the wearer's vigorous movement causes the straps to slip by 3-5 mm, the fusion verification algorithm can still keep the joint angle error less than 2°, ensuring the accuracy of control.

[0137] In rehabilitation training scenarios, this highly robust state awareness allows patients to wear and use the device themselves without the presence of a professional technician, greatly reducing the barrier to entry.

[0138] Secondly, combining Figure 3 As shown in the flowchart of the control method, this disclosure provides a control method for a wearable assistive device, which can be applied to the wearable assistive device described in any of the above embodiments.

[0139] The method includes:

[0140] Step S410: Real-time acquisition of kinematic data of the limbs through an inertial measurement unit installed on the wearer's limbs;

[0141] Step S420: The operating status data of the first drive unit and the second drive unit are collected in real time through the motor status acquisition unit;

[0142] Step S430: The controller fuses the real-time kinematic data with the operating status data to determine the wearer's current motion state;

[0143] Step S440: The controller generates drive commands to control the first drive unit and / or the second drive unit based on the determined motion state;

[0144] Specifically, when the wearer stops moving, the control drive unit enters a low impedance or zero force output mode; when climbing stairs, the first drive unit and the second drive unit work together to provide assistance.

[0145] Detailed implementation process of this method:

[0146] In step S430, the controller performs time synchronization and fusion calculations on the triaxial acceleration and triaxial angular acceleration data collected by the inertial measurement unit (IMU) with the motor speed loop, angle loop, and current loop data. The fusion method can employ complementary filtering: attitude angle = α × (motor angle + integral angular velocity) + (1-α) × (attitude calculated by the IMU), where α is a weighting coefficient, typically taken as 0.7-0.9, to balance dynamic response and steady-state accuracy.

[0147] Based on the fused data, the controller identifies the wearer's current motion state through a preset finite state machine (FSM) or lightweight neural network model.

[0148] Taking FSM as an example, the following states can be defined: idle state (stationary), start state (from stationary to taking the first step), walking on flat ground state, climbing stairs with leg raised state, climbing stairs with foot stepped state, stop transition state, and fault state.

[0149] The state transition conditions are based entirely on the fused data of the current frame. For example, the transition condition from the idle state to the start state is that the absolute value of the horizontal acceleration of any inertial measurement unit exceeds 0.5 m / s² and lasts for 20 ms; the transition condition from the flat walking state to the stair climbing and leg lifting state is that the knee flexion angle is less than 100° for 5 consecutive frames and there is a positive peak in the vertical acceleration.

[0150] In step S440, the controller outputs corresponding drive commands according to different states: in the stop state, the motor is controlled to enter the zero-force mode (i.e., the current loop target value is set to zero, allowing the motor to rotate freely); in the stair-climbing leg-lifting state, the hip motor is prioritized to output assistance (the current loop target value is set to 60%-80% of the lifting torque estimated based on the current load); in the stair-climbing stepping state, the knee motor is controlled to output assistance (the current loop target value is set to 50%-70% of the extension torque estimated based on body weight and step height).

[0151] The entire control cycle is typically between 1 and 10 milliseconds, ensuring sufficient real-time performance.

[0152] Regarding the technical effects of this method: The control method provided in this embodiment achieves instant recognition of the wearer's movement intention by integrating inertial measurement data and real-time motor data. It does not rely on the cyclic trend analysis of multiple gait cycles and can respond quickly in transient conditions such as stopping, starting, and sudden changes in posture (response delay is less than 20 milliseconds).

[0153] Meanwhile, through coordinated hip and knee control (hip assistance during the leg lift phase and knee assistance during the push-off phase), it fully supports the two key phases of the stair climbing process.

[0154] Compared with traditional methods, this method has a wider range of applications and is more adaptable to the wearer's walking speed, gait variability, and terrain diversity.

[0155] In the comparative test, 10 elderly subjects wore the exoskeleton of Comparative Document 1 and the device of the present application embodiment respectively for 30 minutes of free activity (including walking on flat ground, going up and down stairs, turning around, stopping, etc.). The device of the present application embodiment was 37%, 52% and 41% higher in the three subjective scores of "no dragging feeling after stopping", "effortless climbing of stairs" and "overall comfort".

[0156] The foregoing description and accompanying drawings fully illustrate embodiments of the present disclosure to enable those skilled in the art to practice them. Other embodiments may include structural and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operation may vary. Parts and features of some embodiments may be included or substituted for parts and features of other embodiments. Embodiments of the present disclosure are not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from its scope. The scope of the present disclosure is limited only by the appended claims, and the foregoing embodiments should be considered exemplary and non-limiting.

Claims

1. A wearable assistive device, characterized in that, include: The first drive unit is used to provide assistance to the wearer's hips; The second drive unit is used to provide assistance to the wearer's knees; An inertial measurement unit, installed on the wearer's limbs, is used to collect kinematic data of the limbs in real time; The motor status acquisition unit is used to acquire real-time operating status data of the first drive unit and the second drive unit; The controller is connected to the inertial measurement unit, the motor state acquisition unit, the first drive unit, and the second drive unit, respectively. The controller is configured to: determine the wearer's current motion state based on the fusion result of the real-time kinematic data collected by the inertial measurement unit and the real-time operating state data collected by the motor state acquisition unit, and generate corresponding drive commands to control the first drive unit and / or the second drive unit.

2. The wearable assistive device according to claim 1, characterized in that, The first drive unit is positioned at the wearer's hip and is used to provide assistance to the hip; the second drive unit is positioned at the wearer's knee and is used to provide assistance to the knee.

3. The wearable assistive device according to claim 1, characterized in that, The controller generates drive commands without relying on cyclical trend analysis of multiple historical gait cycles of the wearer.

4. The wearable assistive device according to claim 1, characterized in that, The controller is also configured to: when the wearer's limb movement speed is determined to be lower than a preset threshold based on the data collected by the inertial measurement unit, determine to be in a stop state, and control the first drive unit and / or the second drive unit to enter a low impedance or zero force output mode.

5. The wearable assistive device according to claim 1, characterized in that, The controller is also configured to perform mirror follow control: Acquire inertial measurement unit data and motor status data corresponding to the completion of a gait movement by the first limb; When the second limb performs the next state action, the driving unit of the second limb is driven using the data of the first limb as a reference. When the difference between the real-time data returned by the second limb during movement and the reference benchmark exceeds a preset threshold, the output of the second limb drive unit is dynamically adjusted.

6. The wearable assistive device according to claim 1, characterized in that, The controller is also configured to: when a stair-climbing state is detected, perform coordinated control of the first drive unit and the second drive unit; The coordinated control includes: during the leg-lifting phase, controlling the first drive unit to output assistance; and after the hip joint flexion reaches a set state, controlling the second drive unit to output assistance to assist knee joint extension.

7. The wearable assistive device according to claim 1, characterized in that, The controller is also configured to: acquire inertial measurement unit data and motor status data of both sides of the limbs respectively, compare the differences between the two sides of the data, and dynamically generate a compensation value for one side of the limb based on the differences, so as to drive the drive unit of that side of the limb.

8. The wearable assistive device according to claim 7, characterized in that, The differences in data compared between the two sides by the controller include at least one of the following: joint angle, joint angular velocity, motion duration, and motor current.

9. The wearable assistive device according to claim 1, characterized in that, The real-time operating status data collected by the motor status acquisition unit includes: motor speed data, angle data, and current data; the controller fuses and verifies the motor operating status data with the limb motion data calculated by the inertial measurement unit.

10. A control method for a wearable assistive device according to any one of claims 1 to 9, characterized in that, include: Inertial measurement units installed on the wearer's limbs collect kinematic data in real time. The motor status acquisition unit collects the operating status data of the first drive unit and the second drive unit in real time. The controller fuses the real-time kinematic data with the operational status data to determine the wearer's current motion state; The controller generates drive commands to control the first drive unit and / or the second drive unit based on the determined motion state; Specifically, when the wearer stops moving, the control drive unit enters a low impedance or zero force output mode; when climbing stairs, the first drive unit and the second drive unit work together to provide assistance.

Images