A wearable exoskeleton device and an exoskeleton control method suitable for rehabilitation training of infants with congenital cerebral palsy

Abstract

The application belongs to the technical field of wearable exoskeleton devices, and provides a wearable exoskeleton device suitable for rehabilitation training of infants with congenital cerebral palsy and an exoskeleton control method. The device comprises a wearable exoskeleton device main body, hip joint drives symmetrically arranged on the left and right sides, knee joint drives, ankle joint drives, a controller and multiple groups of sensors. The controller stores joint target trajectory data corresponding to multiple training modes, retrieves corresponding trajectory data based on the selected training mode and generates control instructions to actively drive the affected lower limbs to move. The multiple groups of sensors comprise electromyography sensors, inertial measurement units and pressure sensors. The controller identifies abnormal movement modes such as combined responses and extensor spasticity based on sensor signals and dynamically reduces the drive control gain to release abnormal stress in a compliant manner. The application realizes multi-mode active guidance training and asymmetric assistance for infants with congenital cerebral palsy, and has the ability of abnormal mode identification and safe compliant response.

Description

Technical Field

[0001] This invention belongs to the field of wearable exoskeleton technology, specifically, it relates to a wearable exoskeleton device and exoskeleton control method suitable for rehabilitation training of infants with congenital cerebral palsy. Background Technology

[0002] Cerebral palsy is mainly characterized by unilateral limb motor dysfunction, with decreased muscle strength, abnormal muscle tone, and limited joint movement in the affected upper and lower limbs, leading to asymmetrical movement patterns in children. These patterns mainly include: 1. Spastic type (accounting for 60%~70%, the most common): limb stiffness, high muscle tone, scissor gait, toe walking, and pigeon-toed walking; 2. Athetoid type: involuntary twisting, distorted movements, facial movements, slurred speech, slightly better when quiet; 3. Ataxia type: poor balance, unsteady gait, inaccurate hand and foot movements, unsteady standing, and poor coordination; 4. Rigid type: generalized stiffness, minimal movement, limbs like sticks that are difficult to bend or straighten; 5. Tremor type: constant shaking of limbs, extremely poor fine motor skills; 6. Hypotonic type: generalized weakness, inability to lift head, inability to sit or stand, resembling a "soft noodle"; 7. Mixed type: the simultaneous presence of two or more types of symptoms.

[0003] In terms of walking ability, the typical characteristics of children with congenital cerebral palsy are extremely slow walking speed, asymmetrical gait, poor balance, and in severe cases, they may not be able to walk independently at all.

[0004] Currently, rehabilitation methods for children with congenital cerebral palsy mainly include the following categories: The first category is traditional physical rehabilitation methods, such as therapist-assisted massage, passive joint range of motion training, and neurodevelopmental therapy. These methods rely on the therapist's techniques and experience, and the intensity and frequency of training are limited by labor costs. It is also difficult to ensure the consistency and accuracy of each training movement. Moreover, for the rehabilitation of children's walking ability, the effect of traditional massage methods is relatively limited, and it is difficult to provide highly repetitive and high-intensity gait training stimulation.

[0005] The second type is functional electrical stimulation rehabilitation training, which involves applying electrical stimulation to specific muscle groups during the child's walking process to assist muscle contraction. However, this method has a clear age limit and is usually only applicable to children over 7 years old. For children under 7 years old, their neuromuscular system is not yet fully developed, and electrical stimulation may cause discomfort or even damage. Therefore, this method is generally not recommended or cannot be used in clinical practice, resulting in a significant gap in technical means for gait rehabilitation training for children under 7 years old with congenital cerebral palsy.

[0006] The third category is lower limb exoskeleton rehabilitation devices. Existing lower limb exoskeleton devices are mainly designed for adult stroke patients with hemiplegia or spinal cord injury. Their control strategy is usually bilateral symmetrical assistance, that is, applying the same or fixed proportion of motor assistance to both lower limbs. However, children with congenital cerebral palsy have unique pathological characteristics: there is a significant difference in motor ability between the affected side and the healthy side. Bilateral symmetrical assistance cannot effectively inhibit compensation on the healthy side and strengthen training on the affected side. Summary of the Invention

[0007] The purpose of this invention is to provide a wearable exoskeleton device suitable for rehabilitation training of infants with congenital cerebral palsy, so as to solve the technical problems existing in the prior art.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a wearable exoskeleton device suitable for rehabilitation training of infants with congenital cerebral palsy, comprising a main body of the wearable exoskeleton device, and hip joint actuators, knee joint actuators, and ankle actuators symmetrically arranged on the left and right lower limb parts of the main body of the wearable exoskeleton device; further comprising a controller communicatively connected to all actuators, the controller comprising: The affected side setting unit receives the affected side setting command and determines the affected side of the current wearer. The storage unit stores joint target trajectory data that corresponds one-to-one with multiple training modes, including flat ground walking mode, in-place walking mode, side walking mode, and backward walking mode. The mode selection unit is used to receive mode selection instructions and determine the current training mode. The data retrieval unit is used to retrieve joint target trajectory data corresponding to the current training mode. The control command unit generates control commands based on the joint target trajectory data, driving the hip joint actuator, knee joint actuator, and ankle actuator to coordinate their movements, thereby actively moving the wearer's affected lower limb and completing the action sequence defined in the training mode.

[0009] Furthermore, the joint target trajectory data includes trajectory data of the affected side and trajectory data of the healthy side, and the trajectory data includes the sagittal plane flexion-extension trajectory of the hip joint, the sagittal plane flexion-extension trajectory of the knee joint, and the dorsiflexion / plantarflexion trajectory of the ankle joint.

[0010] Furthermore, for the same joint under the same training mode, the peak auxiliary torque defined by the affected side trajectory data is N times the peak auxiliary torque defined by the healthy side trajectory data, and the joint motion angle range defined by the affected side trajectory data is M times the joint motion angle range defined by the healthy side trajectory data, where 1 < N < 3, 1 < M < 2.

[0011] Furthermore, it also includes multiple sets of sensors disposed on the main body of the wearable exoskeleton device for detecting the movement status of the wearer's affected and healthy sides; the multiple sets of sensors include: Multiple electromyography (EMG) sensors are respectively arranged on the main body of the wearable exoskeleton device at positions corresponding to the rectus femoris, biceps femoris, tibialis anterior, and gastrocnemius muscles of the wearer's bilateral lower limbs, and are used to detect the degree of muscle activation and pre-spasticity signals of key muscle groups on the affected and healthy sides. An inertial measurement unit, including a three-axis accelerometer and a three-axis gyroscope, is arranged in the torso, thigh and lower leg segments of the wearable exoskeleton device body to detect the attitude angle, angular velocity and linear acceleration of each segment of the wearer's torso and lower limbs in real time. Multiple pressure sensors are installed at the binding connection points where the wearable exoskeleton device body contacts the wearer's limbs, to detect the pushing or pulling force actively applied by the wearer to the wearable exoskeleton device body.

[0012] Furthermore, the controller also includes an information receiving unit for receiving detection signals from the multiple sets of sensors in real time; an anomaly identification unit for identifying, based on the detection signals, whether the wearer exhibits an abnormal movement pattern triggered by an intention to exert force; and a control adjustment unit for dynamically adjusting the control gain of the driver when the abnormal movement pattern is identified.

[0013] Furthermore, the abnormal movement patterns include combined reactions and extensor spasms.

[0014] In a second aspect, the present invention provides an exoskeleton control method for rehabilitation training of children with congenital cerebral palsy, applied to the wearable exoskeleton device described in the first aspect above, comprising: (a) Configuration phase: Store joint target trajectory data corresponding one-to-one with multiple training modes, wherein the multiple training modes include at least flat ground walking mode, stationary walking mode, side walking mode and backward walking mode; (b) Control phase: Receives the affected side setting command to determine the affected side of the current wearer; Receive mode selection instructions and determine the current training mode; Perform the following control steps: Retrieve the joint target trajectory data corresponding to the current training mode; Based on the target trajectory data of the joints, control commands are generated to drive the hip joint actuator, knee joint actuator, and ankle actuator to move in a coordinated manner, so as to actively drive the wearer's lower limb movement on the affected side and complete the action sequence defined by the training mode.

[0015] Furthermore, it also includes a method for dynamically adjusting the control gain of the driver based on the combined reaction: S101: Receives detection signals from electromyography (EMG) sensors, inertial measurement units (IMUs), and pressure sensors in real time; S102: When the amplitude of the electromyographic signal of the muscle group exerting force on the affected lower limb exceeds the force threshold, and at the same time the amplitude of the electromyographic signal of the muscle group exerting force on the healthy lower limb exceeds the linkage threshold, it is initially determined to be a combined reaction. S103: If the human-machine interaction force signal detected by the pressure sensor on the healthy side does not show active force application characteristics in sync, it is confirmed as a combined reaction triggered by force from the affected side. S104: Reduce the control gain of the corresponding driver to reduce the tracking stiffness of the driver for the joint target trajectory data; Simultaneously, the electromyographic signals, inertial measurement unit signals, and pressure signals are continuously monitored to determine whether the combined reaction has been eliminated.

[0016] Furthermore, it also includes a method for dynamically adjusting the control gain of the actuator based on extensor spasm: S201: Receives detection signals from the electromyography sensor and the inertial measurement unit in real time; S202: When the inertial measurement unit detects that the extension angular velocity of the affected knee joint exceeds a preset angular velocity threshold, and the electromyography sensor detects that the amplitude of the electromyography signal of the quadriceps femoris on the affected side exceeds the spasm threshold for a preset duration, it is determined to be extensor spasm. S203: Reduce the control gain of the knee joint actuator to reduce the tracking stiffness of the knee joint actuator on the joint target trajectory data, while maintaining the normal tracking of the hip joint actuator and ankle actuator on their respective joint target trajectory data. Simultaneously, the electromyographic signals and inertial measurement unit signals are continuously monitored to determine whether the extensor spasm has been eliminated.

[0017] Once it is determined that the combined reaction or the extensor spasm has been eliminated and the stable state has lasted for a preset time, the control gain of the driver is gradually restored to the normal level, and the original training mode is continued.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention uses a controller to pre-store joint target trajectory data corresponding to multiple training modes. It can actively drive the wearer's affected lower limb to move along a predetermined trajectory according to the selected training mode, providing highly repetitive and highly consistent multi-mode gait rehabilitation training for children with congenital cerebral palsy. It fills the technical gap that children under 7 years old cannot use functional electrical stimulation for gait training.

[0019] This invention distinguishes joint target trajectory data into trajectory data of the affected side and trajectory data of the healthy side, and makes the range of motion and auxiliary torque of the trajectory data of the affected side greater than that of the healthy side, thereby achieving asymmetric auxiliary control. It can effectively suppress the compensatory behavior of the healthy side that is common in the training of children with cerebral palsy, strengthen the intensity of motor training of the affected side, and thus promote the recovery of motor function of the affected side in a targeted manner, overcoming the shortcomings of traditional bilateral symmetrical assistance of exoskeletons.

[0020] This invention integrates multiple sensors, including electromyography, inertial measurement unit and pressure sensor, to perceive the physiological and motor state of the wearer's affected and healthy sides in real time. Based on the fusion data of multiple sensors, the controller can accurately identify abnormal movement patterns unique to children with cerebral palsy, such as combined reactions and extensor spasticity.

[0021] After identifying abnormal movement patterns, this invention dynamically reduces the control gain of the actuator, thereby reducing the tracking stiffness of the actuator on the joint target trajectory data. This allows for the compliant release of abnormal forces, preventing secondary damage to children in a state of spasticity or abnormal reflexes caused by rigid trajectory tracking, and significantly improving safety. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the main body of the wearable exoskeleton device in this invention. Figure 1 .

[0023] Figure 2 This is a schematic diagram of the main body of the wearable exoskeleton device in this invention. Figure 2 .

[0024] Figure 3 This is a system block diagram of the controller and sensor in this invention.

[0025] Figure 4 This is a flowchart illustrating Embodiment 2 of the present invention.

[0026] Figure 5 This is a physical reference for the main body of the wearable exoskeleton device in this invention. Figure 1 .

[0027] Figure 6 This is a physical reference for the main body of the wearable exoskeleton device in this invention. Figure 2 .

[0028] The corresponding names of the attached figures are as follows: 1. Controller; 2. Hip joint actuator; 3. Knee joint actuator; 4. Ankle actuator; 5. Thigh length adjuster; 6. Lower leg length adjuster; 7. Handle. Detailed Implementation

[0029] To enable those skilled in the art to have a clearer understanding of the present invention, the present invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described below are merely for illustrative purposes and to facilitate understanding. The technical solutions provided by the present invention are not limited to those provided in the following embodiments, nor should they limit the scope of protection of the present invention.

[0030] In the description of this invention, it should be noted that the term "affected side" refers to the side of the limb motor function disorder in children with congenital cerebral palsy, and "healthy side" refers to the side of the limb motor function that is relatively normal; the term "sagittal plane" refers to the midline plane that divides the human body into left and right parts, and "sagittal plane flexion and extension" refers to the back-and-forth swinging movement of the limbs in this plane; the term "dorsiflexion" refers to the bending of the foot towards the front of the lower leg, and "plantar flexion" refers to the bending of the foot towards the sole; the term "control gain" refers to the proportional coefficient between the controller output and the input signal. A decrease in gain means that a smaller control quantity is output under the same input deviation, and the system stiffness is correspondingly reduced. Example 1

[0031] like Figures 1-3 As shown, this embodiment provides a wearable exoskeleton device suitable for rehabilitation training of infants with congenital cerebral palsy. The device includes a main body 8, which comprises a body support, a pelvic support, a thigh support, a lower leg support, and a foot support. Each support is equipped with a strap for securing the main body 8 to the corresponding part of the wearer's body. The body support fits against the wearer's back, the pelvic support surrounds both sides of the wearer's waist, the thigh and lower leg supports extend along the front of the wearer's thigh and lower leg, respectively, and the foot support supports the wearer's feet. Each support uses a lightweight aluminum alloy frame combined with flexible padding. The straps feature an adjustable Velcro design to accommodate the wearing needs of children of different body types. Preferably, the thigh support and calf support of the wearable exoskeleton device body 8 further include: a thigh length adjuster 5, a device for adjusting the thigh length (hip joint driven - knee joint driven); and a calf length adjuster 6, a device for adjusting the calf length (knee driven - ankle driven), to accommodate the wearing needs of children of different body types. Two symmetrical handles 7 may be provided on the top of the wearable exoskeleton device body 8, which assistants can grip when supporting the product. A stop button electrically connected to the controller 1 described below may be provided on the handles 7.

[0032] On the left and right lower limbs of the wearable exoskeleton device body 8, hip joint actuators 2, knee joint actuators 3 and ankle actuators 4 are symmetrically arranged. Each actuator includes a DC brushless servo motor and a reducer connected to it. The output shaft of the motor drives the corresponding joint to rotate after the torque is amplified by the reducer.

[0033] The hip joint actuator 2 is installed at the joint connecting the pelvic support and the thigh support, and is used to drive the wearer's hip joint to complete flexion and extension movements in the sagittal plane. The knee joint actuator 3 is installed at the joint connecting the thigh support and the lower leg support, and is used to drive the knee joint to complete flexion and extension movements in the sagittal plane. The ankle actuator 4 is installed at the joint connecting the lower leg support and the foot support, and is used to drive the ankle joint to complete dorsiflexion and plantarflexion movements. The installation positions of the above actuators are such that the output shaft of each actuator is basically aligned with the physiological rotation axis of the wearer's corresponding joint, thereby driving the wearer's hip, knee, and ankle joints to complete flexion and extension movements in the sagittal plane under the drive of the controller 1. Preferably, each actuator integrates an angle encoder and a torque sensor, which are used to provide real-time feedback on the current joint angle and output torque, respectively, forming a closed-loop control of position and torque.

[0034] The wearable exoskeleton device also includes a controller 1, which is fixedly mounted on the back plate of the main body 8 of the wearable exoskeleton device. The controller 1 communicates with all drivers via a CAN bus to issue control commands and transmit sensor data. In this embodiment, the core processor of the controller 1 uses an ARM Cortex-M7 series microcontroller to ensure the computing power required for real-time control. The controller 1 is also equipped with a touch screen as a human-machine interface and has a built-in rechargeable lithium battery to power the entire system.

[0035] In this embodiment, the controller 1 is functionally divided into multiple units, and the specific configuration of each unit is as follows: 1. Affected side setting unit The affected side setting unit is used to receive the affected side setting instruction and determine the affected side of the current wearer. In this embodiment, the touch screen of the controller 1 provides an affected side selection interface, which displays two option buttons: "left hemiplegia" and "right hemiplegia". Before each training session, the therapist or parent selects the corresponding option based on the clinical diagnosis to generate the affected side setting instruction. After receiving the instruction, the affected side setting unit writes the affected side information (left or right) into the global configuration parameters of the controller 1 for all subsequent control logic to reference. This manual setting method avoids automatic recognition errors caused by the low cognitive cooperation and atypical spontaneous motor performance of young children, ensuring the accuracy and reliability of the affected side information.

[0036] 2. Storage unit The storage unit uses a large-capacity Flash memory to store joint target trajectory data corresponding to multiple training modes. In this embodiment, the multiple training modes include seven modes in total: flat ground walking mode, in-place walking mode, side walking mode, backward walking mode, standing mode, sitting mode, and squatting mode.

[0037] The joint target trajectory data for each training mode is generated from gait data of normal children selected from a clinical gait analysis database, and then adapted by rehabilitation physicians based on the joint range of motion and motor ability of children with congenital cerebral palsy. Specifically, the joint target trajectory data consists of discrete sampling sequences of joint angles changing over time, with a sampling period of 10 milliseconds. Each set of joint target trajectory data contains a data sequence with six channels: left hip joint sagittal plane flexion-extension target angle, left knee joint sagittal plane flexion-extension target angle, left ankle joint dorsiflexion / plantarflexion target angle, right hip joint sagittal plane flexion-extension target angle, right knee joint sagittal plane flexion-extension target angle, and right ankle joint dorsiflexion / plantarflexion target angle. Each sampling point includes a target angle value (unit: degrees) and a target angular velocity value (unit: degrees / second). The controller uses an interpolation algorithm to achieve a smooth transition between adjacent sampling points during the control process.

[0038] Preferably, the aforementioned joint target trajectory data includes trajectory data from the affected side and trajectory data from the healthy side. The trajectory data from the affected side is used to drive the actuator on the affected side, while the trajectory data from the healthy side is used to drive the actuator on the side with relatively normal motor function. Taking the flat-ground walking mode as an example, both the trajectory data from the affected side and the trajectory data from the healthy side contain data sequences from the aforementioned six channels. The difference lies in the following: for the same channel (such as the same joint in the same training mode), the peak assist torque defined by the trajectory data from the affected side is N times that of the trajectory data from the healthy side (where N=1.5), and the range of joint motion angles defined by the trajectory data from the affected side is M times that of the trajectory data from the healthy side (where M=1.3). This asymmetrical trajectory design allows the exoskeleton to specifically strengthen the motor output of the affected limb during training, while allowing the healthy side to primarily function as an auxiliary support, effectively suppressing compensatory behavior on the healthy side.

[0039] 3. Mode selection unit The mode selection unit is used to receive mode selection instructions and determine the current training mode; the touch screen of controller 1 presents a training mode selection menu to the user before each training session begins, listing all available training modes for therapists or parents to choose from; after receiving the selection instruction input by the user through the touch screen, the mode selection unit writes the current training mode identifier into the running status variable of controller 1.

[0040] 4. Data retrieval unit The data retrieval unit is used to retrieve the joint target trajectory data corresponding to the current training mode. Specifically, the data retrieval unit reads the current training mode identifier in the running status variable, queries the preset mode and data mapping table in the storage unit according to the identifier, obtains the starting address and data length of the corresponding joint target trajectory data in the Flash memory, and then loads the data block into the RAM of the controller 1 for real-time access by the control instruction unit.

[0041] 5. Control command unit The control command unit generates control commands based on the joint target trajectory data loaded into RAM by the data retrieval unit, and drives the hip joint actuator 2, knee joint actuator 3 and ankle actuator 4 to coordinate their movements, so as to actively drive the wearer's affected lower limb to complete the action sequence defined by the training mode.

[0042] The control command unit adopts a three-loop cascaded control architecture of position, speed, and torque. The inner loop is the torque loop, the middle loop is the speed loop, and the outer loop is the position loop. The position loop takes the target angle in the joint target trajectory data as input and the current angle fed back by the angle encoder inside each driver as feedback. After calculation by the proportional-derivative controller, it outputs the target angular velocity command. The speed loop takes the target angular velocity output by the position loop as input and the actual angular velocity obtained by differentiating the current angle fed back by the angle encoder inside each driver as feedback. After calculation by the proportional-integral controller, it outputs the target torque command. The torque loop takes the target torque output by the speed loop as input and the actual torque fed back by the torque sensor inside the driver as feedback. After calculation by the proportional controller, it outputs the pulse width modulation duty cycle signal of the drive motor. The control gain of each loop can be adjusted as needed.

[0043] It should be noted that, in this embodiment, the proportional-derivative controller, the proportional-integral controller, and the proportional controller are all stored in the program memory of the controller 1 in the form of program code, and are called and executed by the central processing unit of the controller 1 in each control cycle.

[0044] During training, the control instruction unit executes the above control in a fixed control cycle (e.g., 1 millisecond), calculates and issues control instructions in real time, so that each actuator drives the wearer's lower limbs to complete the target action along a predetermined trajectory.

[0045] In this embodiment, the wearable exoskeleton device also includes multiple sets of sensors disposed on the main body 8 of the wearable exoskeleton device for detecting the movement status of the wearer's affected and healthy sides. The multiple sets of sensors specifically include: multiple electromyography sensors, an inertial measurement unit, and multiple pressure sensors. The specific settings of each sensor are as follows: Multiple electromyography (EMG) sensors 5 employ Ag / AgCl surface electrodes and are respectively arranged on the inner side of the flexible binding strap of the main body of the device, corresponding to four key muscle groups of the wearer's bilateral lower limbs: rectus femoris (front of the thigh, responsible for knee extension and hip flexion), biceps femoris (posterior of the thigh, responsible for knee flexion and hip extension), tibialis anterior (front of the lower leg, responsible for ankle dorsiflexion), and gastrocnemius (posterior of the lower leg, responsible for ankle plantarflexion); a total of 8 EMG sensors are arranged on both sides. Each EMG sensor collects the surface EMG signal of the corresponding muscle group. After the signal is pre-amplified, bandpass filtered (20-500Hz), and full-wave rectified, the real-time EMG amplitude envelope signal is output, which is used to detect the degree of muscle activation and pre-spasticity signals of the corresponding muscle group; The inertial measurement unit (IMU) comprises multiple sensor nodes, each integrating a triaxial accelerometer and a triaxial gyroscope. These nodes are located on the torso (back panel), left thigh, right thigh, left lower leg, and right lower leg segments of the main body 8 of the device, totaling five nodes. Each node detects the triaxial acceleration and triaxial angular velocity of its corresponding segment in real time. A complementary filtering algorithm is used to fuse the accelerometer and gyroscope data to calculate the real-time attitude angles (pitch, roll, yaw), angular velocity, and linear acceleration of each segment. The pitch and roll angles detected by the torso nodes are used to assess the tilt of the wearer's torso in the sagittal and coronal planes. The attitude angles detected by the lower limb nodes are used to determine the actual angles of each joint in space. The angular velocity and linear acceleration data are used to assess the dynamic characteristics of each segment's motion. Each sensor node of the IMU is only responsible for raw data acquisition, directly sending the raw triaxial acceleration and triaxial angular velocity data to the controller 1 for processing to complete the calculation of the attitude angles of each node. Multiple pressure sensors 7 are flexible piezoresistive thin-film pressure sensors. Preferably, each has a thickness of no more than 0.5 mm. The pressure sensors are respectively set at the binding connection points where the main body of the device contacts the wearer's limbs. Specifically, they include: one on the front and one on the back of the thigh binding strap (4 in total on both sides), one on the front and one on the back of the calf binding strap (4 in total on both sides), and one on the forefoot and one on the heel of the foot support plate (4 in total on both sides). The pressure sensors on the thigh and calf binding straps are used to detect the pushing or pulling force actively applied by the wearer to the main body of the device, i.e., the magnitude and direction of the human-machine interaction force; the pressure sensors on the foot support plate are used to detect the contact force and contact time between the wearer's foot and the ground.

[0046] In this embodiment, the controller 1 further includes an information receiving unit, an anomaly identification unit, and a control adjustment unit, which are used to monitor the wearer's status in real time during training, identify abnormal movement patterns, and dynamically adjust the control gain.

[0047] The information receiving unit is used to receive the detection signals from the above-mentioned multiple sensors in real time, namely the electromyographic amplitude envelope signal of the electromyographic sensor, the attitude angle / angular velocity / acceleration signal of the inertial measurement unit, and the human-computer interaction force and plantar pressure signal of the pressure sensor; the information receiving unit synchronizes and buffers the data of each sensor according to a unified timestamp for the anomaly identification unit to read. Based on sensor data provided by the information receiving unit, the anomaly recognition unit identifies whether the wearer exhibits an abnormal movement pattern triggered by an intention to exert force. In this embodiment, the abnormal movement patterns include combined reactions and extensor spasms. The specific recognition logic for the two abnormal movement patterns is described below: 1. Identification of combined reactions Associated reactions are abnormal movement patterns in children with congenital cerebral palsy. The neurophysiological mechanism is as follows: when a child attempts to use the affected limb to complete a certain movement, due to the impaired inhibitory function of the central nervous system on motor excitation, the excitation spreads to the contralateral side and even the motor neuron pool of the contralateral upper limb, resulting in involuntary, task-irrelevant muscle activation and coordinated movements in the healthy limb. For example, when a child tries to bend the hip joint on the affected side to take a step, the rectus femoris muscle on the healthy side may simultaneously exhibit involuntary contraction. The presence of associated reactions not only interferes with the establishment of normal movement patterns but may also lead to the continuous reinforcement of abnormal movement patterns.

[0048] The anomaly detection unit identifies combined reactions based on sensor detection signals: Step 1: Monitor the electromyographic signal amplitude of the bilateral lower limb muscle groups (rectus femoris and tibialis anterior) in real time. When the electromyographic signal amplitude of the affected lower limb muscle group exceeds the preset force threshold (e.g., 30% of the electromyographic amplitude during maximum voluntary contraction), trigger pathological linkage monitoring.

[0049] Step 2: At the same time as the triggering in Step 1, check whether the amplitude of the electromyographic signal of the corresponding force-generating muscle group in the healthy lower limb exceeds the preset linkage threshold (e.g., 15% of the electromyographic amplitude during maximum voluntary contraction). If the electromyographic signal of the affected side exceeds the force-generating threshold and the electromyographic signal of the healthy side also exceeds the linkage threshold, it is initially determined to be a combined reaction. The linkage threshold of the healthy side is set relatively low here because the involuntary activation of the healthy side in the combined reaction is usually weaker than the active force exertion of the affected side.

[0050] Step 3: To avoid misjudgment (e.g., the child is indeed actively exerting force on the healthy side at the same time), cross-validation is performed using the human-machine interaction force signal detected by the pressure sensor on the healthy side. Specifically, check whether the human-machine interaction force detected by the pressure sensor at the binding strap on the healthy side synchronously exhibits active force characteristics, that is, whether the interaction force is applied along the movement direction of the healthy side and the amplitude exceeds the active force threshold. If the interaction force on the healthy side does not synchronously exhibit the above active force characteristics, it is confirmed that the activation of the healthy side muscle is not from the active movement intention of the healthy side, but a combined response triggered by the force exerted by the affected side.

[0051] 2. Identification of extensor muscle spasm Extensor spasticity is another abnormal movement pattern of the lower limbs in children with cerebral palsy. It is usually triggered by rapid traction or active force and manifests as sudden and sustained high-tension contraction of extensor muscle groups such as the quadriceps femoris, resulting in forced hyperextension of the knee joint and / or plantar flexion of the ankle joint. In severe cases, it can cause pain and joint damage.

[0052] The anomaly detection unit identifies extensor spasm based on signals detected by an inertial measurement unit and an electromyography sensor. Step 1: Monitor the dynamic movement of the affected knee joint in real time. When the extension angular velocity of the affected knee joint exceeds the preset angular velocity threshold (e.g., 150° / s), it indicates that the knee joint is undergoing rapid, non-physiological extension movement, triggering pathological spasm monitoring. Step 2: Simultaneously with the triggering in Step 1, check the electromyography (EMG) sensor signal of the affected quadriceps femoris muscle to determine whether the EMG signal amplitude continuously exceeds the preset spasticity threshold (e.g., 50% of the EMG amplitude during maximum voluntary contraction) and the duration reaches the preset duration (e.g., 500 milliseconds). If the EMG signal amplitude continuously exceeds the spasticity threshold and reaches the preset duration, it is determined to be extensor spasticity. The duration condition is set here to distinguish the high amplitude EMG caused by spasticity from the short-term high amplitude EMG during normal exertion.

[0053] When the abnormality identification unit determines that an abnormal motion pattern has occurred, the control adjustment unit dynamically adjusts the control gain of the driver to avoid discomfort or damage to the child in an abnormal state caused by rigid trajectory tracking control.

[0054] 1. Control and adjustment of combined reactions When the anomaly detection unit confirms a combined reaction, the control adjustment unit performs the following adjustments: immediately reduces the control gain of all actuators, specifically by multiplying the proportional gain and differential gain of the position loop by a compliance factor α (e.g., α=0.3). After the gain is reduced, the target torque amplitude output by the control command unit decreases accordingly, and the tracking stiffness of the actuators for the joint target trajectory data decreases, allowing the wearer's affected and healthy limbs to deviate from the preset trajectory to a certain extent, thereby releasing the abnormal stress caused by the combined reaction and preventing the exoskeleton from clashing with the child's abnormal movement patterns. During this process, the actuators do not completely stop working, but follow the wearer's limb movements in a low-stiffness, compliant manner, thus providing a safety protection function.

[0055] 2. Control and adjustment of extensor muscle spasm When the anomaly detection unit determines that extensor spasm has occurred, the control adjustment unit performs the following adjustments: It selectively reduces the control gain of the knee actuator 3 by multiplying the proportional gain and differential gain of the knee actuator position loop by a compliance factor α (e.g., α = 0.2); simultaneously, it maintains normal tracking of the target trajectory data of the hip actuator 2 and ankle actuator 4, i.e., it keeps the control gains of the hip and ankle joints unchanged. This selective gain reduction strategy allows the affected knee joint to passively flex to release extensor spasm, while the hip and ankle joints maintain a preset support posture, preventing the child from falling due to sudden knee instability.

[0056] After the control adjustment unit reduces the control gain, the anomaly detection unit continuously monitors the corresponding sensor signals to determine whether the abnormal motion pattern has been eliminated. (1) For combined reactions: continuously monitor electromyographic signals and pressure signals on the affected and healthy sides. When the amplitude of the electromyographic signal of the muscle group exerting force on the affected side drops below the force threshold, or the amplitude of the electromyographic signal of the muscle group exerting force on the healthy side drops below the linkage threshold, and this state continues for a preset safe recovery time (e.g., 1 second), it is determined that the combined reaction has been eliminated.

[0057] (2) For extensor spasm: continuously monitor the electromyographic signal of the quadriceps femoris on the affected side and the angular velocity of the knee joint. When the amplitude of the electromyographic signal of the quadriceps femoris drops below the spasm threshold and the angular velocity of the knee joint returns to the normal range, and this state continues for a preset safe recovery time (e.g., 1 second), it is determined that the extensor spasm has been eliminated.

[0058] Once the abnormality identification unit determines that the abnormal motion pattern has been eliminated and the stable state has been maintained for a preset safe recovery time, the control adjustment unit gradually restores the control gain of the driver to the normal level. Specifically, the recovery method is as follows: the compliance factor α is increased by 0.1 every 100 milliseconds in a ramp-increment manner until α is restored to 1.0 (i.e., normal gain is restored) to avoid discomfort to the wearer caused by sudden gain changes. After the gain is restored, the control command unit continues to execute the joint target trajectory data of the original training mode, smoothly transitioning from the current joint position to the preset trajectory to ensure the continuity of training. Example 2

[0059] like Figure 4 As shown, this embodiment also provides an exoskeleton control method for rehabilitation training of children with congenital cerebral palsy. The method is applied to the wearable exoskeleton device provided in Embodiment 1. The controller stores joint target trajectory data corresponding to multiple training modes one by one. The multiple training modes include at least flat walking mode, stationary walking mode, side walking mode and backward walking mode. The joint target trajectory data of each mode is pre-stored in the Flash memory of the controller. The specific generation method has been detailed in the device embodiment.

[0060] The control method is implemented by controller 1, as follows: The system receives the affected side setting instruction and determines the affected side of the current wearer. The therapist or parent manually selects the affected side as the left or right side via the touch screen of the controller 1. The controller 1 writes this information into the global configuration parameters. This step is performed once at the start of each training session or when the wearer is changed.

[0061] Upon receiving the mode selection instruction, the current training mode is determined; the therapist or parent selects the training mode to be used in this training session from the mode menu via the touch screen, and the controller 1 writes the selected mode identifier into the running status variable.

[0062] After the above is completed, perform the following control steps: Step 1: Retrieve the joint target trajectory data corresponding to the current training mode; Controller 1 indexes and loads the corresponding trajectory data from the memory into RAM according to the mode identifier in the running status variable.

[0063] Step 2: Based on the joint target trajectory data, generate control commands to drive the hip joint actuator, knee joint actuator, and ankle actuator to coordinate their movements, so as to actively drive the wearer's affected lower limb movement and complete the action sequence defined in this training mode; the control command generation adopts a three-ring cascaded control architecture, which has been detailed in the device embodiment.

[0064] During the training process, controller 1 executes the abnormal movement pattern monitoring and control adjustment sub-process in parallel, including the combined response processing sub-process and the extensor spasticity processing sub-process: 1. Combined reaction processing sub-process: Step S101: Receive detection signals from the electromyography (EMG) sensor, inertial measurement unit (IMU), and pressure sensor in real time. The EMG sensor provides the EMG amplitude envelope signals of the rectus femoris, biceps femoris, tibialis anterior, and gastrocnemius muscles of both lower limbs. The IMU provides the posture angle, angular velocity, and acceleration of each segment. The pressure sensor provides the human-machine interaction force at the binding point and the plantar pressure signal.

[0065] Step S102: When the amplitude of the electromyographic signal of the muscle group (rectus femoris and / or tibialis anterior) of the affected lower limb exceeds the preset force threshold, and the amplitude of the electromyographic signal of the muscle group of the healthy lower limb exceeds the preset linkage threshold, it is initially determined to be a combined reaction.

[0066] Step S103: Cross-validate the human-machine interaction force signal detected by the pressure sensor on the healthy side. If the interaction force does not show active force characteristics synchronously, that is, no push or pull force is detected along the movement direction of the healthy side and the amplitude exceeds the active force threshold, the possibility of active force exertion on the healthy side is ruled out, and it is confirmed as a combined reaction triggered by force exertion on the affected side.

[0067] Step S104: After confirming the joint reaction, immediately reduce the control gain of the corresponding actuator, that is, multiply the proportional gain and differential gain of the position loop by the compliance factor α (e.g., α=0.3), so that the tracking stiffness of the actuator for the joint target trajectory data is reduced, and abnormal stress is released in a compliant manner.

[0068] After the gain is reduced, the electromyographic signal, inertial measurement unit signal and pressure signal are continuously monitored to determine whether the combined reaction has been eliminated. The specific determination criteria are: the amplitude of the electromyographic signal of the muscle group exerting force on the affected side drops below the force threshold, or the amplitude of the electromyographic signal of the muscle group exerting force on the healthy side drops below the linkage threshold, and this state continues for a preset safe recovery time (e.g., 1 second).

[0069] 2. Sub-process for treating extensor muscle spasm: Step S201: Receive detection signals from the electromyography sensor and the inertial measurement unit in real time.

[0070] Step S202: When the inertial measurement unit detects that the extension angular velocity of the affected knee joint exceeds the preset angular velocity threshold (e.g., 150° / s), and the electromyography sensor detects that the amplitude of the electromyography signal of the affected quadriceps muscle continuously exceeds the spasticity threshold (e.g., 50% of the electromyography amplitude during maximum voluntary contraction) for a preset duration (e.g., 500 milliseconds), it is determined to be extensor spasticity.

[0071] Step S203: After determination, reduce the control gain of the knee joint actuator, multiply the proportional gain and differential gain by the compliance factor α (e.g., α=0.2) to allow passive flexion of the affected knee to release spasticity; at the same time, maintain normal tracking of the target trajectory data of the hip joint actuator 2 and the ankle actuator 4, that is, keep the control gain of the hip and ankle joints unchanged and maintain the support posture.

[0072] After the gain is reduced, the electromyography signal and the inertial measurement unit signal are continuously monitored to determine whether the extensor spasm has been eliminated. The specific determination criteria are: the amplitude of the quadriceps electromyography signal drops below the spasm threshold and the knee joint angular velocity returns to the normal range, and this state continues for a preset safe recovery time (e.g., 1 second).

[0073] Once the abnormal motion mode has been eliminated and the stable state has been maintained for a preset safe recovery time, the control adjustment unit gradually restores the control gain of the driver to the normal level. The specific recovery method is as described in the scheme provided in Example 1, and will not be repeated here.

[0074] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A wearable exoskeleton device suitable for rehabilitation training of infants with congenital cerebral palsy, comprising a main body (8) of the wearable exoskeleton device, and hip joint actuators (2), knee joint actuators (3), and ankle actuators (4) symmetrically arranged on the lower limbs on both sides of the main body (8); characterized in that, It also includes a controller (1) that is communicatively connected to all drives, the controller (1) comprising: The affected side setting unit receives the affected side setting command and determines the affected side of the current wearer. The storage unit stores joint target trajectory data that corresponds one-to-one with multiple training modes, including flat ground walking mode, in-place walking mode, side walking mode, and backward walking mode. The mode selection unit is used to receive mode selection instructions and determine the current training mode. The data retrieval unit is used to retrieve joint target trajectory data corresponding to the current training mode. The control command unit generates control commands based on the target trajectory data of the joints, and drives the hip joint actuator (2), knee joint actuator (3) and ankle actuator (4) to coordinate their movements, so as to actively drive the wearer's affected lower limb to complete the action sequence defined by the training mode.

2. The wearable exoskeleton device for rehabilitation training of infants with congenital cerebral palsy according to claim 1, characterized in that: The joint target trajectory data includes trajectory data of the affected side and trajectory data of the healthy side. The trajectory data includes the sagittal plane flexion-extension trajectory of the hip joint, the sagittal plane flexion-extension trajectory of the knee joint, and the dorsiflexion / plantarflexion trajectory of the ankle joint.

3. The wearable exoskeleton device for rehabilitation training of infants with congenital cerebral palsy according to claim 2, characterized in that: For the same joint under the same training mode, the peak auxiliary torque defined by the affected side trajectory data is N times the peak auxiliary torque defined by the healthy side trajectory data, and the joint motion angle range defined by the affected side trajectory data is M times the joint motion angle range defined by the healthy side trajectory data, where 1 < N < 3 and 1 < M < 2.

4. The wearable exoskeleton device for rehabilitation training of infants with congenital cerebral palsy according to claim 1, characterized in that: It also includes multiple sets of sensors disposed on the main body (8) of the wearable exoskeleton device for detecting the movement status of the wearer's affected and healthy sides; the multiple sets of sensors include: Multiple electromyography (EMG) sensors are respectively arranged on the main body (8) of the wearable exoskeleton device at positions corresponding to the rectus femoris, biceps femoris, tibialis anterior, and gastrocnemius muscles of the wearer's bilateral lower limbs, and are used to detect the degree of muscle activation and pre-spasticity signals of key muscle groups on the affected and healthy sides. An inertial measurement unit, including a three-axis accelerometer and a three-axis gyroscope, is arranged in the torso, thigh and lower leg segments of the wearable exoskeleton device body (8) to detect the attitude angle, angular velocity and linear acceleration of each segment of the wearer's torso and lower limbs in real time. Multiple pressure sensors are installed at the binding connection between the wearable exoskeleton device body (8) and the wearer's limbs to detect the pushing or pulling force actively applied by the wearer to the wearable exoskeleton device body (8).

5. The wearable exoskeleton device for rehabilitation training of infants with congenital cerebral palsy according to claim 4, characterized in that: The controller (1) further includes an information receiving unit for receiving detection signals from the multiple sets of sensors in real time; an anomaly identification unit for identifying whether the wearer exhibits an abnormal movement pattern triggered by the intention to exert force based on the detection signals; and a control adjustment unit for dynamically adjusting the control gain of the driver when the abnormal movement pattern is identified.

6. The wearable exoskeleton device for rehabilitation training of infants with congenital cerebral palsy according to claim 5, characterized in that: The abnormal movement patterns include combined reactions and extensor spasms.

7. An exoskeleton control method for rehabilitation training of children with congenital cerebral palsy, applied to the wearable exoskeleton device according to any one of claims 1 to 6, characterized in that, include: (a) Configuration phase: Store joint target trajectory data corresponding one-to-one with multiple training modes, wherein the multiple training modes include at least flat ground walking mode, stationary walking mode, side walking mode and backward walking mode; (b) Control phase: Receives the affected side setting command to determine the affected side of the current wearer; Receive mode selection instructions and determine the current training mode; Perform the following control steps: Retrieve the joint target trajectory data corresponding to the current training mode; Based on the target trajectory data of the joints, control commands are generated to drive the hip joint actuator, knee joint actuator, and ankle actuator to move in a coordinated manner, so as to actively drive the wearer's lower limb movement on the affected side and complete the action sequence defined by the training mode.

8. The exoskeleton control method according to claim 7, characterized in that, It also includes methods for dynamically adjusting the control gain of the driver based on the combined reaction: S101: Receives detection signals from electromyography (EMG) sensors, inertial measurement units (IMUs), and pressure sensors in real time; S102: When the amplitude of the electromyographic signal of the muscle group exerting force on the affected lower limb exceeds the force threshold, and at the same time the amplitude of the electromyographic signal of the muscle group exerting force on the healthy lower limb exceeds the linkage threshold, it is initially determined to be a combined reaction. S103: If the human-machine interaction force signal detected by the pressure sensor on the healthy side does not show active force application characteristics in sync, it is confirmed as a combined reaction triggered by force from the affected side. S104: Reduce the control gain of the corresponding driver to reduce the tracking stiffness of the driver for the joint target trajectory data.

9. The exoskeleton control method according to claim 7, characterized in that, It also includes methods for dynamically adjusting the control gain of the actuator based on extensor spasm: S201: Receives detection signals from the electromyography sensor and the inertial measurement unit in real time; S202: When the inertial measurement unit detects that the extension angular velocity of the affected knee joint exceeds a preset angular velocity threshold, and the electromyography sensor detects that the amplitude of the electromyography signal of the quadriceps femoris on the affected side exceeds the spasm threshold for a preset duration, it is determined to be extensor spasm. S203: Reduce the control gain of the knee joint actuator to reduce the tracking stiffness of the knee joint actuator on the joint target trajectory data, while maintaining normal tracking of the hip joint actuator and ankle actuator on their respective joint target trajectory data.

10. The exoskeleton control method according to claim 8 or 9, characterized in that, The electromyography signal, inertial measurement unit signal, and pressure signal are continuously monitored. When it is determined that the abnormal movement pattern has been eliminated and the stable state has lasted for a preset time, the control gain of the driver is gradually restored to the normal level, and the original training mode is continued.

Images