Knee joint orthopedic rehabilitation brain-computer interface exoskeleton power-assisted control method and system

Abstract

The invention discloses a knee joint orthopedic rehabilitation brain-computer interface exoskeleton power-assisted control method and system, and relates to the technical field of knee joint rehabilitation. The method comprises the following steps: based on an electroencephalogram electrode cap and a knee joint angle sensor, analyzing a corresponding relation between an electroencephalogram main peak and a joint action, combining plantar pressure change, screening gait characteristics, judging an intention signal and an action trend, proofreading exoskeleton execution, and obtaining a closed-loop action offset index. According to multi-dimensional trend analysis of the action execution process, electroencephalogram intention judgment is dynamically associated with the knee joint action process, intention signals and action state changes are comprehensively analyzed through trend synchronous judgment and continuous behavior cooperation, and layer-by-layer linkage of intentions, behaviors and the execution process is achieved; the control output has real-time adaptation and compliant coordination ability, the training stage can adapt to multi-source signal changes, the false triggering probability is effectively reduced, the action continuity is guaranteed, and the man-machine coordination consistency and the rehabilitation training safety guarantee ability are improved.

Description

Technical Field

[0001] This invention relates to the field of knee joint rehabilitation technology, and in particular to a brain-computer interface exoskeleton-assisted control method and system for knee joint orthopedic rehabilitation. Background Technology

[0002] Knee rehabilitation involves a range of techniques and methods to restore knee function, alleviate pain, and improve patients' quality of life. Knee rehabilitation techniques mainly include physical therapy, post-surgical rehabilitation, and treatment using external assistive devices. Among these, traditional exoskeleton-assisted control methods in knee orthopedic rehabilitation refer to techniques that use exoskeleton devices to assist in the control of the knee joint. These typically include a control system that uses sensors and algorithms to monitor the patient's movements in real time and adjusts the exoskeleton's movement patterns according to their needs, thereby providing precise rehabilitation training support.

[0003] Existing methods mostly rely on single action data as the basis for execution. Action judgment only focuses on surface action signals and sensor changes, lacking deep collaborative recognition between brain movement intention and the actual process of action. This makes it difficult for assisted control to accurately match the user's active participation in movement. During training, it is easily affected by environmental interference, signal drift and non-standard movements, and there is a risk of false triggering and sudden changes. The safety and smoothness of the action process are insufficient, and it is unable to maintain reliable human-machine collaborative stable output under complex and changing training scenarios. Summary of the Invention

[0004] To address the technical problems existing in the prior art, this invention provides a brain-computer interface exoskeleton assistance control method and system for knee joint orthopedic rehabilitation. The technical solution is as follows:

[0005] On the one hand, a brain-computer interface exoskeleton-assisted control method for knee joint orthopedic rehabilitation is provided, including the following steps: S1: Based on the EEG electrode cap and knee joint angle sensor, analyze the correspondence between the EEG main peak change and the joint movement record, and simultaneously compare the change time periods of the two types of signals to determine whether the main peak fluctuation occurs at the same time as the joint movement, and obtain the EEG co-discrimination feature. S2: Based on the aforementioned EEG co-discrimination features, monitor the synchronous trend of joint movement changes and foot pressure distribution changes during walking, compare the direction of trajectory change with the direction of force change, identify continuous movement intervals, and obtain gait behavior co-parameters. S3: Based on the gait behavior coordination parameters, determine the consistency between the EEG signal intention characteristics and the knee joint movement trend, analyze the direction of data change within the same time period, correct the exoskeleton execution state, and obtain the motion command fusion factor. S4: Based on the motion command fusion factor, start the exoskeleton knee joint execution, compare the consistency between the motor running direction and the joint movement direction, record the change process of the movement trajectory over time, and obtain the knee joint execution trajectory range. S5: Based on the knee joint execution trajectory range, compare the action state fed back by the knee joint angle sensor with the target action, analyze the difference trend between the feedback position and the action endpoint, compensate and adjust the offset action, and obtain the closed-loop action offset index.

[0006] Optionally, the EEG co-discrimination features include peak response features, synchronization association features, and temporal coordination features; the gait behavior coordination parameters include stride length features, stride frequency features, and pressure distribution features; the motor command fusion factor includes action synchronization features, trend consistency features, and command coordination features; the knee joint execution trajectory range includes action continuity features, trajectory morphology features, and response amplitude features; and the closed-loop action offset index includes action offset features, compensation correction features, and continuous stability features.

[0007] Optionally, the steps for obtaining the EEG collaborative discrimination features specifically include: S101: Based on the EEG electrode cap and knee joint angle sensor, analyze EEG signals and joint movement data, correspond the main peak time point of the EEG signal motion-related frequency band with the change time of knee joint movement, judge the synchronization performance of the data on the time axis, and obtain the brain-motor synchronization event group. S102: Based on the brain-motor synchronization event group, compare the direction of change of the main peak of brain electrical activity and the direction of change of knee joint movement, determine whether the trends of the two are synchronized, identify data combinations with consistent trends, and obtain the neuro-joint trend cluster. S103: Based on the aforementioned neural joint trend cluster, aggregate EEG fluctuation temporal features, action response rhythm features, and data overlap intervals, classify and integrate the data according to the features, and obtain EEG collaborative discrimination features.

[0008] Optionally, the steps for obtaining the gait behavior coordination parameters specifically include: S201: Based on the EEG collaborative discrimination features, analyze the data sequence output by the knee joint angle sensor, group the changing direction of continuous angle rotation segments, determine the start and end positions and the order of direction changes of each rotation segment, identify segments with changing characteristics within the walking cycle, aggregate the rotation cycle according to the change sequence, and obtain a joint gait structure segment group. S202: Based on the joint gait structure segment group, compare it with the data collected by the plantar pressure sensor, analyze the movement trajectory of the force center of each section of the plantar pressure, determine the correspondence between the trajectory change direction and the knee joint rotation direction, identify the synchronously changing data segments, and obtain the segment set corresponding to the gait pressure joint. S203: Based on the segment set corresponding to the step pressure joint, determine the segments with consistent trends, count the time overlap interval between the knee joint movement state and the foot force state, classify the state switching process, identify the gait stage according to the joint action state type, and obtain the gait behavior coordination parameters.

[0009] Optionally, the step of obtaining the motion command fusion factor specifically includes: S301: Based on the gait behavior coordination parameters, analyze the EEG electrode cap signal and the motion data of the knee joint angle sensor in the corresponding time period, compare the direction of change of the data in each gait cycle, determine the correspondence between the change pattern of the neural signal and the rotation trend of the joint movement, count the segments with consistent change trends, and obtain the brain-joint trend coordination sequence. S302: Based on the brain joint trend coordination sequence, determine the direction of exoskeleton execution unit movement corresponding to each trend consistency interval, compare the synchronization relationship between the exoskeleton unit movement and the knee joint angle change trend, analyze the correspondence between signal direction and movement direction in each segment, identify synchronization segments, and obtain execution trend synchronization patch group. S303: Based on the execution trend synchronization patch group, calculate the coordination of EEG signal trend, knee joint rotation trend and exoskeleton execution direction in the same time interval, analyze the synchronization and direction consistency of each type of data, and obtain the motion command fusion factor.

[0010] Optionally, the step of obtaining the knee joint execution trajectory range specifically includes: S401: Based on the motion command fusion factor, analyze the action sequence of the motor drive device, determine whether the action direction required by the action command matches the action direction output by the drive device, identify the action segment with the same direction, and obtain the action direction matching segment. S402: Based on the motion direction matching segment, compare the motion process executed by the exoskeleton knee joint, monitor the correspondence between the output motion segment of each motor and the joint motion direction, determine whether the change process of the motion path shows a consistent trend with the motion executed by the knee joint, and obtain the joint path collaborative mapping set. S403: Based on the joint path collaborative mapping set, determine the time overlap interval between the motor-driven motion path and the knee joint motion sequence, analyze the morphological characteristics and amplitude coordination relationship of the path in each interval, classify and summarize the trajectory range during the execution of each motion segment, and obtain the knee joint execution trajectory interval.

[0011] Optionally, the steps for obtaining the closed-loop action offset index are as follows: S501: Based on the knee joint execution trajectory range, compare it with the target action state fed back by the knee joint angle sensor, analyze the spatial correspondence between the position of the feedback action and the end point of the motor trajectory range, determine the offset between each group of positions, and obtain the action space offset parameter group. S502: Based on the motion space offset parameter group, determine the offset phenomenon between the feedback position of each group and the end point of the motor trajectory interval, analyze the distribution of the offset segment, adjust the motor output motion sequence, correct the joint motion for the detected offset motion segment, and obtain the offset compensation execution sequence. S503: Based on the offset compensation execution sequence, monitor the performance in continuous action feedback, determine the offset change of each action sequence after compensation, analyze the continuity and stability of the corrected action, summarize the parameters adjusted for each action, and obtain the closed-loop action offset index.

[0012] Optionally, the main peak change represents the amplitude point in the motion-related frequency band of the EEG waveform, which is a characteristic signal of motion intention, and the change time period represents the time point when the main peak change of the EEG signal and the knee joint angle change occur respectively.

[0013] Optionally, the intention feature refers to the EEG features collected by the EEG electrode cap that reflect the patient's active motor imagination or intention, and the joint movement trend refers to the direction of movement shown in the knee joint angle data, including a continuous bending or extension trend.

[0014] On the other hand, a brain-computer interface exoskeleton assistive control system for knee joint orthopedic rehabilitation is provided. This system is applied to a brain-computer interface exoskeleton assistive control method for knee joint orthopedic rehabilitation, including: The EEG action discrimination module, based on the EEG electrode cap and knee joint angle sensor, analyzes the correspondence between the changes in the main peak of EEG and the joint action recording, and simultaneously compares the change periods of the two types of signals to determine whether the main peak fluctuation occurs at the same time as the joint action, thereby obtaining EEG collaborative discrimination features. The gait synchronization recognition module, based on the aforementioned EEG collaborative discrimination features, monitors the synchronization trend of joint movement changes and foot pressure distribution changes during walking, compares the trajectory change direction with the force change direction, identifies continuous movement intervals, and obtains gait behavior coordination parameters. The intention trend fusion module, based on the gait behavior coordination parameters, determines the consistency between the intention features of the EEG signal and the trend of knee joint movement, analyzes the direction of data change within the same time period, corrects the exoskeleton execution state, and obtains the motion command fusion factor. The trajectory generation module, based on the motion command fusion factor, initiates the execution of the exoskeleton knee joint, compares the consistency between the motor running direction and the joint movement direction, records the change process of the movement trajectory over time, and obtains the knee joint execution trajectory range. The closed-loop compensation module, based on the knee joint execution trajectory range, compares the action state fed back by the knee joint angle sensor with the target action, analyzes the difference trend between the feedback position and the action endpoint, compensates and adjusts the offset action, and obtains the closed-loop action offset index.

[0015] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, based on multi-dimensional trend analysis of the action execution process, the EEG intention judgment is dynamically associated with the knee joint movement process. Through trend synchronization judgment and continuous behavior coordination, the intention signal and the change of action state are comprehensively analyzed to achieve layer-by-layer linkage of intention, behavior and execution process. The control output has real-time adaptability and compliant coordination ability. During the training phase, it can adapt to changes in multi-source signals, improve the matching degree of active intention, effectively reduce the probability of false triggering, ensure the continuity of action, and improve the consistency of human-machine collaboration and the safety guarantee of rehabilitation training. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the main steps of the present invention; Figure 2 This is a flowchart of steps S1 of the present invention; Figure 3 This is a flowchart of steps S2 of the present invention; Figure 4 This is a flowchart of steps S3 of the present invention; Figure 5 This is a flowchart of step S4 of the present invention; Figure 6 This is a flowchart of steps S5 of the present invention; Figure 7 This is a system block diagram of the present invention. Detailed Implementation

[0018] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0019] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0020] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0021] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] This invention provides a brain-computer interface exoskeleton-assisted control method for knee joint orthopedic rehabilitation, referencing... Figure 1 As shown, it includes the following steps: S1: Based on the EEG electrode cap and knee joint angle sensor, analyze the correspondence between the EEG main peak change and the joint movement record, and simultaneously compare the change time periods of the two types of signals to determine whether the main peak fluctuation occurs at the same time as the joint movement, and obtain the EEG co-discrimination feature. Specifically, based on EEG electrode caps and knee joint angle sensors, the changes in the main peak in scalp EEG signals and knee joint movement records are analyzed. The timing of changes in brain signals and movement data is compared synchronously. By judging whether the main peak fluctuation is accompanied by joint movement, the synergistic phenomenon of signal and movement is summarized, and EEG synergistic discrimination features are obtained.

[0024] S2: Based on the aforementioned EEG co-discrimination features, monitor the synchronous trend of joint movement changes and foot pressure distribution changes during walking, compare the direction of trajectory change with the direction of force change, identify continuous movement intervals, and obtain gait behavior co-parameters. Specifically, based on EEG collaborative discrimination features, the output data of the knee joint angle sensor is screened, the changes in joint movement and the changes in force distribution of the foot pressure sensor are monitored during walking, the joint rotation trajectory and pressure distribution change trends are compared, and the gait state is recorded by identifying the trend-synchronized action range, thus obtaining gait behavior coordination parameters.

[0025] S3: Based on the gait behavior coordination parameters, determine the consistency between the EEG signal intention characteristics and the knee joint movement trend, analyze the direction of data change within the same time period, correct the exoskeleton execution state, and obtain the motion command fusion factor. Specifically, based on gait behavior coordination parameters, the intention signals and joint movement trends of the EEG electrode cap and knee joint angle sensor are judged, the change direction of the two data in the same time period is analyzed, the exoskeleton execution state is corrected, and the brain signal changes and movement direction are jointly judged to obtain the motor command fusion factor.

[0026] S4: Based on the motion command fusion factor, start the exoskeleton knee joint execution, compare the consistency between the motor running direction and the joint movement direction, record the change process of the movement trajectory over time, and obtain the knee joint execution trajectory range. Specifically, based on the motion command fusion factor, the action sequence of the motor drive device is adjusted, and the exoskeleton knee joint is activated according to the action requirements in the command. The consistency between the motor operation and the joint action direction is compared. By recording the change process between the motor action path and the joint execution action, the knee joint execution trajectory range is obtained.

[0027] S5: Based on the knee joint execution trajectory range, compare the action state fed back by the knee joint angle sensor with the target action, analyze the difference trend between the feedback position and the action endpoint, compensate and adjust the offset action, and obtain the closed-loop action offset index. Specifically, based on the knee joint execution trajectory range, the target motion in the knee joint angle sensor motion feedback is compared, the correspondence between the feedback position and the end point of the motor range is analyzed, the motion with deviation is compensated and adjusted, the continuous motion of the knee joint is continuously tracked, the feedback correction situation is summarized, and the closed-loop motion deviation index is obtained.

[0028] In this embodiment of the invention, the EEG coordinating discrimination features include peak response features, synchronous correlation features, and temporal coordination features; the gait behavior coordination parameters include stride length features, stride frequency features, and pressure distribution features; the motor command fusion factors include action synchronization features, trend consistency features, and command coordination features; the knee joint execution trajectory range includes action continuity features, trajectory morphology features, and response amplitude features; and the closed-loop action offset index includes action offset features, compensation correction features, and continuous stability features.

[0029] In S1, the main peak change represents the point of maximum amplitude in the motor-related frequency band of the EEG waveform, and its timing, shape or amplitude changes significantly, which is a characteristic signal related to motor intention; the knee joint movement record represents the joint angle, rotation direction and its continuous sequence of changes over time collected in real time by the knee joint angle sensor, reflecting the actual movement process of the knee joint; the change time represents the specific time point when the EEG signal main peak change and the knee joint angle change occur respectively, and the two are compared through a synchronous time axis; the main peak fluctuation represents the appearance, disappearance, amplitude change or position shift of the main peak, which is a reflection of the brain signal in the state of motor intention; the synergistic phenomenon indicates that the EEG signal characteristic changes and the knee joint movement changes show a synchronous and continuous state on the time axis.

[0030] In S2, the output data represents the continuous angle sequence generated by the knee joint angle sensor during gait, including information such as amplitude, direction, and time; joint movement changes represent the continuous changes in knee joint angle data, such as bending and extension; force distribution changes represent the changes in force on each zone of the plantar pressure sensor over time within the gait cycle, such as the force transfer between the heel and forefoot; joint rotation trajectory represents the curve formed by the knee joint angle changing over time, reflecting the direction and rhythm of knee joint movement; pressure distribution change trend represents the direction of change of plantar pressure data over time, such as the gradual transfer of force from the hindfoot to the forefoot; the trend-synchronized action interval represents the time period in which knee joint rotation and plantar pressure changes change synchronously within the same time window; gait state represents the stage of gait movement within the synchronous action interval, such as stepping, support phase, or swing phase.

[0031] In S3, the intention signal represents the EEG characteristics collected by the EEG electrode cap that reflect the patient's active motor imagination or intention; the joint movement trend represents the direction of movement shown in the knee joint angle data, such as the trend of continuous bending or extension; the change direction represents the change trend of brain signals and joint angles within the same time period, including the direction of motor intention and the actual direction of movement; the exoskeleton execution status represents the current operating status of the exoskeleton drive module, including whether an action is executed, joint position and other real-time information; the joint discrimination represents the collaborative analysis based on the change trend of brain signals and joint data within the same time period to determine whether the motor intentions of the two are consistent.

[0032] In S4, the motor drive unit represents the drive unit of the exoskeleton knee joint, including the motor and transmission mechanism, which is responsible for executing knee joint movements; the movement requirements represent the specific requirements for the direction and amplitude of knee joint movements in the motion command; consistency represents whether the direction of the motor's execution of the movement is consistent with the direction required by the motion command; the motor movement path represents the trajectory of the motor during the drive process, reflecting the changes in rotation sequence and amplitude; the change process represents the working process of the motor drive component over time, including parameters such as displacement and direction.

[0033] In S5, the target motion in the motion feedback represents the target state information recorded by the knee joint angle sensor after the motion is completed; the motor interval endpoint represents the termination position point reached after the motor motion is completed; the correspondence represents the positional correspondence and deviation trend between the target motion and the motor endpoint motion; the compensation adjustment represents the correction of the actual joint motion by readjusting the motor output according to the motion deviation; the continuous knee joint motion represents the motion sequence formed when the knee joint completes multiple motion cycles continuously; and the feedback correction status represents the deviation correction status between the actual knee joint motion and the target motion after the compensation adjustment.

[0034] like Figure 2 As shown, the specific steps for obtaining EEG collaborative discrimination features are as follows: S101: Based on the EEG electrode cap and knee joint angle sensor, analyze EEG signals and joint movement data, correspond the main peak time point of the EEG signal motion-related frequency band with the change time of knee joint movement, judge the synchronization performance of the data on the time axis, and obtain the brain-motor synchronization event group. Multiple channels of electroencephalogram (EEG) signals were collected from the wearer's scalp surface, and the knee joint angle variation curve during continuous gait was recorded simultaneously. The EEG signal sampling frequency was set to 250 data points per second, with no fewer than 8 sampling channels. An angle sensor synchronously output the knee joint angle at a rate of 100 data points per second. The EEG signals were then divided into fixed time windows for segmented processing. Within each time window, the point with the largest amplitude in the signal waveform was retrieved, its occurrence time was extracted and recorded, forming a set of main peak time point sequences. Simultaneously, the knee joint angle variation values ​​were differentially processed to obtain the angle abrupt change time points. The two sets of time points were compared and aligned according to a unified time axis to check each... Whether the peak time point appears within a certain time window before or after a certain joint change point. If the time interval between a certain peak time point and the joint movement change is less than the set standard time value, it is recorded as a pair of synchronous events. For example, when a user performs a stepping action, the peak of their EEG appears at a certain time point at the beginning of the gait. Then, the knee joint angle quickly transitions from a flexed state to an extended state. After detecting the angle change, it is found that the time point of the movement change and the peak time point of the EEG are less than the set maximum allowable time interval. Then, the combination of these time points is determined to be a set of synchronous events. Multiple similar synchronous time pairs are recorded as event sets to form a brain-motor synchronous event group, which is used for subsequent trend analysis and movement intention recognition.

[0035] S102: Based on the brain-motor synchronization event group, compare the direction of change of the main peak of brain electrical activity and the direction of change of knee joint movement, determine whether the trends of the two are synchronized, identify data combinations with consistent trends, and obtain the neuro-joint trend cluster. The operation compares the direction of change of the EEG peak with the direction of change of knee joint movement in each group of events. First, the waveform height change trend of each EEG peak at consecutive time points is extracted. The relationship between the peak heights of the peaks in consecutive time periods is used to determine whether it is an upward or downward trend. At the same time, the change trend of knee joint angle between two adjacent movement points is examined. If the joint angle value increases from small to large, it is judged as an extension trend, and vice versa. The directions of change of the two are marked as the same or opposite. Then, all event pairs are classified into groups with consistent or inconsistent trends, and the number information of events with consistent trends is recorded to form a neuro-joint trend. In practice, a trend cluster is a data aggregation structure. For example, when a patient is undergoing gait training, if the main peak of the EEG signal shows a continuous increase over a period of time, and the knee joint angle also continuously changes in the direction of extension during this period, it indicates that the EEG signal and the joint movement change in the same direction during that time period. This consistent trend event is recorded and categorized. When multiple similar consistent trend events show high consistency in terms of time distribution, signal change amplitude, and movement direction, they are combined into a trend cluster. This trend cluster contains all data combinations that show the matching of the direction of brain signal and joint movement changes, forming a data aggregation structure.

[0036] S103: Based on the trend cluster of neural joints, aggregate the temporal features of EEG fluctuations, the rhythmic features of action response, and the overlapping intervals of data. Classify and integrate the data according to the features to obtain EEG collaborative discrimination features; The included data is further processed to extract key features. First, an overall fluctuation feature analysis is performed on the timing, location, and amplitude changes of the EEG peak within a cluster to assess its stability and fluctuation intensity throughout the entire movement cycle. Then, the movement response time intervals of the knee joint within the corresponding time period of the event are statistically analyzed, and the average interval between consecutive movements is extracted and categorized. Simultaneously, the regularity of movement occurrence is examined, such as whether the time intervals between movements remain basically stable. If the fluctuations are small within an acceptable range, it is marked as a rhythmically stable trend. Furthermore, the degree of overlap between the EEG peak occurrence time and the joint movement occurrence time in each trend cluster is extracted. If this overlap time is within a set... If the time exceeds the lower limit, it is considered that there is an effective temporal coordination relationship between the data points. Data segments with high coordination are extracted and classified as effective overlapping intervals. Finally, the above-mentioned EEG fluctuation features, action rhythm features, and temporal overlap features are integrated to construct a set of three-dimensional feature description data structures, representing the performance of the trend cluster under multiple feature dimensions. For example, if the EEG fluctuations in a trend cluster remain within a stable range, the action interval time is highly regular, and most of the main peaks overlap with the action time intervals, then this set of data will be identified as a feature combination with strong coordination. This result is characterized as EEG coordination discrimination features, serving as an important basic data input for recognizing motor intentions.

[0037] like Figure 3 As shown, the specific steps for obtaining gait behavior coordination parameters are as follows: S201: Based on EEG co-discrimination features, analyze the data sequence output by the knee joint angle sensor, group the changing direction of continuous angle rotation segments, determine the start and end positions and the order of direction changes of each rotation segment, identify segments with changing characteristics within the walking cycle, and aggregate the rotation cycle according to the change sequence to obtain a joint gait structure segment group. The system retrieves time-series data collected by a knee joint angle sensor during continuous gait. The angle values ​​at each moment are sorted chronologically to form a complete angle change curve. Then, the angle sequence is scanned segment by segment within a set time window. The continuous difference sequence between angle values ​​within each window is calculated. The direction of angle rotation is determined by the sign of the difference. When multiple consecutive differences remain positive or negative, a segment with the same direction of rotation is defined. The time points corresponding to the start and end of the rotation segment are used as the boundaries of the segment. The duration, direction of rotation, and start and end times of each segment are recorded. Multiple segments are organized chronologically to form an angle rotation segment list. This list is then aggregated according to movement trends, grouping multiple segments that occur consecutively within a single walking cycle and have the same or alternating direction. The system combines segments to determine whether there is a trend shift from flexion to extension or from extension to flexion between adjacent segments. If there is a shift in the rotational direction between two segments, the gait cycle boundary is split according to the shift point. Based on this, the rhythm of angle sequence changes, directional sequence, and fluctuation range of maximum angle value within each gait cycle are analyzed to identify key movement segments within the adjacent movement direction switching interval. These segments are then recombined according to their change sequence to form a structural segment group that stably presents joint movements in each cycle during continuous walking. For example, in a test subject's 5-second walking data, the knee joint angle continuously rises from 35 degrees to 72 degrees, then slides down to 40 degrees and rises again. This change sequence is divided into three independent directional segments and classified into two complete rotation cycles, forming a joint gait structural segment group.

[0038] S202: Based on the joint gait structure segment group, compare with the data collected by the plantar pressure sensor, analyze the movement trajectory of the force center of each zone of the foot, determine the correspondence between the trajectory change direction and the knee joint rotation direction, identify the synchronously changing data segments, and obtain the segment set corresponding to the gait pressure joint; The start and end time ranges recorded within each rotation cycle are extracted, and the plantar pressure data matrix output by the plantar pressure sensor within the same time period is extracted simultaneously. This data includes the pressure values ​​of the heel, arch, and forefoot at each time point. The center pressure position of each frame of data is extracted according to the plantar segment location to form a center trajectory sequence. For each plantar pressure center position point, the direction of the spatial coordinate difference between the preceding and following time points is analyzed to obtain the movement trajectory direction of the plantar pressure center. Then, the movement direction is correlated with the change direction of the knee joint angle during the same period. When the plantar pressure center moves along the forefoot direction while the knee joint angle shows a continuous upward trend, or when the pressure center... If the joint angle decreases as the heart moves backward, it is considered that the two directions are the same, and this time period is recorded as a segment with consistent direction. All structural segments are compared with the direction of the foot trajectory one by one, and the start and end times, joint rotation direction, pressure center change direction, and overlap duration are recorded. When the direction is consistent throughout the overlap time period, the data segment is marked as a synchronous change segment. In actual operation, for example, if the test subject's joint angle increases from 45 degrees to 65 degrees in the time period from 3.2s to 3.6s, and the pressure center of the foot transitions from the heel to the forefoot in a forward direction with a continuous trend, then this time period is determined to be a synchronous change segment. All similar segments are combined into a segment set corresponding to the step pressure joint.

[0039] S203: Based on the segment set corresponding to the step pressure joint, determine the segments with consistent trends, count the time overlap interval between the knee joint motion state and the foot force state, classify the state switching process, identify the gait stage according to the joint motion state type, and obtain the gait behavior coordination parameters. All data segments are sequentially screened. For each pair of segments where plantar pressure and knee joint movement direction correspond, the start and end times are further determined to ensure a continuous trend with the preceding and following segments. If the time interval between the start time of the current segment and the end time of the previous segment is within a set time tolerance range (e.g., a tolerance of 0.2 seconds), the two segments are considered to be able to be spliced ​​into a continuous trend segment. All splicable segments are merged into a group of segments with consistent continuous trends. Based on this, the overlap duration of knee joint movement and plantar pressure changes on the time axis in each merged segment is calculated. Each overlap interval is classified according to the rate of change, direction of change of knee joint angle value, and regional transfer speed of plantar pressure, and each state is further classified. The transition process is categorized into specific gait phases. For example, when the angle gradually increases and the center of pressure moves from the hind foot to the forefoot, this state is marked as the stepping phase. When the angle remains stable and the pressure remains concentrated in the midfoot area, it is determined to be the support phase. When the angle continues to decrease and the pressure gradually shifts towards the heel, it is marked as the swing phase. Each data segment is marked according to the above state classification results, and the gait phase information of each time period is output to construct gait behavior coordination parameters. For example, if the test subject continuously increases the angle between 4.0 seconds and 4.8 seconds, and the pressure gradually shifts from the heel to the forefoot and remains overlapping for about 0.3 seconds, it is marked as the stepping phase and used as one of the gait behavior coordination parameters for that time period.

[0040] like Figure 4 As shown, the specific steps for obtaining the motion command fusion factor are as follows: S301: Based on gait behavior coordination parameters, analyze the EEG electrode cap signal and knee joint angle sensor motion data in the corresponding time period, compare the direction of data change in each gait cycle, determine the correspondence between the change pattern of neural signals and the rotation trend of joint movements, and statistically analyze the segments with consistent change trends to obtain the brain-joint trend coordination sequence. The time range corresponding to each gait stage is extracted, and within this time range, the raw multi-channel signal data recorded by the EEG electrode cap and the angle change sequence recorded by the knee joint angle sensor are extracted. Differential analysis is performed on the amplitude trend of the EEG signals on the same time axis to determine whether the main peak in the continuous waveform is rising, falling, or remaining stable. Simultaneously, the knee joint angle sequence is differentially processed over time to confirm whether its rotation trend within this time range is continuously increasing, decreasing, or remaining constant. Then, the direction of change of the EEG main peak is compared one-to-one with the direction of change of the knee joint angle; when the trends of the two are consistent, they are marked as positive. Coordinating segments, if in opposite directions, are marked as reversed or irrelevant segments. Then, the above comparison operation is performed on all time segments within each gait cycle, and time segments with continuous consistent trends are integrated into a trend-consistent segment. In actual testing, for example, if the wearer's EEG peak continuously increases during the stride phase, and the knee joint angle continuously increases from 45 degrees to 70 degrees, then the data in this phase is judged to be trend-consistent. If there are more than three similar trend-consistent segments within the entire gait cycle, they are statistically analyzed and numbered to form multiple time segment combinations where the EEG change trend is consistent with the joint rotation direction, and these are combined into a brain-joint trend coordination sequence.

[0041] S302: Based on the brain joint trend coordination sequence, determine the direction of exoskeleton execution unit movement corresponding to each trend consistent interval, compare the synchronization relationship between exoskeleton unit movement and knee joint angle change trend, analyze the correspondence between signal direction and movement direction in each segment, identify synchronization segments, and obtain execution trend synchronization patch group. The start and end times of each trend-consistent segment are retrieved, and the motion control commands and motor feedback data of the exoskeleton execution unit are extracted within the corresponding time period. The motor drive state of this segment is recorded as a sequence of motion directions. The sequence is then differentially processed according to time to determine whether the execution direction is clockwise or counterclockwise. This motion direction is then compared item by item with the direction of knee joint angle change within the same time period. If the exoskeleton motor execution direction is consistent with the knee joint angle direction (i.e., the motor output is positive drive when the angle increases, or negative drive when the angle decreases), then the exoskeleton execution state and knee joint movement are considered synchronized within that time period, and this segment is marked as a synchronized segment. Conversely, if the directions are inconsistent... If the signal is out of sync, it is then compared again with the main peak trend of the EEG signal during that time period. If the main peak trend of the EEG, the knee joint rotation trend, and the exoskeleton execution direction all maintain the same direction of change, then the segment is recorded as a segment where the signal and the execution action are completely corresponding. In the test sample, for example, if the main peak of the EEG continues to rise, the joint angle rises from 55 degrees to 68 degrees, the motor output state is continuously positive drive and the torque feedback continues to increase during a certain period of time from t=6.1s to t=6.4s, then the segment is determined to be a synchronous segment of the execution trend and is included in the synchronous segment group of the execution trend. All segments that meet the synchronization characteristics of the three factors constitute the complete sequence of the segment group.

[0042] S303: Based on the execution trend synchronization patch group, calculate the coordination between the EEG signal trend, the knee joint rotation trend and the exoskeleton execution direction in the same time interval, analyze the synchronization and directional consistency of each type of data, and obtain the motion command fusion factor. For each synchronized segment, the trend signs of three data types—the main peak change in EEG, knee joint angle rotation, and exoskeleton execution direction—are calculated within the time interval. The consistency of these three data types in that segment is determined by whether their sign values ​​are the same. The proportion of time during which this consistency persists throughout the entire time interval is then calculated. If the proportion of time during which the three trends are completely consistent exceeds a set threshold for the total segment duration (e.g., 80%), the segment is recorded as a high-coordination segment. Subsequently, the consistency rate and duration of each direction are summarized for each segment, and the overall synchronization ratio (the proportion of segments with synchronized directions of the three data types to the total observation time) is calculated. If this ratio is within a set high-coordination range (e.g., above 0.75), it is considered a highly synchronized stage. The corresponding directional characteristics, duration, and rate of change (referring to a certain...) are then... The rate of change of a signal or parameter over time can also be interpreted as "slope" or "the speed of increase or decrease," reflecting whether the data trend is rising or falling, or changing quickly or slowly. These key indicators form a set of control input parameters (referring to the multi-dimensional feature parameter set for subsequent control of exoskeleton movements, used for motion control). This is marked as the fusion parameter set for the current time period. In a complete experiment, when the directional synchronization ratio of a segment exceeds a set threshold (e.g., 0.8), it is marked as a high-coordination segment. Subsequently, key indicators such as directional consistency rate, synchronization duration, and trend change rate are extracted from all high-coordination segments to form a set of control input feature parameters. If there are three high-coordination segments in a certain experiment, with synchronization ratios of 0.83, 0.85, and 0.88 respectively, then three sets of fusion feature parameters can be formed as the motion command fusion features for this experimental stage.

[0043] like Figure 5 As shown, the specific steps for obtaining the knee joint execution trajectory range are as follows: S401: Based on the motion command fusion factor, analyze the action sequence of the motor drive device, determine whether the action direction required by the action command matches the action direction output by the drive device, identify the action segment with the same direction, and obtain the action direction matching segment. Extract the time period corresponding to each fusion factor and three data points: the EEG trend direction, the knee joint movement trend, and the expected execution direction of the exoskeleton. Sequentially compare the recorded movement direction in the fusion factor with the actual output execution direction of the exoskeleton motor within the same time period. During the comparison, a one-to-one match is performed between the expected movement direction of the fusion factor at each time point and the current output direction of the motor. If the two direction markers match, the time point is recorded as a consistent direction point; otherwise, it is marked as a deviation point. Consecutive consistent direction points are grouped into a movement segment, and the start and end times, duration, and direction marker of each segment are recorded. During the judgment process, if the direction marker is positive and the motor is currently... If the output is clockwise, it is considered consistent. If the direction marker is negative and the motor output is counterclockwise, it is also considered consistent. Otherwise, it is inconsistent. Based on the continuity of the consistent points in the direction, the segment is divided. The minimum duration threshold for consistent direction is set to 0.3 seconds. If the continuous time of the consistent points in a segment is less than this threshold, it is not recorded. For example, in the test, if a user's fusion factor output direction is continuously positive during gait training, and the motor continuously outputs positively for 2.4 seconds to 2.9 seconds, this period of time meets the consistency and minimum duration requirements. This period of time is recorded as a direction matching segment. All segments that meet the consistency of direction and pass the duration judgment are combined to form a motion direction matching segment.

[0044] S402: Based on motion direction matching segments, compare the motion process executed by the exoskeleton knee joint, monitor the correspondence between the output motion segment of each motor and the joint motion direction, determine whether the change process of the motion path shows a consistent trend with the motion executed by the knee joint, and obtain the joint path collaborative mapping set; Extract the motor output angle sequence and the joint sensor output angle change sequence corresponding to the segment. Compare the trends of the two data streams within each segment over the corresponding time period. First, divide the motor angle change into continuous time windows on the time axis and calculate the angle increase / decrease trend within each window. Then, compare the knee joint angle change trend within the same time window to determine if the two change directions are consistent within that window. For example, if the motor angle increases from 20 degrees to 45 degrees within a certain time window, and the knee joint angle increases from 30 degrees to 50 degrees within the same time window, then the direction within that window is considered consistent, and it is recorded as a trend matching segment. Continuously combine such trend-consistent windows into segments with consistent motion trends, and establish... The correlation between the motor output motion path and the actual joint motion trend is established. Then, based on the motor drive parameters and joint sensor feedback values ​​of each consistent trend segment, the angle change amplitude, motion rhythm characteristics, and relative start and end times are marked. The presence of obvious rhythm synchronization or phase shift phenomena is analyzed. In some tests, the exoskeleton motor advances the angle at a rate of 20 degrees per second during the motion execution process, while the knee joint maintains an angle change rate of 18 degrees per second during the process. The fluctuation range does not exceed the set error limit of 2 degrees / second. This segment is considered to have a consistent motion process trend. All time periods where the motor motion and the joint angle execution process trend are consistent are collected to form a joint path co-mapping set.

[0045] S403: Based on the joint path coordination mapping set, determine the time overlap interval between the motor-driven motion path and the knee joint motion sequence, analyze the morphological characteristics and amplitude coordination relationship of the path in each interval, classify and summarize the trajectory range during the execution of each motion segment, and obtain the knee joint execution trajectory interval. For each mapping relationship, the recorded motor motion path and joint angle execution curve are compared temporally. First, the start and end times of the two are checked on the time axis to see if there is an overlap. If the start time of a certain motor motion is within 0.1 seconds before the start of the joint motion and the end time lags by no more than 0.2 seconds, the two are considered to have temporal overlap, and this time period is recorded as the path overlap segment. Then, the motor output angle and the joint sensor feedback angle at each time point within the overlap time period are extracted for difference analysis. The trend of each difference change curve is compared with its corresponding motion trend to determine whether it is a continuous rise, fall, or oscillation. If the trend type is consistent, the path shape of that segment is recorded. The system is designed to ensure consistency between the motor's execution path amplitude and the joint's execution amplitude. Simultaneously, the maximum and minimum angle values ​​within the specified time period are statistically analyzed, and the angle amplitude matching error is calculated. If the error is below a set amplitude error threshold (e.g., 5 degrees), the system considers the motor's execution path amplitude and the joint's execution amplitude to be highly consistent. Time periods that overlap in time, maintain a consistent morphological trend, and have amplitude errors within the threshold are recorded as execution trajectory intervals. For example, in a test sample, within a time period from 3.6 seconds to 4.1 seconds, the motor angle change amplitude is 25 degrees, the joint sensor angle amplitude is 26 degrees, the error is 1 degree, and the waveform type is a single-peak upward trend with the same change trend. This data segment is identified as a trajectory-consistent segment. All similar segments are combined to form the knee joint execution trajectory interval.

[0046] like Figure 6 As shown, the specific steps for obtaining the closed-loop action offset index are as follows: S501: Based on the knee joint execution trajectory range, compare it with the target motion state fed back by the knee joint angle sensor, analyze the spatial correspondence between the position of the feedback motion and the end point of the motor trajectory range, determine the offset between each group of positions, and obtain the motion space offset parameter group. The motor angle output value corresponding to the termination time point of each trajectory segment is extracted and used as the endpoint position of that trajectory segment. Simultaneously, the target motion state, i.e., the actual termination angle value of the knee joint after the execution of that motion segment, is extracted from the feedback data collected by the knee joint angle sensor. A one-to-one correspondence is established between the feedback angle data and the endpoint of the motor motion trajectory. The difference between the motor endpoint angle value and the feedback angle value is calculated, and the spatial offset value for each corresponding position is obtained by directly subtracting the two values. The offset value is marked as positive or negative, and the marking result is used to identify the offset direction. For example, if the motor endpoint angle is 62 degrees and the feedback angle is 58 degrees, the offset value is +4 degrees, indicating that the motor output is too large. Next, each set of offset values ​​is used to form a spatial offset data set. All offset values ​​in this set are statistically analyzed and classified into offset level ranges. For example, an offset value less than or equal to 2 degrees is a slight offset, 2 to 5 degrees is a moderate offset, and greater than 5 degrees is a severe offset. If the absolute value of the offset value exceeds a set offset threshold, such as 3 degrees, then the data set is marked as having an offset. Then, the sample numbers of all samples with offsets are summarized to form a motion spatial offset parameter set. For example, if 8 execution segments are recorded in a training cycle, and the offsets of the feedback endpoint angle and the motor trajectory endpoint in 3 of these segments are +4.5 degrees, -3.8 degrees, and +5.2 degrees, respectively, then these three segments constitute an offset parameter set for the next step of compensation control analysis.

[0047] S502: Based on the motion space offset parameter group, determine the offset phenomenon between the feedback position of each group and the end point of the motor trajectory interval, analyze the distribution of the offset segment, adjust the motor output motion sequence, correct the joint motion for the detected offset motion segment, and obtain the offset compensation execution sequence. The differences in direction and magnitude between the feedback angle value and the trajectory endpoint value in each group are analyzed one by one. First, segments in each group of data where the difference between the motor trajectory endpoint value and the sensor feedback value exceeds the offset threshold are screened out, and their corresponding action segment numbers are recorded. Then, the screened offset segments are arranged in chronological order, and their time positions in the overall training cycle are checked to determine whether the offsets are concentrated in specific time periods, such as the majority of offset segments being concentrated in the early or late stages of the action cycle. If there is a clustering of offset segments, the segment is treated as a whole for output sequence correction. Then, based on the original motor output angle, the offset correction value is added or subtracted according to the offset direction. For example, if the offset value is... If the offset is -4 degrees, the target angle of the motor will be increased by 4 degrees within the action sequence. The adjusted action sequence will be rhythmically corrected to ensure that the corrected action does not produce abrupt changes. The rhythm correction process is completed by keeping the difference in angular velocity between adjacent output points less than 5 degrees / second. The motor output sequence after the offset points are corrected is recorded as the compensated sequence. If the original motor outputs an angle of 40 degrees to 65 degrees between 2.0 seconds and 2.8 seconds, the feedback angle is 61 degrees, and the offset is -4 degrees, then the new compensation sequence will be adjusted to 44 degrees to 69 degrees within the same time period, and the adjusted value will be re-output. All such sequences are summarized to form the offset compensation execution sequence.

[0048] S503: Based on the offset compensation execution sequence, monitor the performance in continuous action feedback, judge the offset changes of each action sequence after compensation, analyze the continuity and stability of the corrected action, summarize the parameters adjusted for each action, and obtain the closed-loop action offset index. For each consecutive set of motor actions and feedback angles, the endpoint of the feedback angle after each compensation is compared with the offset value before compensation to determine whether the offset value has decreased after compensation. If the absolute value of the offset value after compensation is less than the value before compensation, the compensation is marked as effective. Simultaneously, the error value between the feedback angle and the target angle after each set of compensation is recorded and arranged in chronological order. Analysis is performed to determine if the error fluctuation is within ±2 degrees for three or more consecutive sets of actions. If this condition is met, the action feedback is marked as stable. All compensation execution processes are summarized, and compensation is extracted. The difference in the change of the offset value before and after, the correction amount applied for compensation, the execution time range and the final angle difference are used to construct a record entry for each set of compensation parameters and archive them according to the action number to form a set of action adjustment parameters. For example, in a set of data, the first offset is +4.5 degrees, the correction amount is set to -4 degrees, and the offset after compensation is +0.3 degrees. The second compensation is -3.8 degrees, and after correction to +4 degrees, the offset is -0.2 degrees. The error fluctuation range of the three consecutive sets does not exceed ±1 degree, and it is determined that the feedback action correction is stable. All action correction performance and parameter sets are integrated to form a closed-loop action offset index.

[0049] like Figure 7As shown, this embodiment of the invention also provides a brain-computer interface exoskeleton assistive control system for knee joint orthopedic rehabilitation, comprising: The EEG action discrimination module, based on the EEG electrode cap and knee joint angle sensor, analyzes the correspondence between the changes in the main peak of EEG and the joint action recording, and simultaneously compares the change periods of the two types of signals to determine whether the main peak fluctuation occurs at the same time as the joint action, thereby obtaining EEG collaborative discrimination features. The gait synchronization recognition module, based on EEG co-discrimination features, monitors the synchronization trend of joint movement changes and plantar pressure distribution changes during walking, compares the direction of trajectory change with the direction of force change, identifies continuous movement intervals, and obtains gait behavior coordination parameters. The intention trend fusion module, based on gait behavior coordination parameters, judges the consistency between EEG signal intention characteristics and knee joint movement trends, analyzes the direction of data change within the same time period, corrects the exoskeleton execution state, and obtains the motor command fusion factor. The trajectory generation module, based on the motion command fusion factor, initiates the execution of the exoskeleton knee joint, compares the consistency between the motor running direction and the joint movement direction, records the change process of the movement trajectory over time, and obtains the knee joint execution trajectory range. The closed-loop compensation module, based on the knee joint execution trajectory range, compares the action state fed back by the knee joint angle sensor with the target action, analyzes the difference trend between the feedback position and the action endpoint, compensates and adjusts the offset action, and obtains the closed-loop action offset index.

[0050] For ease of explanation, Figure 7 Only the main components of the system are shown. The system of this embodiment can be used to perform... Figure 1 The technical solutions of the method embodiments shown are similar in principle and in effect, and will not be described again here.

[0051] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A brain-computer interface exoskeleton-assisted control method for knee joint orthopedic rehabilitation, characterized in that, The method includes: S1: Based on the EEG electrode cap and knee joint angle sensor, analyze the correspondence between the EEG main peak change and the joint movement record, and simultaneously compare the change time periods of the two types of signals to determine whether the main peak fluctuation occurs at the same time as the joint movement, and obtain the EEG co-discrimination feature. S2: Based on the aforementioned EEG co-discrimination features, monitor the synchronous trend of joint movement changes and foot pressure distribution changes during walking, compare the direction of trajectory change with the direction of force change, identify continuous movement intervals, and obtain gait behavior co-parameters. S3: Based on the gait behavior coordination parameters, determine the consistency between the EEG signal intention characteristics and the knee joint movement trend, analyze the direction of data change within the same time period, correct the exoskeleton execution state, and obtain the motion command fusion factor. S4: Based on the motion command fusion factor, start the exoskeleton knee joint execution, compare the consistency between the motor running direction and the joint movement direction, record the change process of the movement trajectory over time, and obtain the knee joint execution trajectory range. S5: Based on the knee joint execution trajectory range, compare the action state fed back by the knee joint angle sensor with the target action, analyze the difference trend between the feedback position and the action endpoint, compensate and adjust the offset action, and obtain the closed-loop action offset index.

2. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The EEG co-discrimination features include peak response features, synchronous association features, and temporal coordination features; the gait behavior coordination parameters include stride length features, stride frequency features, and pressure distribution features; the motor command fusion factors include action synchronization features, trend consistency features, and command coordination features; the knee joint execution trajectory range includes action continuity features, trajectory morphology features, and response amplitude features; and the closed-loop action offset index includes action offset features, compensation correction features, and continuous stability features.

3. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The specific steps for obtaining the EEG collaborative discrimination features are as follows: S101: Based on the EEG electrode cap and knee joint angle sensor, analyze EEG signals and joint movement data, correspond the main peak time point of the EEG signal motion-related frequency band with the change time of knee joint movement, judge the synchronization performance of the data on the time axis, and obtain the brain-motor synchronization event group. S102: Based on the brain-motor synchronization event group, compare the direction of change of the main peak of brain electrical activity and the direction of change of knee joint movement, determine whether the trends of the two are synchronized, identify data combinations with consistent trends, and obtain the neuro-joint trend cluster. S103: Based on the aforementioned neural joint trend cluster, aggregate EEG fluctuation temporal features, action response rhythm features, and data overlap intervals, classify and integrate the data according to the features, and obtain EEG collaborative discrimination features.

4. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The specific steps for obtaining the gait behavior coordination parameters are as follows: S201: Based on the EEG collaborative discrimination features, analyze the data sequence output by the knee joint angle sensor, group the changing direction of continuous angle rotation segments, determine the start and end positions and the order of direction changes of each rotation segment, identify segments with changing characteristics within the walking cycle, aggregate the rotation cycle according to the change sequence, and obtain a joint gait structure segment group. S202: Based on the joint gait structure segment group, compare it with the data collected by the plantar pressure sensor, analyze the movement trajectory of the force center of each section of the plantar pressure, determine the correspondence between the trajectory change direction and the knee joint rotation direction, identify the synchronously changing data segments, and obtain the segment set corresponding to the gait pressure joint. S203: Based on the segment set corresponding to the step pressure joint, determine the segments with consistent trends, count the time overlap interval between the knee joint movement state and the foot force state, classify the state switching process, identify the gait stage according to the joint action state type, and obtain the gait behavior coordination parameters.

5. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The specific steps for obtaining the motion command fusion factor are as follows: S301: Based on the gait behavior coordination parameters, analyze the EEG electrode cap signal and the motion data of the knee joint angle sensor in the corresponding time period, compare the direction of change of the data in each gait cycle, determine the correspondence between the change pattern of the neural signal and the rotation trend of the joint movement, count the segments with consistent change trends, and obtain the brain-joint trend coordination sequence. S302: Based on the brain joint trend coordination sequence, determine the direction of exoskeleton execution unit movement corresponding to each trend consistency interval, compare the synchronization relationship between the exoskeleton unit movement and the knee joint angle change trend, analyze the correspondence between signal direction and movement direction in each segment, identify synchronization segments, and obtain execution trend synchronization patch group. S303: Based on the execution trend synchronization patch group, calculate the coordination of EEG signal trend, knee joint rotation trend and exoskeleton execution direction in the same time interval, analyze the synchronization and direction consistency of each type of data, and obtain the motion command fusion factor.

6. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The specific steps for obtaining the knee joint execution trajectory range are as follows: S401: Based on the motion command fusion factor, analyze the action sequence of the motor drive device, determine whether the action direction required by the action command matches the action direction output by the drive device, identify the action segment with the same direction, and obtain the action direction matching segment. S402: Based on the motion direction matching segment, compare the motion process executed by the exoskeleton knee joint, monitor the correspondence between the output motion segment of each motor and the joint motion direction, determine whether the change process of the motion path shows a consistent trend with the motion executed by the knee joint, and obtain the joint path collaborative mapping set. S403: Based on the joint path collaborative mapping set, determine the time overlap interval between the motor-driven motion path and the knee joint motion sequence, analyze the morphological characteristics and amplitude coordination relationship of the path in each interval, classify and summarize the trajectory range during the execution of each motion segment, and obtain the knee joint execution trajectory interval.

7. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The specific steps for obtaining the closed-loop action offset index are as follows: S501: Based on the knee joint execution trajectory range, compare it with the target action state fed back by the knee joint angle sensor, analyze the spatial correspondence between the position of the feedback action and the end point of the motor trajectory range, determine the offset between each group of positions, and obtain the action space offset parameter group. S502: Based on the motion space offset parameter group, determine the offset phenomenon between the feedback position of each group and the end point of the motor trajectory interval, analyze the distribution of the offset segment, adjust the motor output motion sequence, correct the joint motion for the detected offset motion segment, and obtain the offset compensation execution sequence. S503: Based on the offset compensation execution sequence, monitor the performance in continuous action feedback, determine the offset change of each action sequence after compensation, analyze the continuity and stability of the corrected action, summarize the parameters adjusted for each action, and obtain the closed-loop action offset index.

8. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The main peak change refers to the amplitude point that appears in the motion-related frequency band of the EEG waveform, which is a characteristic signal of motion intention. The change time period refers to the time point when the main peak change of the EEG signal and the knee joint angle change occur respectively.

9. The method for controlling a brain-computer interface exoskeleton for knee joint orthopedic rehabilitation according to claim 1, characterized in that, The intention features refer to the EEG features collected by the EEG electrode cap that reflect the patient's active motor imagination or intention, and the joint movement trends refer to the direction of movement shown in the knee joint angle data, including the trend of continuous bending or extension.

10. A brain-computer interface exoskeleton assistive control system for knee joint orthopedic rehabilitation, the system being used to implement the brain-computer interface exoskeleton assistive control method for knee joint orthopedic rehabilitation as described in any one of claims 1-9, characterized in that, The system includes: The EEG action discrimination module, based on the EEG electrode cap and knee joint angle sensor, analyzes the correspondence between the changes in the main peak of EEG and the joint action recording, and simultaneously compares the change periods of the two types of signals to determine whether the main peak fluctuation occurs at the same time as the joint action, thereby obtaining EEG collaborative discrimination features. The gait synchronization recognition module, based on the aforementioned EEG collaborative discrimination features, monitors the synchronization trend of joint movement changes and foot pressure distribution changes during walking, compares the trajectory change direction with the force change direction, identifies continuous movement intervals, and obtains gait behavior coordination parameters. The intention trend fusion module, based on the gait behavior coordination parameters, determines the consistency between the intention features of the EEG signal and the trend of knee joint movement, analyzes the direction of data change within the same time period, corrects the exoskeleton execution state, and obtains the motion command fusion factor. The trajectory generation module, based on the motion command fusion factor, initiates the execution of the exoskeleton knee joint, compares the consistency between the motor running direction and the joint movement direction, records the change process of the movement trajectory over time, and obtains the knee joint execution trajectory range. The closed-loop compensation module, based on the knee joint execution trajectory range, compares the action state fed back by the knee joint angle sensor with the target action, analyzes the difference trend between the feedback position and the action endpoint, compensates and adjusts the offset action, and obtains the closed-loop action offset index.

Images