Hemiplegic gait feature parameter extraction method, gait measuring device and electronic equipment

Abstract

The application relates to the field of rehabilitation medicine engineering and provides a hemiplegic gait feature parameter extraction method, a gait measuring device and electronic equipment. The method comprises the following steps: acquiring body movement data of a subject, wherein the body movement data comprises acceleration data and angular velocity data of lower limbs and real-time pressure data of soles; based on the real-time pressure data, determining that the subject has a preset gait event, cross-checking the gait event based on the angular velocity data, and extracting time parameters of the gait of the subject; based on the acceleration data and the angular velocity data, calculating the real-time posture of the subject; calculating the relative flexion and extension angles of the joints of the subject, and extracting the spatial parameters of the gait of the subject. The method is used to solve the defects of the existing single mode sensor in the hemiplegic gait analysis, such as the sensing blind area, integral drift and poor adaptability to pathological gait, the scheme of the application can fuse multi-source sensing information, and adapt to abnormal fluctuations of pathological gait.

Description

Technical Field

[0001] This invention relates to the field of rehabilitation medicine engineering technology, and in particular to a method for extracting gait characteristic parameters of hemiplegic patients, as well as a gait measurement device and electronic equipment. Background Technology

[0002] Patients with hemiplegia due to stroke often exhibit typical kinematic abnormalities such as foot drop, foot inversion, knee hyperextension, and circumduction gait due to damage to the central nervous system. Gait analysis can reveal the key aspects and influencing factors of gait abnormalities and is an important tool for guiding rehabilitation assessment and treatment.

[0003] Existing gait feature parameter extraction methods mainly suffer from the following technical shortcomings: (1) Although the gait detection system based on video images has high measurement accuracy, the equipment is expensive and has strict environmental requirements, making it difficult to promote its application in home environments. (2) Wearable sensors based on a single mode have obvious sensing blind spots: foot pressure detection is out of range during the swing phase and cannot capture the foot contour trajectory; the inertial measurement unit (IMU) faces numerical accumulation error caused by the zero bias drift of the microelectromechanical system, and the spatial displacement diverges severely after long-term integration. (3) Existing feature extraction algorithms for hemiplegic gait are mostly designed based on physiological models of healthy people. They are poorly adapted to pathological compensatory movements such as foot drop, foot inversion, and dragging steps that are unique to hemiplegic patients, and are prone to false detection or missed detection of gait events. Summary of the Invention

[0004] This invention provides a method for extracting hemiplegic gait feature parameters, as well as a gait measurement device and electronic equipment, to solve the defects of existing single-modal sensors in hemiplegic gait analysis, such as perception blind spots, integral drift, and poor adaptability to pathological gait. The solution of this application can integrate multi-source sensor information and adapt to abnormal fluctuations in pathological gait.

[0005] This invention provides a method for extracting gait feature parameters in hemiplegic patients, comprising: Acquire the subject's body motion data, which includes acceleration and angular velocity data of the lower limbs, as well as real-time pressure data of the soles of the feet; Based on the real-time pressure data, a pre-set gait event is determined to have occurred in the subject. The gait event is then cross-validated based on the angular velocity data, and the temporal parameters of the subject's gait are extracted. The subject's real-time attitude is calculated based on the acceleration data and the angular velocity data; Based on the posture data, the relative flexion and extension angles of the subject's joints are calculated, and the spatial parameters of the subject's gait are extracted. The spatial parameters include at least one or more of the following: stride length, gait speed, ground contact foot inclination angle, ground takeoff foot inclination angle, and plantar pressure center trajectory.

[0006] According to the method for extracting hemiplegic gait feature parameters provided by the present invention, the real-time pressure data includes a set of pressure data acquired by several pressure sensors; The determination of a pre-defined gait event in the subject based on the real-time pressure data includes: Based on the pressure data changes acquired by each pressure sensor in the pressure dataset, it is determined that the subject has experienced a pre-set gait event.

[0007] According to the hemiplegic gait feature parameter extraction method provided by the present invention, the step of calculating the real-time posture of the subject based on the acceleration data and the angular velocity data includes: The real-time attitude quaternion of the subject's lower limbs was calculated using a strapdown inertial navigation algorithm combined with a Kalman filter algorithm. The angular velocity data collected in the sensor coordinate system is projected to the geographic coordinate system using the direction cosine matrix corresponding to the real-time attitude quaternion.

[0008] The method for extracting hemiplegic gait feature parameters according to the present invention further includes: Obtain heading data of the subject's lumbar center of gravity region; Using the heading data as a global reference, a waist reference frame is constructed; The real-time yaw angles of each limb of the subject are mapped to the waist reference frame, and the yaw angles of the acceleration data and angular velocity data are calibrated collaboratively.

[0009] According to the hemiplegic gait feature parameter extraction method provided by the present invention, the step of calculating the real-time posture of the subject based on the acceleration data and the angular velocity data further includes: When the subject's gait meets the set requirements, the velocity vector of the subject's foot in the geographic coordinate system is reset to zero based on the time parameter.

[0010] According to the method for extracting hemiplegic gait feature parameters provided by the present invention, after acquiring the subject's body movement data, the method further includes: The acceleration and angular velocity data are subjected to static zero-bias correction and filtered using a second-order Butterworth low-pass filter. The real-time pressure data is converted into physical pressure values ​​using a conductivity-pressure linear transformation model.

[0011] The present invention also provides a hemiplegic gait measurement device, comprising: The central control unit is used to execute the method for extracting hemiplegic gait feature parameters; A distributed inertial measurement unit is used to acquire acceleration and angular velocity data of the subject's lower limbs, as well as heading data of the lumbar center of mass region; A distributed pressure sensor array is used to acquire real-time pressure data of the soles of the subjects' feet.

[0012] According to the hemiplegic gait measurement device provided by the present invention, the distributed inertial measurement unit includes at least seven nodes, wherein the first node is deployed in the lumbar center of mass region of the subject, the second node is deployed in the lateral aspect of the left thigh of the subject, the third node is deployed in the lateral aspect of the right thigh of the subject, the fourth node is deployed in the anterior aspect of the tibia of the left lower leg of the subject, the fifth node is deployed in the anterior aspect of the tibia of the right lower leg of the subject, the sixth node is deployed in the dorsum of the left foot of the subject, and the seventh node is deployed in the dorsum of the right foot of the subject.

[0013] According to the hemiplegic gait measurement device provided by the present invention, the distributed pressure sensor array includes at least five pressure sensors, wherein the first pressure sensor is deployed in the heel region of the subject, the second pressure sensor is deployed in the base of the fifth metatarsal bone of the subject, the third pressure sensor is deployed in the head region of the fifth metatarsal bone of the subject, the fourth pressure sensor is deployed in the head region of the first metatarsal bone of the subject, and the fifth pressure sensor is deployed in the big toe region of the subject.

[0014] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements any of the above-described methods for extracting hemiplegic gait feature parameters.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for extracting hemiplegic gait feature parameters.

[0016] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-described methods for extracting hemiplegic gait feature parameters.

[0017] The hemiplegic gait feature parameter extraction method provided by this invention integrates lower limb inertial sensor (acceleration + angular velocity) and plantar pressure distribution data to construct a multi-source information cross-verification mechanism, which significantly improves the reliability of gait event detection and temporal parameter extraction. At the same time, based on posture calculation and joint angle analysis, it realizes the comprehensive quantification of spatial parameters such as stride length, gait speed, foot tilt angle and pressure center trajectory. It can simultaneously acquire high-precision temporal-spatial multidimensional gait features in natural walking state without relying on complex vision systems or fixed walkways, providing an objective, portable and quantifiable technical means for the rehabilitation assessment of hemiplegic patients. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the method for extracting hemiplegic gait feature parameters provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the spatial deployment of the distributed inertial measurement unit provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the spatial deployment of the pressure sensor array provided in an embodiment of the present invention; Figure 4 This is a flowchart illustrating the time parameter extraction method provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the spatial pose calculation process provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the physical structure of the electronic device provided in an embodiment of the present invention.

[0020] in: 1 - First node; 2 - Second node; 3 - Third node; 4 - Fourth node; 5 - Fifth node; 6 - Sixth node; 7 - Seventh node; 8 - First pressure sensor; 9 - Second pressure sensor; 10 - Third pressure sensor; 11 - Fourth pressure sensor; 12 - Fifth pressure sensor. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0022] Figure 1 This is a flowchart illustrating the method for extracting hemiplegic gait feature parameters provided in an embodiment of the present invention.

[0023] like Figure 1 As shown, this embodiment provides a method for extracting hemiplegic gait feature parameters, including: Step 101: Obtain the subject's body motion data, which includes acceleration and angular velocity data of the lower limbs, as well as real-time pressure data of the soles of the feet; Step 102: Based on the real-time pressure data, determine that the subject has experienced a pre-set gait event, perform cross-verification of the gait event based on the angular velocity data, and extract the subject's gait time parameters; In practical applications, the gait events provided in this embodiment may include heel strike, full foot flat, heel lift-off, and toe lift-off.

[0024] In practical applications, the sagittal angular velocity signal of the inertial measurement unit of the foot can be collected synchronously. The trough before the peak point of the angular velocity is logically aligned with the moment when the total force of the pressure array returns to zero to confirm the toe-off event; the zero-crossing point after the peak point of the angular velocity is logically aligned with the moment of pressure jump to confirm the heel-striking event.

[0025] Step 103: Calculate the real-time attitude of the subject based on the acceleration data and the angular velocity data; Step 104: Based on the posture data, calculate the relative flexion and extension angles of the subject's joints, and extract the spatial parameters of the subject's gait. The spatial parameters include at least one or more of the following: stride length, gait speed, ground contact foot inclination angle, ground lift foot inclination angle, and plantar pressure center trajectory.

[0026] In practical applications, the above feature parameter extraction method can be implemented by the central control unit of the hemiplegic gait measurement device. The hemiplegic gait measurement device can also include a distributed inertial measurement unit and a distributed pressure sensor array. The distributed inertial measurement unit is used to acquire the acceleration and angular velocity data of the subject's lower limbs, as well as the heading data of the lumbar center of mass region. The distributed pressure sensor array is used to acquire the real-time pressure data of the subject's soles.

[0027] Figure 2This is a schematic diagram of the spatial deployment of the distributed inertial measurement unit provided in an embodiment of the present invention.

[0028] Figure 3 This is a schematic diagram of the spatial deployment of the pressure sensor array provided in an embodiment of the present invention.

[0029] like Figure 2 As shown, the distributed inertial measurement unit includes at least seven nodes, wherein the first node 1 is deployed in the lumbar center of mass region of the subject, the second node 2 is deployed in the lateral aspect of the left thigh of the subject (midpoint of the vastus lateralis muscle), the third node 3 is deployed in the lateral aspect of the right thigh of the subject, the fourth node 4 is deployed in the anterior aspect of the tibia of the left lower leg of the subject (below the tibial tuberosity), the fifth node 5 is deployed in the anterior aspect of the tibia of the right lower leg of the subject, the sixth node 6 is deployed in the dorsum of the left foot of the subject (midpoint of the metatarsal region), and the seventh node 7 is deployed in the dorsum of the right foot of the subject.

[0030] In practical applications, the seven nodes can be WT601WIFI nine-axis attitude sensors, and the sampling frequency of each sensor can be set to 100Hz, with the measurement range being acceleration. 8g, angular velocity 2000 ° / s.

[0031] like Figure 3 As shown, the distributed pressure sensor array includes at least five pressure sensors, wherein the first pressure sensor 8 is deployed in the heel region of the subject, the second pressure sensor 9 is deployed in the base of the fifth metatarsal bone of the subject, the third pressure sensor 10 is deployed in the head region of the fifth metatarsal bone of the subject, the fourth pressure sensor 11 is deployed in the head region of the first metatarsal bone of the subject, and the fifth pressure sensor 12 is deployed in the big toe region of the subject.

[0032] In practical applications, the pressure sensor on the sole of the foot can be an MD30-60 flexible piezoresistive thin-film pressure sensor. The data collected by each pressure sensor is connected to the central control unit through an analog-to-digital conversion module, and the pressure data can be uploaded to the central control unit in real time via WIFI wireless communication.

[0033] All sensor nodes of the above distributed inertial measurement unit and distributed pressure sensor array can build a star-shaped wireless communication network with the host computer through the UDP protocol. The host computer realizes the time synchronization and aggregation of multi-source heterogeneous data by using a multi-threaded dual-port concurrent listening method.

[0034] In the exemplary embodiment, before collecting the subject's body motion data, the subject remains in a static standing posture, and the system executes a static calibration procedure: each posture sensor node collects 2000 sets of static data to calculate the gyroscope zero bias value and stores it as compensation parameters; each pressure sensor collects the initial mean and standard deviation under no-load conditions as the initial benchmark for the adaptive threshold algorithm.

[0035] After acquiring the subjects' body movement data, the following steps are also included: The acceleration and angular velocity data are subjected to static zero-bias correction and filtered using a second-order Butterworth low-pass filter. The real-time pressure data is converted into physical pressure values ​​using a conductivity-pressure linear transformation model, conforming to the following formula: in, This is the physical pressure value obtained after conversion. This represents the real-time conductance value of the i-th channel. The converted pressure signal is processed by a low-pass filter with a cutoff frequency of 5Hz to filter out high-frequency vibrations and transient noise.

[0036] Figure 4 This is a flowchart illustrating the time parameter extraction method provided in this embodiment of the invention.

[0037] like Figure 4 As shown, it may include the following steps: Step 401: During the subject's walking process, the data segment identified as the swing phase in the previous gait cycle can be extracted through a sliding window. The mean and standard deviation of the pressure in each channel are dynamically updated, and then the activation threshold for the next cycle is calculated, satisfying the following formula: in, The average pressure. With a pressure standard deviation and a safety factor K of 3, it can cover 66.7% of noise fluctuations.

[0038] Step 402, define the binarized state of each pressure sensor. It conforms to the following formula: Step 403: Determine that the subject has experienced a pre-set gait event. The determination logic is as follows: Heel strike (HS) detection: For patients with foot drop, OR logic compensation is added to the second pressure sensor (S2) and the third pressure sensor (S3) at the lateral node of the foot. When the first pressure sensor (S1) is not triggered but either state of S2 or S3 changes from 0 to 1, it is determined that the foot is in initial contact with the ground and is marked as "inversion compensation ground contact mode". Full foot flat (FF) detection: When the fourth pressure sensor (S4) and the fifth pressure sensor (S5) of the hind foot area node (S1, S2) and the forefoot area node are simultaneously in a pressure-activated state, it is determined to be a full foot flat. Heel lift-off (HO) detection: When the pressure on the heel is released (S1 state is zero) while the forefoot remains under pressure (S4 and / or S5 state is 1), it is determined that the heel is lifted off the ground. Toe-off (TO) detection: When all five sensors on the foot return to zero, it is determined that the foot has completely left the ground.

[0039] Step 404: Perform cross-verification of gait events based on angular velocity data. Specifically, the sagittal plane angular velocity signal from the dorsum of the foot attitude sensor can be acquired simultaneously. An adaptive peak detection method is used to identify local peaks and zero crossings in the angular velocity waveform.

[0040] Toe-off verification: TO corresponds to the first local trough before the local peak of angular velocity. This kinematic inflection point is logically aligned with the moment when the total force on the pressure array returns to zero, thus accurately locking the start of the swing phase. If the time difference between the two exceeds a preset threshold, it is judged as a "drag artifact" and is not counted. Heel strike verification: Hs is the first signal zero-crossing point after the local peak point of angular velocity. If the system detects pressure triggering but the angular velocity is still in the high dynamic rotation range (amplitude greater than the preset motion threshold), it is determined to be a pressure artifact caused by pathological dragging and is not included in the support phase.

[0041] Step 405, Gait Time Parameter Extraction. In practical applications, extracting the subject's gait time parameters means, assuming that the times corresponding to HS, FF, HO, and TO in the k-th gait cycle are respectively... , , as well as Then the following event parameters can be extracted: Support duration in the early stages: Support for the medium term: Support duration at the end of the period: Swing period duration: Gait cycle: Step frequency: After the above steps, the complete extraction of gait time parameters from the original pressure and angular velocity signals was completed, providing basic data for subsequent gait period segmentation and rhythm analysis.

[0042] In an exemplary embodiment, calculating the subject's real-time attitude based on the acceleration data and the angular velocity data includes: The real-time attitude quaternion of the subject's lower limbs was calculated using a strapdown inertial navigation algorithm combined with a Kalman filter algorithm. The angular velocity data collected in the sensor coordinate system is projected to the geographic coordinate system using the direction cosine matrix corresponding to the real-time attitude quaternion.

[0043] Figure 5 This is a schematic diagram of the spatial pose calculation process provided in an embodiment of the present invention.

[0044] like Figure 5 As shown, in practical applications, spatial pose calculation may include the following steps: Step 501, Attitude Calculation. In practical applications, quaternions can be used to represent sensor attitude. Let the geographic coordinate system be G (East-North-Sky) and the sensor coordinate system be S. By fusing accelerometer, gyroscope, and magnetometer data through Kalman filtering, the attitude quaternions of each sensor relative to the geographic coordinate system are recursively solved. .

[0045] Step 502, acceleration coordinate transformation and gravity compensation.

[0046] Specifically, the direction cosine matrix corresponding to the attitude quaternion can be used to project the acceleration data collected in the sensor coordinate system to the geographic coordinate system, satisfying the following formula: in, Let be the direction cosine matrix corresponding to the attitude quaternion. For acceleration data, Let T be the gravitational acceleration vector [0, 0, -6.8] in the geographic coordinate system.

[0047] Step 503, zero-speed correction.

[0048] When the subject is in the support phase from foot flat (FF) to heel off the ground (HO), the physical constraint of foot contact with the ground is used to reset the foot's velocity vector in the geographic coordinate system to zero, satisfying the following formula: This zero-speed correction step can be performed once per gait cycle, thus effectively cutting off the drift error accumulated in the previous oscillation cycle.

[0049] Step 504, Displacement Integral Specifically, the trapezoidal integral method can be used to calculate the instantaneous spatial displacement of the foot, satisfying the following formula: Step 505: Global heading coordination correction for the waist area.

[0050] Specifically, the heading of the waist reference node can be used as a global reference to coordinately calibrate the yaw angles of other limb nodes. Let each limb node be relative to its initial coordinate system. The relative attitude quaternions satisfy the following formula: in, The quaternion represents the installation bias between the sensor and the limb link.

[0051] Then, the real-time relative yaw angles of the waist node and limb node i can be extracted to construct a heading correction operator. Thus, the corrected body posture satisfies the following formula: This embodiment replaces the absolute yaw information that diverges over time with a stable waist heading, effectively reducing the impact of geomagnetic interference on the yaw angles of each node of the lower limbs and ensuring the physical consistency of the horizontal projection when calculating the stride length.

[0052] In the exemplary embodiment, after the waist heading cooperative correction of the above embodiments, the posture quaternion of the subject's proximal limbs (such as the thigh) can be: The posture quaternion of the distal limb (such as the lower leg) can be... According to the chain rule of quaternion transformations, the relative rotation quaternion between two adjacent limbs can satisfy the following formula: in, For the conjugate of the proximal limb posture quaternion, This is a quaternion multiplication. The core advantage of this operation is that the quaternions of both limbs involved in the operation have been pre-mapped to a unified lumbar heading reference frame, and the relative quaternions obtained have eliminated the influence of overall turning and center of gravity swaying on local joint angles during the subject's walking path.

[0053] The relative quaternion can then be converted into a direction cosine matrix. It satisfies the following formula: Since the physiological movements of the subjects' lower limb joints are mainly flexion and extension reciprocating motions in the sagittal plane, only the rotational component around the joint's transverse axis can be extracted. Specifically: Hip flexion-extension angle: calculated from the relative posture matrix of the pelvis and thigh, satisfying the following formula: ; Knee flexion-extension angle: calculated from the relative posture matrix of the thigh and lower leg, satisfying the following formula: ; Ankle flexion angle: calculated from the relative posture matrix of the lower leg and foot, satisfying the following formula: .

[0054] Wherein, c32 and c33 are the elements of the third row and second column and the third row and third column of the direction cosine matrix, respectively. Through the above steps, the extraction of all gait space parameters from the corrected posture quaternion and plantar pressure data is completed, including stride length, gait speed, ground contact foot inclination angle, ground takeoff foot inclination angle, plantar pressure center trajectory, and sagittal angles of the hip, knee, and ankle joints.

[0055] In summary, the hemiplegic gait feature parameter extraction method provided in this application has at least the following beneficial effects: (1) By integrating plantar pressure sensors and distributed inertial measurement units, the perception blind spots of single-modal sensors are overcome, and full-dimensional information acquisition of hemiplegic gait from dynamics to kinematics is realized. (2) For typical pathological signs of hemiplegia such as foot drop and foot inversion, a five-point specific pressure sensor layout and an adaptive state machine with logic compensation were designed. Combined with IMU angular velocity cross-verification, the detection accuracy of key gait events under pathological gait was significantly improved. (3) Through the dual strategy of zero-speed correction and waist global heading coordination correction, the integral drift and geomagnetic interference in the inertial navigation solution are effectively suppressed, ensuring the physical consistency of spatial trajectory and joint angle reconstruction. (4) The sagittal angle of the joint can be directly calculated by the quaternion chain transformation of adjacent limb segments without the need for complex inverse kinematics optimization. It has high computational efficiency and is suitable for embedded real-time processing.

[0056] The present invention also provides a verification experiment to verify the effectiveness of the solution provided in this application.

[0057] Experimental conditions: Six healthy subjects (4 males, 2 females, aged 22.5 ± 2.2 years) and one typical stroke hemiplegic patient were recruited. Subjects wore the hemiplegic gait measurement device provided in this application and walked back and forth in a straight line on a straight path at a comfortable speed of their own choosing.

[0058] Comparison benchmark: The OpenCap label-free visual motion capture system was used as the reference benchmark. The system outputs lower limb joint angle data through OpenSim inverse kinematics calculation.

[0059] Experimental results (time parameters): The extraction results of gait time parameters of healthy subjects showed that the average gait cycle was distributed between 1.05s and 1.17s, with the support phase accounting for about 60.6% and the swing phase accounting for about 36.1%, which was highly consistent with the clinical physiological norms.

[0060] The results of extracting time parameters from hemiplegic patients showed that the system identified the following typical pathological features: compensatory prolongation of the swing phase on the affected side. The swing phase duration on the affected side of the patient was 0.55s, which was significantly longer than that on the healthy side (0.43s). In contrast, the mean swing phase duration on the left and right sides of the healthy control group was 0.48s, indicating that the affected limb was forced to prolong the swing time in order to complete foot clearance. The healthy side exhibited compensatory prolongation in the mid-stance of gait, with the patient's healthy side showing a mid-stance duration of 0.56 s, significantly longer than that of the affected side (0.47 s). In contrast, the mean mid-stance durations for both the left and right sides in the healthy control group were 0.26-0.30 s. This indicates that the healthy side provides stable support for the swing of the affected side by prolonging the single-limb weight-bearing time, which is a typical gait compensation strategy. These results are consistent with the clinical gait manifestations of hemiplegic patients, validating the effectiveness of this invention in extracting gait time parameters.

[0061] Experimental Results (Spatial Parameters): Verification results of spatial parameter extraction accuracy in healthy subjects showed that the solution accuracy of the method of this invention and the OpenCap reference system for various joint angles was as follows: hip joint correlation coefficient R² = 0.653~0.676, root mean square error RMSE = 2.25~2.71°; knee joint correlation coefficient R² = 0.635~0.642, RMSE = 3.46~4.18°; ankle joint correlation coefficient R² = 0.848~0.871, RMSE = 2.48~3.22°. All of these errors are within clinically acceptable ranges, verifying the measurement accuracy of this invention under routine conditions.

[0062] The above results of spatial parameter extraction from hemiplegic patients show that the system captured the differences in joint angles between them and healthy controls throughout the complete gait cycle: at the ankle level, the affected side of the patient exhibited typical foot drop characteristics, with a peak dorsiflexion angle of only about 5° during the swing phase, while that of the healthy controls was about 15°, indicating significantly limited clearance ability; at the knee level, the affected side of the patient exhibited rigid gait characteristics, with a peak flexion angle of about 40° during the swing phase, significantly lower than that of the healthy controls (about 60°); at the hip level, the unaffected side of the patient exhibited compensatory hyperflexion, with a peak flexion angle of about 35° during the swing phase, while that of the healthy controls was about 30°, showing a typical hip levitation compensation pattern.

[0063] In summary, under normal gait conditions, the joint angles extracted by this invention show a high degree of consistency with the OpenCap gold standard system (hip and knee joint R2>0.63, RMSE<4.2°), verifying the measurement accuracy of this invention under normal conditions. Under hemiplegic pathological gait conditions, this invention successfully identified typical pathological features such as prolonged swing phase on the affected side, compensatory prolonged mid-stance phase on the healthy side, and loss of ankle dorsiflexion, decreased knee flexion, and compensatory excessive flexion of the hip on the healthy side. The effective extraction of the above pathological features is attributed to the multi-source sensor fusion architecture adopted by this invention. Specifically, the five-point plantar pressure layout and adaptive state machine enable accurate identification of abnormal landing patterns such as foot drop and foot inversion. Zero-velocity correction and global lumbar heading collaborative correction effectively suppress integral drift and geomagnetic interference in inertial navigation calculations. Quaternion chain transformation between adjacent limb segments achieves physically consistent calculation of joint angles. The experimental results show that the present invention overcomes the perception blind spot and integral drift problems of single modal sensors in hemiplegic gait analysis through multi-source sensor fusion and collaborative correction, and provides a reliable technical means for the accurate extraction of gait feature parameters of hemiplegic patients.

[0064] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute a hemiplegic gait feature parameter extraction method, which includes: Acquire the subject's body motion data, which includes acceleration and angular velocity data of the lower limbs, as well as real-time pressure data of the soles of the feet; Based on the real-time pressure data, a pre-set gait event is determined to have occurred in the subject. The gait event is then cross-validated based on the angular velocity data, and the temporal parameters of the subject's gait are extracted. The subject's real-time attitude is calculated based on the acceleration data and the angular velocity data; Based on the posture data, the relative flexion and extension angles of the subject's joints are calculated, and the spatial parameters of the subject's gait are extracted. The spatial parameters include at least one or more of the following: stride length, gait speed, ground contact foot inclination angle, ground takeoff foot inclination angle, and plantar pressure center trajectory.

[0065] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0066] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the hemiplegic gait feature parameter extraction method provided by the above methods, the method including: Acquire the subject's body motion data, which includes acceleration and angular velocity data of the lower limbs, as well as real-time pressure data of the soles of the feet; Based on the real-time pressure data, a pre-set gait event is determined to have occurred in the subject. The gait event is then cross-validated based on the angular velocity data, and the temporal parameters of the subject's gait are extracted. The subject's real-time attitude is calculated based on the acceleration data and the angular velocity data; Based on the posture data, the relative flexion and extension angles of the subject's joints are calculated, and the spatial parameters of the subject's gait are extracted. The spatial parameters include at least one or more of the following: stride length, gait speed, ground contact foot inclination angle, ground takeoff foot inclination angle, and plantar pressure center trajectory.

[0067] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the hemiplegic gait feature parameter extraction method provided by the above methods, the method comprising: Acquire the subject's body motion data, which includes acceleration and angular velocity data of the lower limbs, as well as real-time pressure data of the soles of the feet; Based on the real-time pressure data, a pre-set gait event is determined to have occurred in the subject. The gait event is then cross-validated based on the angular velocity data, and the temporal parameters of the subject's gait are extracted. The subject's real-time attitude is calculated based on the acceleration data and the angular velocity data; Based on the posture data, the relative flexion and extension angles of the subject's joints are calculated, and the spatial parameters of the subject's gait are extracted. The spatial parameters include at least one or more of the following: stride length, gait speed, ground contact foot inclination angle, ground takeoff foot inclination angle, and plantar pressure center trajectory.

[0068] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0069] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for extracting gait feature parameters in hemiplegic patients, characterized in that, include: Acquire the subject's body motion data, which includes acceleration and angular velocity data of the lower limbs, as well as real-time pressure data of the soles of the feet; Based on the real-time pressure data, a pre-set gait event is determined to have occurred in the subject. The gait event is then cross-validated based on the angular velocity data, and the temporal parameters of the subject's gait are extracted. The subject's real-time attitude is calculated based on the acceleration data and the angular velocity data; Based on the posture data, the relative flexion and extension angles of the subject's joints are calculated, and the spatial parameters of the subject's gait are extracted. The spatial parameters include at least one or more of the following: stride length, gait speed, ground contact foot inclination angle, ground takeoff foot inclination angle, and plantar pressure center trajectory.

2. The method for extracting hemiplegic gait feature parameters according to claim 1, characterized in that, The real-time pressure data includes a set of pressure data acquired by several pressure sensors; The determination of a pre-defined gait event in the subject based on the real-time pressure data includes: Based on the pressure data changes acquired by each pressure sensor in the pressure dataset, it is determined that the subject has experienced a pre-set gait event.

3. The method for extracting hemiplegic gait feature parameters according to claim 1, characterized in that, The calculation of the subject's real-time attitude based on the acceleration data and the angular velocity data includes: The real-time attitude quaternion of the subject's lower limbs was calculated using a strapdown inertial navigation algorithm combined with a Kalman filter algorithm. The angular velocity data collected in the sensor coordinate system is projected to the geographic coordinate system using the direction cosine matrix corresponding to the real-time attitude quaternion.

4. The method for extracting hemiplegic gait feature parameters according to claim 1, characterized in that, Also includes: Obtain heading data of the subject's lumbar center of gravity region; Using the heading data as a global reference, a waist reference frame is constructed; The real-time yaw angles of each limb of the subject are mapped to the waist reference frame, and the yaw angles of the acceleration data and angular velocity data are calibrated collaboratively.

5. The method for extracting hemiplegic gait feature parameters according to claim 1, characterized in that, The process of calculating the subject's real-time attitude based on the acceleration and angular velocity data further includes: When the subject's gait meets the set requirements, the velocity vector of the subject's foot in the geographic coordinate system is reset to zero based on the time parameter.

6. The method for extracting hemiplegic gait feature parameters according to claim 1, characterized in that, The process of acquiring the subject's body movement data then includes: The acceleration and angular velocity data are subjected to static zero-bias correction and filtered using a second-order Butterworth low-pass filter. The real-time pressure data is converted into physical pressure values ​​using a conductivity-pressure linear transformation model.

7. A hemiplegic gait measurement device, characterized in that, include: The central control unit is used to execute the hemiplegic gait feature parameter extraction method as described in any one of claims 1-6; A distributed inertial measurement unit is used to acquire acceleration and angular velocity data of the subject's lower limbs, as well as heading data of the lumbar center of mass region; A distributed pressure sensor array is used to acquire real-time pressure data of the soles of the subjects' feet.

8. The hemiplegic gait measurement device according to claim 7, characterized in that, The distributed inertial measurement unit includes at least seven nodes, wherein the first node is deployed in the lumbar center of mass region of the subject, the second node is deployed in the lateral left thigh of the subject, the third node is deployed in the lateral right thigh of the subject, the fourth node is deployed in the anterior tibial region of the left lower leg of the subject, the fifth node is deployed in the anterior tibial region of the right lower leg of the subject, the sixth node is deployed in the dorsum of the left foot of the subject, and the seventh node is deployed in the dorsum of the right foot of the subject.

9. The hemiplegic gait measurement device according to claim 7, characterized in that, The distributed pressure sensor array includes at least five pressure sensors, wherein the first pressure sensor is deployed in the heel region of the subject, the second pressure sensor is deployed in the base of the fifth metatarsal bone of the subject, the third pressure sensor is deployed in the head region of the fifth metatarsal bone of the subject, the fourth pressure sensor is deployed in the head region of the first metatarsal bone of the subject, and the fifth pressure sensor is deployed in the big toe region of the subject.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the hemiplegic gait feature parameter extraction method as described in any one of claims 1-6.

Images