Search

CN-122004792-A - Human body balance function assessment method and system adapting to multitasking and complex scene thereof

CN122004792ACN 122004792 ACN122004792 ACN 122004792ACN-122004792-A

Abstract

The invention provides a human body balance function assessment method and system adapting to multiple tasks and complex scenes of the multiple tasks, comprising the steps of collecting plantar pressure distribution data of a subject in static standing and dynamic walking tasks and bipedal six-degree-of-freedom pose data calculated based on visual inertia instant positioning and mapping technology, reconstructing an overall pressure center track of the whole process of the dynamic and static tasks through time synchronization, rigid body transformation and weighted fusion algorithm, realizing continuous accurate pressure center calculation facing the whole domain, respectively extracting multidimensional assessment indexes for the dynamic and static tasks, adaptively calculating dynamic balance parameters in complex scenes such as turning, obstacle detouring, ascending and descending stairs and the like, mapping the multidimensional assessment indexes to multiple physiological function submodels, calculating comprehensive efficacy indexes of the submodels based on a normal model database of healthy people, and outputting a subject balance control capability assessment result. The invention realizes dynamic and static joint evaluation and complex scene self-adaption, and can accurately position the mechanism root of balance dysfunction.

Inventors

  • REN WEIYAN
  • LIANG SHANGHUA
  • PU FANG

Assignees

  • 北京航空航天大学

Dates

Publication Date
20260512
Application Date
20260414

Claims (10)

  1. 1. A human balance function assessment method adapted to multitasking and complex scenarios thereof, the method comprising: The method comprises the steps of collecting multi-source heterogeneous data of a subject in a multi-task evaluation process comprising a static standing task and a dynamic walking task, wherein the multi-source heterogeneous data comprise plantar pressure distribution data obtained through plantar pressure sensors arranged on feet and bipedal six-degree-of-freedom pose data calculated based on visual inertia real-time positioning and mapping technology through a self-positioning tracker arranged on the feet; performing time synchronization on the plantar pressure distribution data and the bipedal six-degree-of-freedom pose data to obtain a time-aligned synchronous data set; based on the synchronous data set, mapping the plantar pressure distribution data into a unified space coordinate system through rigid body transformation, and reconstructing the integral pressure center track of the subject in the static standing task and the dynamic walking task by combining a weighted fusion algorithm to realize continuous and accurate pressure center calculation facing the whole domain; Based on the synchronous data set and/or the integral pressure center track, extracting a multi-dimensional evaluation index aiming at the static standing task and the dynamic walking task respectively, and constructing a feature vector set based on the multi-dimensional evaluation index; the multi-dimensional evaluation index comprises a static balance parameter, a dynamic balance parameter, a frequency spectrum characteristic, a lower limb joint moment estimation value and a multi-sense organ stimulation response characteristic, wherein the dynamic balance parameter is adaptively calculated under each complex scene based on gait event detection and advancing direction estimation for the dynamic walking task, the complex scene at least comprises turning, obstacle detouring and ascending and descending stairs, the frequency spectrum characteristic is obtained by carrying out frequency spectrum analysis on the integral pressure center track, decomposing the integral pressure center track into a plurality of characteristic frequency bands corresponding to different physiological regulation functions and calculating the power characteristic of each frequency band, the lower limb joint moment estimation value is reversely calculated and obtained based on integral pressure center track data of a supporting period in a gait cycle, and the multi-sense organ stimulation response characteristic is obtained by applying controllable sense organ stimulation to a subject and quantifying the response of the subject under different stimulation conditions, and comprises response lag time and response gain; Mapping the feature vector set to a plurality of preset physiological function sub-models, calculating the comprehensive efficiency index of each physiological function sub-model by taking a normal model database of healthy people as a reference, and outputting the balance control capability assessment results of the subject in different physiological function dimensions.
  2. 2. The human balance function assessment method of adaptive multitasking and its complex scenarios according to claim 1, characterized in that extracting multidimensional assessment index based on the synchronous dataset and/or the global pressure center trajectory comprises: the static balance parameters are obtained based on the calculation of the reconstructed overall pressure center track in the static standing assessment task, and comprise the gravity center moving speed, the total path length of the pressure center track, the maximum displacement of the pressure center and the enveloping area; The dynamic balance parameters are obtained based on gait events detected in the walking balance evaluation task and the estimated travelling direction through calculation, and the dynamic balance parameters comprise step length, stride, step width, step height, step frequency, step speed, gait cycle, support phase proportion, swing phase proportion, double support phase proportion, gait variability and gait symmetry index; The spectrum analysis adopts wavelet decomposition to decompose the integral pressure center track into a plurality of characteristic frequency bands, wherein the characteristic frequency bands comprise a low frequency band of 0-0.3Hz, a medium frequency band of 0.3-1Hz and a high frequency band of 1-3Hz, and the characteristic frequency bands correspond to a vision adjusting function, a vestibular adjusting function and a proprioception adjusting function respectively; The response lag time is obtained by calculating the time difference between the stimulation excitation time and the acceleration of the pressure center exceeding a standard deviation threshold value which is 3 times that of a base line, the response gain is obtained by establishing a stimulation-response transfer function model and calculating the ratio of system output to system input; The lower limb joint moment estimation value is obtained by reversely calculating based on the whole pressure center track data of the supporting period in the gait cycle, and specifically comprises the steps of identifying and extracting the whole pressure center track data of the supporting period in the gait cycle, constructing a high-dimensional space-time input vector containing the dynamic characteristics of the current frame and the historical frame information in the preset step length, introducing a multi-scale sliding window to segment and extract the characteristics, and realizing the prediction of the ankle joint, knee joint and hip joint moment through a cascade random forest model.
  3. 3. The adaptive multitasking and its complex scenario human body balance function assessment method as claimed in claim 1, characterized in that mapping the plantar pressure distribution data into a unified space coordinate system by rigid body transformation based on the synchronous data set, and reconstructing the overall pressure center track of the subject in the whole process of static standing and dynamic walking by combining a weighted fusion algorithm, comprising: Based on the bipedal six-degree-of-freedom pose data in the synchronous data set, mapping a preset sensor local coordinate to a world coordinate system through rigid transformation to obtain an absolute coordinate of each pressure sensor in the world coordinate system, wherein the expression of the rigid transformation is as follows: ; Wherein, the Represent the first Frame(s) Side foot first Absolute coordinates of the individual sensors in the world coordinate system; Representing a rotation matrix obtained by conversion of the rotation quaternion; Representing local coordinates of the sensor; Representing the fixed structure offset from the tracker center to the origin of the insole local coordinate system; Representing the position coordinates of the tracker in the world coordinate system; introducing a self-adaptive weight coefficient, and calculating an integral pressure center under a world coordinate system by combining the projected absolute coordinates of the sensor and the real-time pressure value, wherein the calculation formula is as follows: ; ; Wherein, the 、 Respectively represent the first The global pressure center of the frame is in the world coordinate system Shaft and method for producing the same Position coordinates on the shaft; Representing the filtered pressure value; representing adaptive weight coefficients; 、 respectively representing the sensor in the world coordinate system Shaft and method for producing the same And (5) axis coordinates.
  4. 4. The adaptive multitasking and its complex scenario human balance function assessment method of claim 1, characterized in that when a subject is doing the dynamic walking task, the gait event detection is done based on the synchronized dataset and the global center of pressure trajectory, comprising: calculating the total pressure of the sole of the single side, wherein the expression is as follows: ; Wherein, the Represent the first Frame(s) Total plantar pressure of the lateral foot; Representing the filtered pressure value; Setting a landing threshold and a landing threshold, and detecting gait events by adopting a double-threshold judging mechanism: when the single-side plantar total pressure is larger than or equal to the grounding threshold value and the plantar total pressure of the previous frame is smaller than the grounding threshold value, judging that the single-side plantar total pressure is a grounding event, recording the current moment as the grounding time and recording the current foot position as the grounding position; When the single-side plantar total pressure is smaller than or equal to the ground leaving threshold value and the plantar total pressure of the previous frame is larger than the ground leaving threshold value, judging that the single-side plantar total pressure is a ground leaving event, and recording that the current moment is the ground leaving time; calculating the geometric center point of the bipedal, wherein the expression is as follows: ; Wherein, the Represent the first A bipedal geometric center point under a frame world coordinate system; 、 Respectively representing the position coordinates of the left foot tracker and the right foot tracker in a world coordinate system; constructing a sliding window with preset duration, performing polynomial fitting on a bipedal geometric center point sequence in the window, reconstructing a traveling path of a subject, and calculating a normalized tangent vector of a fitting curve at the current moment to obtain a unit traveling direction vector.
  5. 5. The method for assessing the balance of human body function of adaptive multitasking and its complex scenarios according to claim 1, characterized in that said physiological function sub-models include vision adjusting function sub-model, vestibular adjusting function sub-model, proprioceptive adjusting function sub-model, motor control and executive function sub-model and anti-interference and adaptive control function sub-model.
  6. 6. The method for evaluating the human body balance function of adaptive multitasking and its complex scenario according to claim 1, wherein calculating the comprehensive performance index of each physiological function sub-model comprises: Taking a normal model database of healthy people as a reference, and carrying out standardization processing on characteristic components corresponding to each physiological function sub-model by adopting a Z-Score standardization algorithm; and carrying out fusion calculation on the normalized feature vectors by adopting a multi-factor weighted fusion algorithm to obtain the comprehensive efficiency index of each physiological function sub-model.
  7. 7. A human balance function assessment system adapted to multitasking and complex scenarios thereof, the system comprising: the plantar pressure acquisition module is arranged on both feet of the subject and used for acquiring plantar pressure distribution data of the subject in the evaluation process; The biped pose tracking module comprises two tracker modules which are respectively arranged on the biped of the subject and are used for acquiring biped pose data with six degrees of freedom; a multi-sensory stimulation module for applying controllable sensory stimulation to the subject, the sensory stimulation including at least one of visual stimulation, auditory stimulation, and proprioceptive stimulation; a data processing module in communication with the plantar pressure acquisition module, the bipedal pose tracking module, and the multisensory stimulation module, respectively, the data processing module configured to perform the steps of the method of any one of claims 1 to 6; The diagnosis decision module is in communication connection with the data processing module and is used for mapping the feature vector set to a plurality of preset physiological function sub-models, calculating the comprehensive efficiency index of each physiological function sub-model by taking a normal model database of healthy people as a reference, and outputting the balance control capability assessment result of the subject in different physiological function dimensions.
  8. 8. The adaptive multitasking and its complex scenario human body balance function assessment system of claim 7, characterized in that said plantar pressure acquisition module comprises a plurality of pressure sensors disposed within the insole, said plurality of pressure sensors being disposed in the sole front and heel areas, respectively; The tracker module is rigidly arranged in the heel area of the biped shoe of the subject and keeps a fixed relative position relation with the plantar pressure acquisition module, each tracker is internally provided with a visual sensor and an inertial measurement unit, and six-freedom-degree pose data of the corresponding foot are calculated in real time based on visual inertial instant positioning and mapping technology.
  9. 9. The adaptive multitasking and its complex scenario human balance function assessment system of claim 7, characterized in that said multisensory stimulation module comprises: A visual stimulus sub-module comprising a virtual reality head display for providing a programmable controllable visual stimulus to interfere with visual information of a subject; An auditory stimulus sub-module comprising a three-dimensional spatial audio system for providing a programmable controllable auditory stimulus to interfere with auditory spatial reference information of a subject; A proprioceptive stimulus sub-module for providing a programmable controllable proprioceptive stimulus to interfere with proprioceptive information of a subject, the proprioceptive stimulus comprising at least one of a supporting surface disturbance stimulus and a complex terrain stimulus.
  10. 10. The adaptive multitasking and its complex scenario human balance function assessment system of claim 7, further comprising a safety protection module communicatively connected with said data processing module and said multisensory stimulation module for receiving a dynamic stability margin output by said data processing module and controlling said multisensory stimulation module to terminate stimulation output when said dynamic stability margin is below a preset safety threshold.

Description

Human body balance function assessment method and system adapting to multitasking and complex scene thereof Technical Field The invention relates to the technical field of medical rehabilitation and health monitoring, in particular to a human body balance function assessment method and system adapting to multiple tasks and complex scenes thereof. Background The ability of the body to balance is the basic physiological function of maintaining stable body posture, and its decline is the primary risk factor for the elderly to fall unexpectedly. The balance capacity is objectively and accurately evaluated, and the method has important significance for disease rehabilitation and injury prevention. Currently, the main tool for clinical assessment of balance function is the computerized dynamic gesturing system (Computerized Dynamic Posturography, CDP). The system performs quantitative analysis based on the track of the pressure center (Center of Pressure, COP), but relies on a fixed pressure test bench or a pressure pavement, can only perform static standing assessment in a laboratory environment, and cannot effectively measure dynamic activities such as walking, turning, going up and down stairs and the like. To break through the limitations of stationary devices, researchers have proposed wearable evaluation schemes. Firstly, based on the scheme of pressure monitoring shoe-pad, can obtain plantar pressure distribution in succession, but can't obtain the relative pose of both legs, can only measure plantar COP, is difficult to integrate bipedal pressure into human whole COP. Secondly, based on the scheme of an inertial sensor (Inertial Measurement Unit, IMU), the integral of the IMU has accumulated drift, and the existing algorithm has mutual exclusion between dynamic and static tasks, namely zero-speed update fails in static evaluation, and a gradient descent method is easy to be interfered in dynamic walking. Thirdly, the multi-sensor fusion scheme (IMU+pressure insole) can realize COP estimation in a walking task, but has a drift problem in static estimation, is mostly limited to straight walking, and is invalid in model in complex scenes such as turning, obstacle detouring and the like. In addition, the existing Virtual Reality evaluation schemes mostly take Virtual Reality (VR) equipment as a visual feedback tool, lack accurate control on stimulus parameters such as visual flow, spatial audio and the like, and have single evaluation indexes, so that the contributions of different subsystems such as vision, vestibular sensation, proprioception and the like to balance control are difficult to distinguish. Therefore, the existing balance function evaluation technical scheme has the problems of scene splitting, dynamic and static mutual exclusion, complex scene analysis deficiency, incapability of quantifying multi-sensory stimulation, single functional diagnosis dimension and the like. Disclosure of Invention In view of this, the embodiment of the invention provides a human body balance function evaluation method and system adapting to multitasks and complex scenes thereof, so as to solve the problems that equipment and scene fracture in the existing balance function evaluation technical scheme cannot cover real life scenes, dynamic and static evaluation tasks are mutually exclusive, the same system is difficult to consider accurate measurement of static standing and dynamic walking, gait analysis capability loss in complex scenes cannot quantify real motion characteristics such as turning obstacle detouring, multisensory stimulation evaluation lacks accurate control and quantitative analysis, and functional diagnosis indexes are single and difficult to position the root cause of a mechanism of dysfunction. In one aspect, the present invention provides a human body balance function assessment method adapted to multitasking and complex scenarios thereof, the method comprising: The method comprises the steps of collecting multi-source heterogeneous data of a subject in a multi-task evaluation process comprising a static standing task and a dynamic walking task, wherein the multi-source heterogeneous data comprise plantar pressure distribution data obtained through plantar pressure sensors arranged on feet and bipedal six-degree-of-freedom pose data calculated based on visual inertia real-time positioning and mapping technology through a self-positioning tracker arranged on the feet; performing time synchronization on the plantar pressure distribution data and the bipedal six-degree-of-freedom pose data to obtain a time-aligned synchronous data set; Based on the synchronous data set, mapping the plantar pressure distribution data into a unified space coordinate system through rigid body transformation, and reconstructing the integral pressure center track of the subject in the whole process of static standing and dynamic walking by combining a weighted fusion algorithm to realize continuous and accurate pressure