CN117379284A - Control method and device for hip joint exoskeleton, terminal equipment and storage medium - Google Patents

Control method and device for hip joint exoskeleton, terminal equipment and storage medium Download PDF

Info

Publication number
CN117379284A
CN117379284A CN202311294490.6A CN202311294490A CN117379284A CN 117379284 A CN117379284 A CN 117379284A CN 202311294490 A CN202311294490 A CN 202311294490A CN 117379284 A CN117379284 A CN 117379284A
Authority
CN
China
Prior art keywords
hip joint
joint angle
angle sequence
moment
hip
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311294490.6A
Other languages
Chinese (zh)
Inventor
冷雨泉
张原文
付成龙
熊靖峰
杜豪
刘海峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Southwest University of Science and Technology
Original Assignee
Southwest University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Southwest University of Science and Technology filed Critical Southwest University of Science and Technology
Priority to CN202311294490.6A priority Critical patent/CN117379284A/en
Publication of CN117379284A publication Critical patent/CN117379284A/en
Pending legal-status Critical Current

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H3/00—Appliances for aiding patients or disabled persons to walk about
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1121—Determining geometric values, e.g. centre of rotation or angular range of movement
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/45—For evaluating or diagnosing the musculoskeletal system or teeth
    • A61B5/4528—Joints
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7232—Signal processing specially adapted for physiological signals or for diagnostic purposes involving compression of the physiological signal, e.g. to extend the signal recording period
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7271—Specific aspects of physiological measurement analysis
    • A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H3/00—Appliances for aiding patients or disabled persons to walk about
    • A61H2003/005—Appliances for aiding patients or disabled persons to walk about with knee, leg or stump rests
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
    • A61H2201/50—Control means thereof
    • A61H2201/5053—Control means thereof mechanically controlled
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61H—PHYSICAL THERAPY APPARATUS, e.g. DEVICES FOR LOCATING OR STIMULATING REFLEX POINTS IN THE BODY; ARTIFICIAL RESPIRATION; MASSAGE; BATHING DEVICES FOR SPECIAL THERAPEUTIC OR HYGIENIC PURPOSES OR SPECIFIC PARTS OF THE BODY
    • A61H2201/00—Characteristics of apparatus not provided for in the preceding codes
    • A61H2201/50—Control means thereof
    • A61H2201/5058—Sensors or detectors
    • A61H2201/5061—Force sensors
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Medical Informatics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Signal Processing (AREA)
  • Physiology (AREA)
  • Artificial Intelligence (AREA)
  • Psychiatry (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Dentistry (AREA)
  • Orthopedic Medicine & Surgery (AREA)
  • Physical Education & Sports Medicine (AREA)
  • Rheumatology (AREA)
  • Rehabilitation Therapy (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Geometry (AREA)
  • Epidemiology (AREA)
  • Pain & Pain Management (AREA)
  • Manipulator (AREA)

Abstract

The invention provides a control method, a device, terminal equipment and a storage medium of a hip joint exoskeleton, which comprise the following steps: collecting a first hip joint angle sequence and moment when a human body moves; performing data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence, wherein the data enhancement processing comprises at least one of noise adding processing on the first hip joint angle sequence, amplitude scaling processing on the first hip joint angle sequence and random resampling processing on the first hip joint angle sequence; data alignment is carried out on the second hip joint angle sequence and the moment so as to construct a training data set; training the initial model based on the training data set to obtain a target model; and (3) carrying out reasoning model conversion on the target model, and realizing the control of the hip joint exoskeleton through the bottom layer control logic. The method and the device can reduce the difference between the training data set and the data in actual application, improve the adaptability of the target model and improve the accuracy of moment prediction.

Description

Control method and device for hip joint exoskeleton, terminal equipment and storage medium
Technical Field
The invention belongs to the technical field of control, and particularly relates to a control method and device for a hip joint exoskeleton, terminal equipment and a storage medium.
Background
With the aggravation of population aging and the increase of the number of patients with dyskinesia, the number of patients with impaired lower limb movement functions is more and more, and the problem of helping the patients to complete normal life activities is urgent to be solved. The hip exoskeleton is capable of assisting the wearer in locomotion by applying a correct assistance torque to the hip exoskeleton assistance strategy.
However, the existing hip joint exoskeleton assistance strategy has weak anti-interference capability in practical application, has poor self-adaptability and cannot realize accurate moment prediction.
Disclosure of Invention
In view of the above, the embodiment of the invention provides a control method, a device, a terminal device and a storage medium for a hip joint exoskeleton, which are used for solving the problems that in the prior art, the interference adaptability of a hip joint exoskeleton assistance strategy to practical application is poor and accurate moment prediction cannot be realized.
A first aspect of an embodiment of the present invention provides a method for controlling a hip exoskeleton, including:
collecting a first hip joint angle sequence and moment when a human body moves;
performing data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence, wherein the data enhancement processing comprises at least one of noise adding processing on the first hip joint angle sequence, amplitude scaling processing on the first hip joint angle sequence and random resampling processing on the first hip joint angle sequence;
data alignment of the second hip angle sequence with the moment to construct a training dataset;
training an initial model based on the training data set to obtain a target model;
and (3) converting the target model into an inference model, and controlling the hip joint exoskeleton through bottom control logic.
According to the method, the first hip joint angle sequence is subjected to data enhancement processing to obtain the second hip joint angle sequence, so that the difference between a training data set and data acquired in actual application can be reduced, the adaptability of a target model to the hip joint angles acquired by the motor encoder in actual application is improved, and the accuracy of moment prediction is improved.
In a possible implementation manner of the first aspect, the acquiring a first hip joint angle sequence and moment when the human body moves includes:
a first hip joint angle sequence and moment when the human body moves are acquired in a motion capture environment.
In a possible implementation manner of the first aspect, the performing data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence includes:
carrying out noise adding treatment on the first hip joint angle sequence to obtain a noisy hip joint angle sequence;
performing amplitude scaling treatment on the noisy hip joint angle sequence to obtain a scaled hip joint angle sequence;
and carrying out random resampling processing on the scaled hip joint angle sequence to obtain a second hip joint angle sequence.
According to the method, through noise adding processing on the hip joint angle sequence, data noise possibly existing when the actual motor encoder is used for measuring the hip joint angle can be simulated, and generalization performance of a target model on noisy data is improved; the amplitude value of the first hip joint angle sequence is scaled, so that the diversity of hip joint angle changes can be increased, and the difference between the measured angle of the motor encoder and the measured angle in the motion capture environment can be reduced; and carrying out random resampling processing on the first hip joint angle sequence, and simulating different angle sampling frequencies under actual conditions. The adaptability of the target model to the hip joint angles acquired in practical application is improved and the accuracy of moment prediction is improved by reducing the difference between the hip joint angles used in training and the hip joint angles acquired in practical application.
In a possible implementation manner of the first aspect, the data aligning the second hip joint angle sequence with the moment to construct a training data set includes:
and aligning the moment of the hip joint angle at the first moment and the moment of the moment at the second moment in the second hip joint angle sequence, and constructing a training data set, wherein the time interval between the first moment and the second moment is a first time interval, and the second moment is after the first moment.
In a possible implementation manner of the first aspect, the randomly resampling the scaled hip angle sequence to obtain a second hip angle sequence includes:
and (3) adopting random up-sampling, interpolating the scaled hip joint angle sequence, and resampling to obtain a second hip joint angle sequence.
In a possible implementation manner of the first aspect, the training the initial model based on the training data set to obtain the target model includes:
constructing an initial model comprising N convolution layers and a full connection layer, wherein the input length is a first input length, N belongs to a first set, and the first input length belongs to a second set;
setting a training hyper-parameter for the initial model, and setting the upper limit of the training times of the epoch corresponding to the training hyper-parameter as m;
dividing the training data set into a training set and a verification set;
training the initial model based on the training set, and obtaining an intermediate model after each training (1 epoch) is completed, and verifying the performance of the intermediate model by using the verification set, wherein the performance comprises prediction accuracy and average single prediction time;
obtaining m intermediate models after finishing the training of m epochs, and selecting the intermediate model with the highest prediction accuracy from the m intermediate models as a candidate model;
repeating the steps to obtain a plurality of candidate models;
and carrying out model screening on the candidate models according to the performance to determine a final target model.
In a possible implementation manner of the first aspect, the performing inference model transformation on the target model, implementing control of the hip exoskeleton through the underlying control logic, includes:
converting the target model into an inference model, and deploying the inference model on an exoskeleton airborne processor;
initializing, starting power assistance when a hip joint exoskeleton wearer stands, and defining the hip joint angles of two legs to be 0 degrees and corresponding to the initial value of a motor encoder;
acquiring the hip joint angle at the current moment, and subtracting the initial value of the motor encoder from the current value of the motor encoder to acquire the hip joint angle at the current moment;
storing the hip joint angle at the current moment in a buffer sequence, wherein the length of the buffer sequence is the first input length of an initial model corresponding to the target model; if the buffer sequence is full, deleting the hip joint angle at the earliest moment in the buffer sequence, and increasing the hip joint angle at the current moment to update the buffer sequence;
inputting the buffer queue into the reasoning model and predicting the moment at a third moment after the first time interval at the current moment;
multiplying the moment by a first coefficient and inputting the first coefficient into a motor controller to generate power assistance;
all the steps after the initialization are repeatedly executed, and the power assistance is continuously generated.
A second aspect of an embodiment of the present invention provides a control device for a data-driven hip exoskeleton, including:
the data acquisition module acquires a first hip joint angle sequence and moment when a human body moves;
the data processing module is used for carrying out data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence, wherein the data enhancement processing comprises at least one of noise adding processing on the first hip joint angle sequence, amplitude scaling processing on the first hip joint angle sequence and random resampling processing on the first hip joint angle sequence;
the data alignment module performs data alignment on the second hip joint angle sequence and the moment to construct a training data set;
the model training module is used for training an initial model based on the training data set to obtain a target model;
and the model deployment module is used for carrying out reasoning model conversion on the target model and realizing the control of the hip joint exoskeleton through the bottom layer control logic.
A third aspect of an embodiment of the present invention provides a terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method as provided in the first aspect when executing the computer program.
A fourth aspect of the embodiments of the present invention provides a storage medium being a computer readable storage medium storing a computer program which, when executed by a processor, implements the steps of the method as provided in the first aspect.
It will be appreciated that the advantages of the second to fourth aspects may be found in the relevant description of the first aspect and are not repeated here.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments or the description of the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic implementation flow chart of a control method of a hip joint exoskeleton provided by an embodiment of the present invention;
FIG. 2 is a CNN overall model diagram of a control method for a hip exoskeleton provided by an embodiment of the present invention;
fig. 3 is a Conv layer specific structure diagram of a control method of a hip joint exoskeleton provided by an embodiment of the present invention;
FIG. 4 is a schematic diagram of exoskeleton control logic flow of a method for controlling a hip joint exoskeleton according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a control device for a data-driven hip exoskeleton according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a control terminal device based on a data-driven hip exoskeleton provided in an embodiment of the present invention;
fig. 7 is a hip model in human motion provided by an embodiment of the present invention.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
It should be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It should also be understood that the term "and/or" as used in this specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
As used in this specification and the appended claims, the term "if" may be interpreted as "when..once" or "in response to a determination" or "in response to detection" depending on the context. Similarly, the phrase "if a determination" or "if a [ described condition or event ] is detected" may be interpreted in the context of meaning "upon determination" or "in response to determination" or "upon detection of a [ described condition or event ]" or "in response to detection of a [ described condition or event ]".
In addition, in the description of the present application and the appended claims, the terms "first," "second," "third," and the like are used merely to distinguish between descriptions and are not to be construed as indicating or implying relative importance.
Reference in the specification to "one embodiment" or "some embodiments" or the like means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," and the like in the specification are not necessarily all referring to the same embodiment, but mean "one or more but not all embodiments" unless expressly specified otherwise. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless expressly specified otherwise.
As shown in fig. 1, an embodiment of the present application provides a control method of a hip joint exoskeleton, including the following steps S10 to S50:
step S10, acquiring a first hip joint angle sequence and moment when a human body moves.
In application, a first hip joint angle sequence and moment when a human body moves can be acquired in a motion capture environment. Specifically, a mark point (mark point) can be attached to the corresponding position of the experimenter's body using the Helen Hayes dotting method. Setting various motion environments such as a flat ground, an ascending slope, a descending slope, an ascending stair, a descending stair and the like, under the action capturing environment, enabling an experimenter to walk in different motion environments at different pace, and collecting motion information of a human body mark point (mark point) and force measuring plate data. According to the motion information of the human body mark points and the force measuring plate data acquired in the motion capturing environment, a first hip joint angle sequence and moment when the human body moves are calculated and generated.
It can be understood that the experimenters can make the collected data richer and have diversity by moving in different movement environments at different pace, so that the network model trained based on the data has better generalization performance and can be applied to various complicated practical situations.
Step S20, performing data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence, where the data enhancement processing includes at least one of noise adding processing on the first hip joint angle sequence, amplitude scaling processing on the first hip joint angle sequence, and random resampling processing on the first hip joint angle sequence.
In application, since the actual hip joint angle sequence is measured by the motor encoder in the actual hip joint exoskeleton, then moment prediction is performed based on the actual hip joint angle sequence as an input of the network model, and the first hip joint angle sequence used for training the network model is obtained by the Helen Hayes dotting method in the motion capture environment, a certain difference exists between the actual hip joint angle sequence measured by the motor encoder and the first hip joint angle sequence, which may cause deviation of the network model for predicting the moment. Therefore, the first hip joint angle sequence needs to be subjected to data enhancement processing to obtain a second hip joint angle sequence which is closer to the actual hip joint angle sequence measured by the motor encoder, so that the performance of the network model in practical application, such as the prediction speed and the prediction accuracy of moment, is improved.
The data enhancement processing may include noise-adding the first hip-joint angle sequence and/or amplitude-magnitude scaling of the first hip-joint angle sequence and/or random resampling of the first hip-joint angle sequence.
Specifically, the noise adding process is performed on the first hip joint angle sequence, which may be to simulate data noise that may exist when the angle is measured by using the motor encoder in practice by adding a random noise n (x) to the first hip joint angle sequence. The random noise generation scheme may be as follows:
n(x i )＝rand(0,1)*k n
wherein rand (0, 1) represents a random number, k, that generates 0-1 n Representing the noise reduction and amplification, given an appropriate k n The generalization performance of the model to noisy data can be improved.
Specifically, the scaling processing of the amplitude value of the first hip joint angle sequence may be performed by multiplying the whole first hip joint angle sequence by a random scaling factor, so as to increase the diversity of the variation of the first hip joint angle sequence and reduce the difference between the measured angle of the motor encoder and the measured angle in the motion capture environment. The scaling of the magnitude may be as follows:
wherein X is i Is a first hip angle sequence, K1 is a random scaling factor,is a hip angle sequence with amplitude scaling.
Specifically, the random resampling process may be a random up-sampling process or a random down-sampling process, and the first hip joint angle sequence is resampled after interpolation or extraction to obtain hip joint angle sequences with different frequencies so as to simulate different angle sampling frequencies under practical conditions. The random resampling avoids the problem that the angular acquisition frequency of the motor encoder is different from the acquisition frequency in the motion capture environment, so that the network model obtains wrong characteristic information, and the moment prediction is inaccurate.
It can be understood that, in order to overcome the difference between the first hip joint angle sequence obtained by the Helen Hayes punctuation method under the motion capture environment for training the network model and the actual hip joint angle sequence measured by the motor encoder in the actual hip joint exoskeleton for moment prediction by the input network model as far as possible, the first hip joint angle sequence can be subjected to multiple-aspect data enhancement processing according to a certain step to obtain a second hip joint angle sequence.
Specifically, the step of performing data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence may be that the step of performing noise adding processing on the first hip joint angle sequence to obtain a noisy hip joint angle sequence; performing amplitude scaling treatment on the noisy hip joint angle sequence to obtain a scaled hip joint angle sequence; and carrying out random resampling processing on the scaled hip joint angle sequence to obtain a second hip joint angle sequence.
And step S30, carrying out data alignment on the second hip joint angle sequence and the moment to construct a training data set.
It will be appreciated that the function to be achieved after the final deployment of the target model on the hip exoskeleton is to predict future joint moments from the actual hip angle sequence measured by the motor encoder, so that moment data of the hip angle at a first time and moment data of a second time need to be aligned, and a training data set for training the network model is constructed, where the time interval between the first time and the second time is a first time interval, and the second time is after the first time.
Specifically, the first time interval can be obtained by measuring the control delay, and the method is that the average time required for obtaining the calculated control quantity is the first time interval by performing complete control for a plurality of times by using a prediction model through a preliminary experiment. For example, assuming a first time interval = 0.01s, the second hip angle sequence is data aligned with the moment to construct a training data set, an initial model with a particular network size, a particular input length is constructed, the initial model is applied to the exoskeleton control system and the average time required to determine the final control quantity is calculated.
And step S40, training an initial model based on the training data set to obtain a target model.
In one implementation manner of this embodiment, training the initial model based on the training data set to obtain the target model includes:
constructing an initial model comprising N convolution layers and a full connection layer, wherein the input length is a first input length, N belongs to a first set, and the first input length belongs to a second set;
setting a training hyper-parameter for the initial model, and setting the upper limit of the training times of the epoch corresponding to the training hyper-parameter as m;
dividing the training data set into a training set and a verification set;
training the initial model based on the training set, and obtaining an intermediate model each time the epoch training is completed, and verifying the performance of the intermediate model by using the verification set, wherein the performance comprises prediction accuracy and average single prediction time;
obtaining m intermediate models after finishing the training of m epochs, and selecting the intermediate model with the highest prediction accuracy from the m intermediate models as a candidate model;
repeating the steps to obtain a plurality of candidate models;
and carrying out model screening on the candidate models according to the performance to determine a final target model.
The initial model may be a Convolutional Neural Network (CNN) model for achieving a mapping relationship τ=f (θ) between hip joint angles and hip joint moments, where θ is the hip joint angle and τ is the hip joint moment. The initial model includes N CNN convolution layers and a fully-connected layer, wherein N belongs to the first set, each convolution layer has 32 output channels and a convolution kernel with a size of 5×1
Specifically, the initial model may be a CNN overall model diagram as shown in fig. 2, where Input is an Input layer, save N is N convolution layers, pooling is a Pooling layer, FC is a full connection layer, and Output is an Input layer. The convolutional layers may include a ReLU activation layer and a BN (BatchNormalization) layer to improve network performance.
Specifically, the specific structure of the convolution layer may refer to the Conv layer specific structure shown in fig. 3, where the convolution kernel size (kernel_size) in the convolution layer is 5, the stride (stride) is 2, the padding (padding) is 2, and the convolution layer is followed by a ReLU layer and a batch normalization layer in sequence.
It will be appreciated that the order of the layers in the CNN may be adjusted, e.g. the ReLU layer and the BN layer may precede the BN layer, with the ReLU layer following. The size, the stride and the filling of the convolution kernel in the convolution layer can be flexibly adjusted according to actual needs.
According to the dynamic expressionWherein I represents an inertia term describing the inertia of the links and joints of the robot, +.>Represents a damping term, g (theta) represents a gravity term, theta is a hip joint angle, tau is a hip joint moment, < + >>Hip joint angular velocity and angular acceleration, respectively. Due to->Can be derived from θ (t) to t, and thus can be a mapping τ=f (θ) between hip moment and hip angle.
In an application, the first set may be {1,2,3,4,5,6,7,8}, the second set may be {40,60,100,200,300,400} (corresponding to a sampling time of 0.2s, 0.3s, 0.5s, 1.0s, 1.5s, 2.0s at a sampling frequency of 200 Hz), and the upper limit m of the number of epoch training may be 500. Corresponding initial models with 48 different sizes and input lengths can be built, 500 intermediate models can be obtained after each initial model is trained by 500 epochs, and for each initial model, the intermediate model with the highest prediction accuracy is selected from the 500 intermediate models to serve as a candidate model corresponding to each initial model, and finally 48 candidate models can be obtained. The final target model is then determined from the 48 candidate models based on the performance of the candidate models, including the prediction accuracy and the average single prediction time. It will be appreciated that there may be only one, or more than one, final object model.
In application, the training super-parameters select an Adam optimizer for training, a dynamic learning rate method is adopted, an initial value is set to be 0.1, a batch size is set to be 512, and the maximum training epoch number is 500. The loss function used in training the initial model based on the training set may be a mean square error loss function (MSE function) or a softmax loss function.
And S50, carrying out reasoning model conversion on the target model, and realizing the control of the hip joint exoskeleton through the bottom layer control logic.
In one implementation manner of this embodiment, the performing the inference model transformation on the target model, and implementing the control of the hip exoskeleton through the underlying control logic includes:
converting the target model into an inference model, and deploying the inference model on an exoskeleton airborne processor;
initializing, starting power assistance when a hip joint exoskeleton wearer stands, and defining the hip joint angles of two legs to be 0 degrees and corresponding to the initial value of a motor encoder;
acquiring the hip joint angle at the current moment, and subtracting the initial value of the motor encoder from the current value of the motor encoder to acquire the hip joint angle at the current moment;
storing the hip joint angle at the current moment in a buffer sequence, wherein the length of the buffer sequence is the first input length of an initial model corresponding to the target model; if the buffer sequence is full, deleting the hip joint angle at the earliest moment in the buffer sequence, and increasing the hip joint angle at the current moment to update the buffer sequence;
inputting the buffer queue into the reasoning model and predicting the moment at a third moment after the first time interval at the current moment;
multiplying the moment by a first coefficient and inputting the first coefficient into a motor controller to generate power assistance;
all the steps after the initialization are repeatedly executed, and the power assistance is continuously generated.
Referring to the exoskeleton control logic flow diagram of fig. 4, initialization is performed under an initialization condition, an initial value of an encoder is an initial value of a motor encoder during initialization, a current value of the encoder is a current value of a motor encoder at a current moment, a hip joint angle is a hip joint angle at the current moment calculated according to the initial value and the current value, the hip joint angle at the current moment is stored in a buffer sequence to update the buffer sequence, the updated buffer sequence is input into an inference model deployed on an onboard processor of the hip joint exoskeleton, torque prediction is performed, a torque is obtained through prediction, the torque is multiplied by a proportionality coefficient K and is input into a motor controller, and assistance is generated, wherein the assistance acts on the hip joint exoskeleton.
It will be appreciated that the initialization may be performed when the hip exoskeleton wearer is standing to initiate assistance, or when the hip exoskeleton wearer is attempting to forcibly restart the hip exoskeleton.
Fig. 7 schematically illustrates a model of a hip joint in human motion, wherein θ l And theta r The flexion/extension angle, i.e. hip angle, representing the vertical of the left and right legs with respect to the direction of gravity, when θ l And the theta r When 0, i.e., standing.
In application, converting the target model into an inference model may be converting the trained target model into an ONNX model, and using ONNX run as an inference engine to accelerate the speed of inference of the model on the exoskeleton on-board processor.
Fig. 6 is a schematic structural diagram of a control terminal device based on a data-driven hip exoskeleton according to an embodiment of the present application. As shown in fig. 6, the control terminal device 2 of the data-driven hip exoskeleton of this embodiment includes: at least one processor 20 (only one is shown in fig. 6), a memory 21 and a computer program 22 stored in the memory 21 and executable on the at least one processor 20, the processor 20 implementing the steps in any of the various hip exoskeleton control method embodiments described above when executing the computer program 22.
The control terminal device 2 based on the data-driven hip joint exoskeleton can be a computing device such as a smart phone, a palm computer and a cloud server. The data-driven hip exoskeleton-based control terminal device may include, but is not limited to, a processor 20, a memory 21. It will be appreciated by those skilled in the art that fig. 6 is merely an example of a data-driven hip exoskeleton-based control terminal device 2 and is not intended to be limiting of the data-driven hip exoskeleton-based control terminal device 2, and may include more or fewer components than shown, or may combine certain components, or different components, such as may also include input and output devices, communication devices, etc.
The processor 20 may be a central processing unit (Central Processing Unit, CPU), and the processor 20 may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field-programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory 21 may in some embodiments be an internal storage unit of the control terminal device 2 based on a data-driven hip exoskeleton, for example a hard disk or a memory of the control terminal device 2 based on a data-driven hip exoskeleton. The memory 21 may in other embodiments also be an external memory device of the control terminal device 2 based on a data-driven hip exoskeleton, such as a plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card) or the like provided on the control terminal device 2 based on a data-driven hip exoskeleton. Further, the memory 21 may also comprise both an internal memory unit and an external memory device of the data-driven hip exoskeleton-based control terminal device 2. The memory 21 is used for storing an operating system, application programs, boot loader (BootLoader), data, other programs, etc., such as program codes of the computer program. The memory 21 may also be used for temporarily storing data that has been output or is to be output.
It should be noted that, because the content of information interaction and execution process between the above devices/units is based on the same concept as the method embodiment of the present application, specific functions and technical effects thereof may be referred to in the method embodiment section, and will not be described herein again.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
The embodiment of the application also provides a network device, which comprises: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, which when executed by the processor performs the steps of any of the various method embodiments described above.
Embodiments of the present application also provide a computer readable storage medium storing a computer program which, when executed by a processor, implements steps that may implement the various method embodiments described above.
Embodiments of the present application provide a computer program product which, when run on a mobile terminal, causes the mobile terminal to perform steps that may be performed in the various method embodiments described above.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present application implements all or part of the flow of the method of the above embodiments, and may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium, where the computer program, when executed by a processor, may implement the steps of each of the method embodiments described above. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include at least: any entity or device capable of carrying computer program code to an apparatus/terminal device, recording medium, computer Memory, read-Only Memory (ROM), random access Memory (RAM, random Access Memory), electrical carrier signals, telecommunications signals, and software distribution media. Such as a U-disk, removable hard disk, magnetic or optical disk, etc. In some jurisdictions, computer readable media may not be electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other manners. For example, the apparatus/terminal device embodiments described above are merely illustrative, e.g., the division of the modules or units is merely a logical function division, and there may be additional divisions in actual implementation, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.

Claims (10)

1. A method of controlling a hip exoskeleton, comprising:
collecting a first hip joint angle sequence and moment when a human body moves;
performing data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence, wherein the data enhancement processing comprises at least one of noise adding processing on the first hip joint angle sequence, amplitude scaling processing on the first hip joint angle sequence and random resampling processing on the first hip joint angle sequence;
data alignment of the second hip angle sequence with the moment to construct a training dataset;
training an initial model based on the training data set to obtain a target model;
and (3) converting the target model into an inference model, and controlling the hip joint exoskeleton through bottom control logic.
2. The method of claim 1, wherein the capturing of the first hip angle sequence and moments during the human motion comprises:
a first hip joint angle sequence and moment when the human body moves are acquired in a motion capture environment.
3. The method for controlling a hip exoskeleton of claim 1, wherein the performing data enhancement processing on the first hip angle sequence to obtain a second hip angle sequence includes:
carrying out noise adding treatment on the first hip joint angle sequence to obtain a noisy hip joint angle sequence;
performing amplitude scaling treatment on the noisy hip joint angle sequence to obtain a scaled hip joint angle sequence;
and carrying out random resampling processing on the scaled hip joint angle sequence to obtain a second hip joint angle sequence.
4. A method of controlling a hip exoskeleton as claimed in claim 3 wherein said data alignment of said second hip angle sequence with said moments to construct a training data set comprises:
and aligning the moment of the hip joint angle at the first moment and the moment of the moment at the second moment in the second hip joint angle sequence, and constructing a training data set, wherein the time interval between the first moment and the second moment is a first time interval, and the second moment is after the first moment.
5. A method of controlling a hip exoskeleton as claimed in claim 3 wherein randomly resampling the scaled hip angle sequence to obtain a second hip angle sequence comprises:
and (3) adopting random up-sampling, interpolating the scaled hip joint angle sequence, and resampling to obtain a second hip joint angle sequence.
6. The method for controlling a hip exoskeleton of any one of claims 1 to 5 wherein training an initial model based on the training data set to obtain a target model comprises:
constructing an initial model comprising N convolution layers and a full connection layer, wherein the input length is a first input length, N belongs to a first set, and the first input length belongs to a second set;
setting a training hyper-parameter for the initial model, and setting the upper limit of the training times of the epoch corresponding to the training hyper-parameter as m;
dividing the training data set into a training set and a verification set;
training the initial model based on the training set, and obtaining an intermediate model each time the epoch training is completed, and verifying the performance of the intermediate model by using the verification set, wherein the performance comprises prediction accuracy and average single prediction time;
obtaining m intermediate models after finishing the training of m epochs, and selecting the intermediate model with the highest prediction accuracy from the m intermediate models as a candidate model;
repeating the steps to obtain a plurality of candidate models;
and carrying out model screening on the candidate models according to the performance to determine a final target model.
7. The method for controlling a hip exoskeleton according to claims 1 to 5, wherein said performing an inference model transformation on the target model, implementing the control of the hip exoskeleton by the underlying control logic, comprises:
converting the target model into an inference model, and deploying the inference model on an exoskeleton airborne processor;
initializing, starting power assistance when a hip joint exoskeleton wearer stands, and defining the hip joint angles of two legs to be 0 degrees and corresponding to the initial value of a motor encoder;
acquiring the hip joint angle at the current moment, and subtracting the initial value of the motor encoder from the current value of the motor encoder to acquire the hip joint angle at the current moment;
storing the hip joint angle at the current moment in a buffer sequence, wherein the length of the buffer sequence is the first input length of an initial model corresponding to the target model; if the buffer sequence is full, deleting the hip joint angle at the earliest moment in the buffer sequence, and increasing the hip joint angle at the current moment to update the buffer sequence;
inputting the buffer queue into the reasoning model and predicting the moment at a third moment after the first time interval at the current moment;
multiplying the moment by a first coefficient and inputting the first coefficient into a motor controller to generate power assistance;
all the steps after the initialization are repeatedly executed, and the power assistance is continuously generated.
8. A data-driven hip exoskeleton-based control device, comprising:
the data acquisition module acquires a first hip joint angle sequence and moment when a human body moves;
the data processing module is used for carrying out data enhancement processing on the first hip joint angle sequence to obtain a second hip joint angle sequence, wherein the data enhancement processing comprises at least one of noise adding processing on the first hip joint angle sequence, amplitude scaling processing on the first hip joint angle sequence and random resampling processing on the first hip joint angle sequence;
the data alignment module performs data alignment on the second hip joint angle sequence and the moment to construct a training data set;
the model training module is used for training an initial model based on the training data set to obtain a target model;
and the model deployment module is used for carrying out reasoning model conversion on the target model and realizing the control of the hip joint exoskeleton through the bottom layer control logic.
9. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 7 when the computer program is executed.
10. A storage medium being a computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor realizes the steps of the method according to any one of claims 1 to 7.
CN202311294490.6A 2023-09-28 2023-09-28 Control method and device for hip joint exoskeleton, terminal equipment and storage medium Pending CN117379284A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311294490.6A CN117379284A (en) 2023-09-28 2023-09-28 Control method and device for hip joint exoskeleton, terminal equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311294490.6A CN117379284A (en) 2023-09-28 2023-09-28 Control method and device for hip joint exoskeleton, terminal equipment and storage medium

Publications (1)

Publication Number Publication Date
CN117379284A true CN117379284A (en) 2024-01-12

Family

ID=89464099

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311294490.6A Pending CN117379284A (en) 2023-09-28 2023-09-28 Control method and device for hip joint exoskeleton, terminal equipment and storage medium

Country Status (1)

Country Link
CN (1) CN117379284A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090104261A (en) * 2008-03-31 2009-10-06 주식회사 피앤에스미캐닉스 Robot for walking training and working method thereof
CN113269048A (en) * 2021-04-29 2021-08-17 北京工业大学 Motor imagery electroencephalogram signal classification method based on deep learning and mixed noise data enhancement
CN113940855A (en) * 2021-10-19 2022-01-18 北京动思创新科技有限公司 Exoskeleton device control method and exoskeleton device
WO2022012491A1 (en) * 2020-07-13 2022-01-20 Angelexo Scientific Co., Ltd Single-lower-limb rehabilitation exoskeleton apparatus and control method
CN116766197A (en) * 2023-07-12 2023-09-19 河北工业大学 Hip joint exoskeleton power-assisted control method

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20090104261A (en) * 2008-03-31 2009-10-06 주식회사 피앤에스미캐닉스 Robot for walking training and working method thereof
WO2022012491A1 (en) * 2020-07-13 2022-01-20 Angelexo Scientific Co., Ltd Single-lower-limb rehabilitation exoskeleton apparatus and control method
CN113269048A (en) * 2021-04-29 2021-08-17 北京工业大学 Motor imagery electroencephalogram signal classification method based on deep learning and mixed noise data enhancement
CN113940855A (en) * 2021-10-19 2022-01-18 北京动思创新科技有限公司 Exoskeleton device control method and exoskeleton device
CN116766197A (en) * 2023-07-12 2023-09-19 河北工业大学 Hip joint exoskeleton power-assisted control method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
许宇宁，黄京，陈恳，付成龙: "双足负重行走中扁担参数对腰部力矩影响的仿真研究", 清华大学学报(自然科学版), 22 April 2019 (2019-04-22) *

Similar Documents

Publication Publication Date Title
CN111401516B (en) Searching method for neural network channel parameters and related equipment
CN110689109B (en) Neural network method and device
GB2561277B (en) Forecasting multiple poses based on a graphical image
US10296804B2 (en) Image recognizing apparatus, computer-readable recording medium, image recognizing method, and recognition apparatus
TWI825596B (en) Circuit, method and non-transitory machine-readable storage devices for performing neural network computations
KR20180057096A (en) Device and method to perform recognizing and training face expression
CN107016342A (en) A kind of action identification method and system
CN113065635A (en) Model training method, image enhancement method and device
CN113159273B (en) Neural network training method and related equipment
WO2023246819A1 (en) Model training method and related device
CN111738403A (en) Neural network optimization method and related equipment
CN111950700A (en) Neural network optimization method and related equipment
CN111950702A (en) Neural network structure determining method and device
CN111428566A (en) Deformation target tracking system and method
CN115238909A (en) Data value evaluation method based on federal learning and related equipment thereof
CN117379284A (en) Control method and device for hip joint exoskeleton, terminal equipment and storage medium
CN111590544A (en) Method and device for determining output force of exoskeleton
CN114485417B (en) Structural vibration displacement identification method and system
CN115952916A (en) Artificial intelligence-based wind power prediction error correction method, device and equipment
CN116079727A (en) Humanoid robot motion simulation method and device based on 3D human body posture estimation
CN115907041A (en) Model training method and device
KR20220122862A (en) Apparatus and Method for Providing a Surgical Environment based on a Virtual Reality
CN110046457B (en) Human body model control method and device, electronic equipment and storage medium
CN115879524A (en) Model training method and related equipment thereof
KR20220066535A (en) Method, Server and System for Recognizing Motion in Video

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination