CN114330561B - A method and device for optimizing sensor deployment based on motor neuron diseases - Google Patents
A method and device for optimizing sensor deployment based on motor neuron diseases Download PDFInfo
- Publication number
- CN114330561B CN114330561B CN202111651576.0A CN202111651576A CN114330561B CN 114330561 B CN114330561 B CN 114330561B CN 202111651576 A CN202111651576 A CN 202111651576A CN 114330561 B CN114330561 B CN 114330561B
- Authority
- CN
- China
- Prior art keywords
- data
- sensor
- motion gesture
- gesture data
- candidate
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 208000005264 motor neuron disease Diseases 0.000 title 1
- 230000033001 locomotion Effects 0.000 claims abstract description 159
- 238000012706 support-vector machine Methods 0.000 claims abstract description 67
- 230000035945 sensitivity Effects 0.000 claims abstract description 42
- 201000010099 disease Diseases 0.000 claims abstract description 36
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims abstract description 36
- 210000005036 nerve Anatomy 0.000 claims abstract description 25
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 238000012549 training Methods 0.000 claims abstract description 16
- 230000002068 genetic effect Effects 0.000 claims abstract description 12
- 210000003141 lower extremity Anatomy 0.000 claims description 17
- 210000001364 upper extremity Anatomy 0.000 claims description 17
- 238000001914 filtration Methods 0.000 claims description 16
- 230000008859 change Effects 0.000 claims description 12
- 230000015654 memory Effects 0.000 claims description 12
- 230000009471 action Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012163 sequencing technique Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 5
- 210000000689 upper leg Anatomy 0.000 claims description 5
- 244000309466 calf Species 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 claims description 4
- 208000012902 Nervous system disease Diseases 0.000 claims description 2
- 208000025966 Neurological disease Diseases 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 14
- 238000010586 diagram Methods 0.000 description 9
- 238000011176 pooling Methods 0.000 description 9
- 210000002414 leg Anatomy 0.000 description 7
- 208000006011 Stroke Diseases 0.000 description 6
- 238000013527 convolutional neural network Methods 0.000 description 6
- 230000002146 bilateral effect Effects 0.000 description 5
- 230000001133 acceleration Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000013145 classification model Methods 0.000 description 3
- 230000000295 complement effect Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000007774 longterm Effects 0.000 description 3
- 230000035772 mutation Effects 0.000 description 3
- 230000001502 supplementing effect Effects 0.000 description 3
- 206010008190 Cerebrovascular accident Diseases 0.000 description 2
- 206010019468 Hemiplegia Diseases 0.000 description 2
- 238000005452 bending Methods 0.000 description 2
- 230000002490 cerebral effect Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 230000005358 geomagnetic field Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 208000027089 Parkinsonian disease Diseases 0.000 description 1
- 206010034010 Parkinsonism Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000001684 chronic effect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 201000001119 neuropathy Diseases 0.000 description 1
- 230000007823 neuropathy Effects 0.000 description 1
- 208000033808 peripheral neuropathy Diseases 0.000 description 1
- 229920000728 polyester Polymers 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention provides a method and a device for optimizing sensor layout based on motor nerve diseases, wherein the method for optimizing sensor layout based on motor nerve diseases comprises the steps of arranging sensors at concerned positions corresponding to target motor nerve diseases, collecting motion gesture data of the concerned positions, training a support vector machine model according to the collected motion gesture data, utilizing the trained support vector machine model to obtain sensitivity of each sensor corresponding to the motion gesture data, acquiring a candidate sensor set according to the sensitivity of each sensor, retraining the trained support vector machine model according to new motion gesture data collected by the candidate sensor set, updating the candidate sensor set by utilizing a genetic algorithm, and obtaining the candidate sensor set with optimal output result accuracy of the retrained support vector machine model and the minimum number of the candidate sensor sets to obtain an optimized sensor set. The motion gesture detection efficiency can be improved.
Description
Technical Field
The invention relates to the technical field of motion detection, in particular to a method and a device for optimizing sensor layout based on motor nerve diseases.
Background
Along with the increasing demands of people for healthy life, various chronic motor nerve diseases, such as post-stroke hemiplegia, become important influencing factors for influencing the quality of life. Therefore, the rehabilitation scheme is adjusted according to the detection result by carrying out long-term motion posture tracking detection on the patient, so that the rehabilitation scheme becomes an efficient means for rehabilitation of patients with motor nerve diseases.
The existing motion gesture detection technology mainly comprises two types, namely an optical motion capturing technology and a motion gesture detection technology based on an inertial detection unit (IMU, inertial Measurement Unit). The former records the spatial variation of each part of the patient, and the positioning is careful, but the equipment is complex, a fixed field is required, and the device is not suitable for long-term measurement. In contrast, the latter is a wearable device, which can measure motion gestures without additional devices, and is more suitable for long-term tracking detection of patients. However, in the existing IMU motion gesture detection system, a sensor needs to be placed for each suspected part, so that the number of the laid sensors is large, the data volume is large, the processing time is long, the motion gesture detection efficiency is reduced, long wearing time is required, and the use process is complex.
Disclosure of Invention
Therefore, the invention aims to provide a method and a device for optimizing sensor layout based on motor nerve diseases so as to improve the detection efficiency of the movement posture.
In a first aspect, an embodiment of the present invention provides a method for optimizing sensor layout based on motor neurological diseases, including:
Arranging a sensor at a concerned position corresponding to the target motor nerve disease, and collecting the motion gesture data of the concerned position;
Training a support vector machine model according to the acquired motion gesture data, and acquiring the sensitivity of each sensor corresponding to the motion gesture data by using the trained support vector machine model;
Acquiring a candidate sensor set according to the sensitivity of each sensor;
retraining the trained support vector machine model according to the new motion gesture data acquired by the candidate sensor set;
and updating the candidate sensor set by using a genetic algorithm, and obtaining the candidate sensor set with the optimal accuracy of the output result of the retrained support vector machine model and the minimum number of the candidate sensor sets to obtain an optimized sensor set.
With reference to the first aspect, an embodiment of the present invention provides a first possible implementation manner of the first aspect, where the training a support vector machine model according to the collected motion gesture data includes:
Filtering the collected motion gesture data by using a high-pass filter to obtain angle data taking 0 as a reference;
For each action, acquiring peak-to-peak data of each angle data according to the angle data corresponding to each sensor of the action;
And carrying out normalization processing on the peak-to-peak value data to obtain normalized data in the range of 0-1, and training a support vector machine model by using the normalized data.
With reference to the first possible implementation manner of the first aspect, an embodiment of the present invention provides a second possible implementation manner of the first aspect, where the acquiring, by using a trained support vector machine model, sensitivity of each sensor corresponding to motion gesture data includes:
Sequentially changing the value of the motion gesture data in the motion gesture data set aiming at each motion gesture data in the motion gesture data set input into the support vector machine model to obtain the motion gesture data set with each changed value;
inputting the motion gesture data set of each change value into a trained support vector machine model to obtain a tag value of the motion gesture data set of the change value;
And calculating the difference value between the maximum value and the minimum value in the label value according to the motion gesture data of each change value, and calculating the ratio of the difference value to the maximum value to obtain the sensitivity of the motion gesture data.
With reference to the second possible implementation manner of the first aspect, the embodiment of the present invention provides a third possible implementation manner of the first aspect, wherein the obtaining a candidate sensor set according to the sensitivity of each sensor includes:
sequencing the sensitivity of each obtained motion gesture data according to the size from large to small, and selecting candidate motion gesture data from the sequencing;
and acquiring a sensor to which the candidate motion gesture data belong to, and obtaining a candidate sensor set.
With reference to the first aspect, the first possible implementation manner of the first aspect, or any one of the first possible implementation manner to the third possible implementation manner of the first aspect, the embodiment of the present invention provides a fourth possible implementation manner of the first aspect, where the method further includes:
Wearing each optimizing sensor in the optimizing sensor set at a corresponding part of the monitoring object to acquire gesture data of the moving object;
performing high-pass filtering and normalization on the gesture data of the moving object;
Grouping the normalized moving object posture data to obtain grouping data, inputting the grouping data into a retraining support vector machine, and obtaining a label value corresponding to the normalized moving object posture data.
With reference to the fourth possible implementation manner of the first aspect, the embodiment of the present invention provides a fifth possible implementation manner of the first aspect, wherein, after the moving object pose data is acquired, before the moving object pose data is subjected to high-pass filtering, the method further includes:
If the left side and the right side of the same part are both provided with the optimizing sensors, the dynamic time warping algorithm is utilized to align the gesture data of the moving object collected by the optimizing sensors at the two sides.
With reference to the fourth possible implementation manner of the first aspect, an embodiment of the present invention provides a sixth possible implementation manner of the first aspect, where the grouping the normalized moving object pose data to obtain grouping data includes:
Dividing the optimized sensors into three groups of upper limbs, lower limbs and a trunk, wherein the trunk comprises an upper chest, a crotch, the upper limbs comprise large arms, small arms and the lower limbs comprise thighs, calves and insteps;
and classifying normalized moving object posture data corresponding to each grouped sensor into one group.
In a second aspect, an embodiment of the present invention further provides a device for optimizing sensor layout based on motor nerve diseases, including:
The data acquisition module is used for arranging a sensor at a concerned position corresponding to the target motor nerve disease and acquiring the motion gesture data of the concerned position;
the sensitivity acquisition module is used for training a support vector machine model according to the acquired motion gesture data and acquiring the sensitivity of each sensor corresponding to the motion gesture data by utilizing the trained support vector machine model;
The primary screening module is used for acquiring a candidate sensor set according to the sensitivity of each sensor;
The retraining module is used for retraining the trained support vector machine model according to the new motion gesture data acquired by the candidate sensor set;
and the optimization module is used for updating the candidate sensor set by utilizing a genetic algorithm, obtaining the candidate sensor set with the optimal accuracy of the output result of the retrained support vector machine model and the minimum number of the candidate sensor sets, and obtaining the optimized sensor set.
In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the steps of the method described above when executing the computer program.
In a fourth aspect, embodiments of the present application provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method described above.
The method and the device for optimizing sensor layout based on the motor neuro diseases are provided, the sensor is arranged at the concerned position corresponding to the target motor neuro diseases, the motion gesture data of the concerned position are collected, the support vector machine model is trained according to the collected motion gesture data, the sensitivity of each sensor corresponding to the motion gesture data is obtained by using the trained support vector machine model, the candidate sensor set is obtained according to the sensitivity of each sensor, the trained support vector machine model is retrained according to the new motion gesture data collected by the candidate sensor set, the candidate sensor set is updated by using a genetic algorithm, the candidate sensor set with the optimal output result accuracy of the retrained support vector machine model and the minimum number of the candidate sensor sets is obtained, and the optimized sensor set is obtained. In this way, the sensor sensitivity is obtained by using the support vector machine model, the sensor is initially selected, and then the optimal sensor is determined from the initially selected sensor based on the genetic algorithm and the support vector machine model, so that the number of the laid sensors can be effectively reduced, the acquired motion gesture data volume is reduced, and the motion gesture detection efficiency is improved.
In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.
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 will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 shows a schematic flow chart of a method for optimizing sensor layout based on motor nerve diseases, which is provided by an embodiment of the invention;
fig. 2 shows a schematic diagram of a device structure for optimizing sensor layout based on motor nerve diseases according to an embodiment of the present invention;
FIG. 3 is a schematic view of a sensor layout part according to an embodiment of the present invention;
FIG. 4 shows a schematic diagram of a sensor structure provided by an embodiment of the present invention;
FIG. 5 shows a schematic diagram of sensor layout based on cerebral stroke optimization provided by an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a computer device 600 according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without making any inventive effort, are intended to be within the scope of the present invention.
For each motor neuropathy, the body part of interest is different, for example, parkinsonism focuses on hand shake and lower leg motor rhythms, post-stroke hemiplegia focuses on the motor pattern of the affected side, and in the part of interest, not every part of interest needs to be provided with a sensor. Therefore, the quantity of the sensors which are arranged at present can be effectively reduced by optimizing the sensors which are arranged at present, so that the time required by the motion gesture data processing is shortened, and the detection efficiency of the motion gesture is improved.
The embodiment of the invention provides a method and a device for optimizing sensor layout based on motor nerve diseases, and the method and the device are described below through the embodiment.
Fig. 1 shows a schematic flow chart of a method for optimizing sensor layout based on motor nerve diseases according to an embodiment of the invention. As shown in fig. 1, the method includes:
Step 101, arranging a sensor at a concerned part corresponding to a target motor nerve disease, and collecting motion posture data of the concerned part;
In the embodiment of the invention, taking a certain target motor nerve disease as an example, the concerned parts are respectively positioned at the rear side of the upper chest, the rear side of the crotch, the front sides of the two large arms, the front sides of the two small arms, the front sides of the two large legs, the front sides of the two small legs and the two insteps on the vamp. Thus, one sensor is worn at each site of interest, for a total of 12 sensors. As an alternative embodiment, a tri-axial gyroscope, a tri-axial accelerometer, and a tri-axial magnetometer are provided within each sensor.
In the embodiment of the invention, by wearing the sensor, the movement posture data of multiple movements of a wearer, such as walking movement posture data, squatting movement posture data, two-hand side lifting movement posture data, elbow bending movement posture data and the like, can be recorded. As an alternative embodiment, the motion gesture data is triaxial motion gesture data, and each sensor respectively acquires triaxial motion gesture data of a corresponding position.
In the embodiment of the invention, the acquired motion gesture data not only comprises the three-axis motion gesture data of the patient with the target motor nerve diseases at each concerned part, but also comprises the three-axis motion gesture data of the healthy person at each concerned part. The three-axis motion posture data of a part of the healthy person at each concerned part is used as test data, and the collected rest motion posture data are used as training data.
Step 102, training a support vector machine model according to the acquired motion gesture data, and acquiring the sensitivity of each sensor corresponding to the motion gesture data by using the trained support vector machine model;
in an embodiment of the present invention, as an optional embodiment, training a support vector machine model according to collected motion gesture data includes:
A11, filtering the acquired motion attitude data by using a high-pass filter to obtain angle data taking 0 as a reference;
In the embodiment of the invention, the motion gesture data is filtered to remove the direct current bias. As an alternative embodiment, taking the left leg and the right leg as examples, the angle data comprise axial rotation data along the bone, lifting angle data and plane swing data.
In the embodiment of the invention, the acceleration data collected by the triaxial gyroscope and the acceleration data collected by the triaxial accelerometer are subjected to Kalman complementary filtering to obtain a pitch angle and a roll angle, and the pitch angle and the roll angle are analyzed with geomagnetic field data collected by the triaxial magnetometer to obtain a triaxial attitude angle.
A12, for each action, acquiring peak-to-peak data of each angle data according to the angle data corresponding to each sensor of the action;
In the embodiment of the invention, the peak-to-peak value is the average variation amplitude of the angle data. For each motion, 12 sensors are arranged, the motion gesture data collected by each sensor is three-axis motion gesture data, and each axis motion gesture data corresponds to one peak value data, so that 12×3=36 peak value data are obtained for the motion.
In the embodiment of the invention, the motion gesture data acquired by each sensor are coaxial in time, namely, each sensor acquires the motion gesture data at the same time for each action, and simultaneously acquires data for a period of time. And extracting the maximum value point and the minimum value point of the motion gesture data of each axis, and finally taking the average value of peak values and peak values.
In the embodiment of the invention, the peak value corresponding to the motion gesture data can represent the motion range of the corresponding joint of the corresponding part, and can represent the motion mode of the tester. Meanwhile, through extracting the peak-to-peak data features corresponding to the motion gesture data of each axis, each three peak-to-peak data features represent one IMU, the input number of SVM models can be reduced as much as possible, and therefore the sensitivity of the features can be analyzed more clearly and conveniently.
A13, carrying out normalization processing on the peak-to-peak value data to obtain normalized data in the range of 0-1, and training a support vector machine model by using the normalized data.
In the embodiment of the invention, the input of the support vector machine (SVM, support Vector Machine) model is normalized data corresponding to 36 peak-to-peak data of each action, and the output is a label value of whether a disease exists or not. For example, a healthy person tag is 0 and a patient tag is 1.
In the embodiment of the invention, aiming at the motion gesture data acquired by each action, the motion gesture data corresponding to the action is taken as the input of a support vector machine model, the label value with the output between zero and one is acquired, and the trained support vector machine model is verified according to the test data.
In the embodiment of the present invention, as an optional embodiment, the acquiring sensitivity of each sensor corresponding to the motion gesture data by using a trained support vector machine model includes:
A21, sequentially changing the value of the motion gesture data in the motion gesture data set aiming at each motion gesture data in the motion gesture data set input into the support vector machine model to obtain the motion gesture data set with each changed value;
A22, inputting the motion gesture data set of each change value into a trained support vector machine model to obtain a tag value of the motion gesture data set of the change value;
In the embodiment of the invention, normalized data corresponding to 36 peak-to-peak data are taken as a motion gesture data set, each peak-to-peak data corresponds to a sub-feature, and for each sub-feature in the motion gesture data set, the value of the sub-feature is changed to obtain a motion gesture data set with a changed value. For example, for a sub-feature with a changed value, the values of the sub-feature may be respectively set to 0, 0.25, 0.5, 0.75, and 1, and other sub-features are unchanged, so that five motion gesture data sets corresponding to the sub-feature may be obtained, and the support vector machine model outputs five results (label values) respectively. Wherein the output result is a value between 0 and 1.
A23, calculating the difference value between the maximum value and the minimum value in the label value according to the motion gesture data of each change value, and calculating the ratio of the difference value to the maximum value to obtain the sensitivity of the motion gesture data.
In the embodiment of the invention, a sensor comprises three types of motion gesture data, so that the sensitivity of the sensor comprises the sensitivity corresponding to the three types of motion gesture data respectively.
Step 103, acquiring a candidate sensor set according to the sensitivity of each sensor;
In an embodiment of the present invention, as an optional embodiment, obtaining a candidate sensor set according to the sensitivity of each sensor includes:
A31, sequencing the sensitivity of each obtained motion gesture data according to the size from large to small, and selecting candidate motion gesture data from the sequencing;
In the embodiment of the invention, 36 motion gesture data correspond to 36 sub-features, and the sensitivities of the 36 sub-features are ordered. As an alternative embodiment, the sub-feature with the sensitivity at the first 20% is selected as the candidate sub-feature.
A32, acquiring a sensor to which the candidate motion gesture data belongs, and obtaining a candidate sensor set.
In the embodiment of the invention, the sensors corresponding to the sub-features with 20% of sensitivity are found, the number of the found m candidate sensors is assumed to be m, and the found m candidate sensors are marked as an initial sensor group (candidate sensor set) S1.
In the embodiment of the invention, a plurality of candidate motion gesture data may belong to the same sensor.
Step 104, retraining the trained support vector machine model according to the new motion gesture data acquired by the candidate sensor set;
In the embodiment of the invention, according to the characteristics (3 m) corresponding to the initial sensor group S1, the motion gesture data (new motion gesture data) which are acquired by the initial sensor group S1 and are not used for supporting the training of the vector machine are taken as input, and the SVM model is retrained.
And 105, updating the candidate sensor set by using a genetic algorithm, and obtaining the candidate sensor set with the optimal accuracy of the output result of the retrained support vector machine model and the minimum number of the candidate sensor sets to obtain an optimized sensor set.
In the embodiment of the present invention, as an optional embodiment, updating a candidate sensor set by using a genetic algorithm to obtain a candidate sensor set with the best accuracy of an output result of a retrained support vector machine model and the least number of candidate sensor sets, so as to obtain an optimized sensor set, including:
a41, randomly selecting a predetermined number of candidate sensors from the candidate sensor set to obtain a candidate sensor group;
In the embodiment of the present invention, k candidate sensors S1, S2 are randomly selected from the initial sensor group S1, sk, and the candidate sensor group P (0) = { S1, S2, sk in the genetic algorithm is formed. Let iteration number t=0, maximum iteration number t=500.
In the embodiment of the present invention, as an alternative embodiment, the number of candidate sensors in the initial sensor set S1 is greater than 5, and the number of candidate sensors in the candidate sensor group should be not less than 3.
A42, aiming at each candidate sensor in the candidate sensor group, obtaining the fitness value of the candidate sensor according to a support vector machine classification model, and obtaining the fitness of the candidate sensor based on the fitness value of the candidate sensor and the fitness value of each candidate sensor in the candidate sensor group;
In the embodiment of the invention, iterative operation is performed to calculate the fitness. For each candidate sensor in the candidate sensor population P (t), an SVM classification model is trained separately, and a corresponding fitness value f (s 1), f (s 2), f (sk) is calculated. The fitness of the candidate sensor si (i=1, 2,..k) is f (si)/(f (s 1) +f (s 2) +f (sm)).
A43, reserving a threshold number of candidate sensors before the fitness in the candidate sensor group, deleting other candidate sensors, supplementing the candidate sensors in the candidate sensor group based on a preset crossover and mutation algorithm, and carrying out the next iteration on the supplemented candidate sensor group;
In the embodiment of the invention, as an alternative embodiment, the candidate sensors with the fitness of the first 30% are reserved, n in total, and the rest (k-n) are discarded. And then, supplementing (k-n) candidate sensors from the candidate sensor set through crossover and mutation operation to obtain an iterated candidate sensor group P (t+1).
In an embodiment of the present invention, as an optional embodiment, based on a preset crossover and mutation algorithm, the method for supplementing candidate sensors in a candidate sensor group includes:
The fitness is taken as the probability of each candidate sensor being selected, two candidate sensors are randomly selected each time according to the probability, 50% of the sensors are taken from the two candidate sensors, and the two candidate sensors are combined into a new sensor group scross, namely a crossing result. Then randomly extracting scross candidate sensors, and replacing the 2 candidate sensors with any two candidate sensors except scross to obtain a mutated result. This operation is performed (k-n) times in total, generating (k-n) new candidate sensors, with the original n candidate sensor groups of candidate sensors.
And A44, if the iteration number reaches a preset iteration number threshold value, or the difference value between the maximum fitness of the candidate sensors in the candidate sensor group and the maximum fitness of the candidate sensors in the candidate sensor group of the previous iteration is not greater than a preset error threshold value, acquiring the candidate sensor group of the iteration as an optimized sensor set.
In the embodiment of the invention, if t=t, or the maximum fitness no longer rises, the candidate sensor group with the maximum current fitness is used as the output of the optimized sensor set S2, and the calculation is terminated, otherwise, let t=t+1, and execute step a42.
In the embodiment of the invention, the problem of optimizing the sensor can be expressed as follows:
s.t.num(s)≥3,S∈D
where s is the solution of the problem, i.e. optimizing the sensor set.
D is the feasible domain, i.e. the candidate sensor set.
Acc SVM is the accuracy of the SVM classification model.
Num is the number of candidate sensors in the optimal sensor set or candidate sensor set.
In the embodiment of the invention, the optimized target f(s) is the ratio of the number of the sensors in the optimized sensor set to the number of all the sensors in the candidate sensor set is minimized while the classification accuracy is highest.
In the embodiment of the invention, the initial sensor group is combined and optimized, the optimized initial sensor group is S1, the S1 is updated by utilizing a genetic algorithm, the aim is to ensure that the classification result accuracy of the SVM model is highest, and the number of sensors in the S1 is minimum. In this way, an optimized sensor set is obtained according to the sensor which has the optimal classification result accuracy of the SVM model and the least number of candidate sensors.
In the embodiment of the invention, after optimization, the optimal sensor set aiming at the target motor nerve disease is finally obtained.
In an embodiment of the present invention, as an optional embodiment, the method further includes:
A51, wearing each optimizing sensor in the optimizing sensor set at a corresponding part of the monitoring object to acquire gesture data of the moving object;
a52, performing high-pass filtering and normalization on the gesture data of the moving object;
In the embodiment of the invention, the offset is removed by high-pass filtering on the posture data of each axis of moving object acquired by the optimizing sensor.
In an embodiment of the present invention, as an optional embodiment, after the moving object pose data is collected, before the moving object pose data is subjected to high-pass filtering, the method further includes:
if the optimized sensors are arranged on the left side and the right side of the same part, the motion object posture data collected by the optimized sensors on the two sides are aligned by utilizing a dynamic time warping (DTW, dynamic Time Warping) algorithm, so that the complexity of data processing is reduced.
And A53, grouping the normalized moving object posture data to obtain grouping data, and inputting the grouping data into a retraining support vector machine to obtain a label value corresponding to the normalized moving object posture data.
In the embodiment of the present invention, as an optional embodiment, grouping normalized gesture data of a moving object to obtain grouped data includes:
Dividing the optimized sensors into three groups of upper limbs, lower limbs and a trunk, wherein the trunk comprises an upper chest, a crotch, the upper limbs comprise large arms, small arms and the lower limbs comprise thighs, calves and insteps;
and classifying normalized moving object posture data corresponding to each grouped sensor into one group.
In the embodiment of the invention, the data contained in the packet is taken as an input data unit for inputting the support vector machine.
In the embodiment of the invention, when the motion gesture is monitored, time sequence data acquired by the optimized sensors are utilized, for example, for m optimized sensors, the sampling time is t, the sampling rate is f, then the total acquired 3*m data sequences are acquired, the length of each data sequence is t f, the acquired data sequences are grouped and then directly input into the support vector machine model, and the information contained in the data sequences can be comprehensively analyzed. For example, peak-to-peak information, stability information, symmetry information, and the like included in the data sequence.
In the embodiment of the invention, the data with stronger association is grouped into one group by grouping the data sequences, so that the data in the group has strong association and the data between the groups has weak association, the support vector machine model structure can be simplified, the operation amount is reduced, and the accuracy of the output result is improved.
In the embodiment of the present invention, as an optional embodiment, inputting the packet data into a retrained support vector machine to obtain a tag value corresponding to the normalized motion object pose data, where the tag value includes:
And extracting features of each group of data through a one-dimensional CNN network, and combining the extracted features corresponding to each group of data in a fully-connected network to obtain a tag value corresponding to the normalized moving object posture data.
In the embodiment of the invention, in each group of data, three-axis time sequence data of all sensors are combined into one group of input, the features are extracted through a one-dimensional CNN network respectively, and the three extracted groups of features are combined in a fully connected network, so that a disease diagnosis result is finally obtained.
In the embodiment of the present invention, as an optional embodiment, for each packet data, features are extracted through a one-dimensional CNN network, and features corresponding to each extracted packet data are combined in a fully connected network, so as to obtain a tag value corresponding to normalized motion object pose data, where the tag value includes:
combining three-axis time sequence data of all sensors in the upper limb group to obtain an upper limb posture data input group, inputting the upper limb posture data input group into an SVM model, obtaining upper limb first convolution characteristics through operation of a first convolution layer, inputting the upper limb first convolution characteristics into a second convolution layer to obtain upper limb second convolution characteristics, and inputting the upper limb second convolution characteristics into a pooling layer to obtain upper limb pooling characteristics;
Combining three-axis time sequence data of all sensors in a lower limb group to obtain a lower limb posture data input group, inputting the lower limb posture data input group into an SVM model, obtaining a first lower limb convolution characteristic through operation of a first convolution layer, inputting the first lower limb convolution characteristic into a second convolution layer to obtain a second lower limb convolution characteristic, and inputting the second lower limb convolution characteristic into a pooling layer to obtain a lower limb pooling characteristic;
combining three-axis time sequence data of all sensors in a trunk group to obtain a trunk gesture data input group, inputting the trunk gesture data into an SVM model, obtaining a trunk first convolution feature through operation of a first convolution layer, inputting the trunk first convolution feature into a second convolution layer to obtain a trunk second convolution feature, and inputting the trunk second convolution feature into a pooling layer to obtain a trunk pooling feature;
and inputting the upper limb pooling feature, the lower limb pooling feature and the trunk pooling feature into a full-connection layer to obtain a label value corresponding to the normalized moving object posture data.
Fig. 2 shows a schematic diagram of a device structure for optimizing sensor layout based on motor nerve diseases according to an embodiment of the present invention. As shown in fig. 2, the apparatus includes:
the data acquisition module 201 is used for arranging a sensor at a concerned part corresponding to the target motor nerve disease and acquiring the motion gesture data of the concerned part;
Fig. 3 shows a schematic diagram of a sensor layout part provided by an embodiment of the invention. As shown in fig. 3, the regions of interest are located on the rear side of the upper chest, the rear side of the crotch, the front side of the bilateral large arms, the front side of the bilateral small arms, the front side of the bilateral large legs, the front side of the bilateral small legs, and the bilateral instep on the upper, respectively, in the embodiment of the present invention.
In the embodiment of the invention, the motion gesture data is triaxial motion gesture data, including but not limited to walking motion gesture data, squatting motion gesture data, double-hand side lifting motion gesture data, elbow bending motion gesture data and the like.
The sensitivity acquisition module 202 is configured to train a support vector machine model according to the acquired motion gesture data, and acquire the sensitivity of each sensor corresponding to the motion gesture data by using the trained support vector machine model;
In an embodiment of the present invention, as an optional embodiment, the sensitivity obtaining module 202 includes:
A filtering unit (not shown in the figure) for filtering the collected motion gesture data by using a high-pass filter to obtain angle data taking 0 as a reference;
The peak-to-peak value calculation unit is used for acquiring peak-to-peak value data of each angle data according to the angle data corresponding to each sensor of each action;
The training unit is used for carrying out normalization processing on the peak-to-peak value data to obtain normalized data in the range of 0-1, and training a support vector machine model by using the normalized data;
the data adjusting unit is used for sequentially changing the value of the motion gesture data in the motion gesture data set aiming at each motion gesture data in the motion gesture data set input into the support vector machine model to obtain the motion gesture data set with each changed value;
The tag value output unit is used for inputting the motion gesture data set of each change value into a trained support vector machine model to obtain the tag value of the motion gesture data set of the change value;
And the sensitivity calculation unit is used for calculating the difference value between the maximum value and the minimum value in the label value according to the motion gesture data of each change value, and calculating the ratio of the difference value to the maximum value to obtain the sensitivity of the motion gesture data.
A primary screening module 203, configured to obtain a candidate sensor set according to the sensitivity of each sensor;
in an embodiment of the present invention, as an optional embodiment, the primary screening module 203 includes:
A sorting unit (not shown in the figure) for sorting the sensitivities of the obtained motion gesture data according to the order from large to small, and selecting candidate motion gesture data from the sorting;
and the primary screening unit is used for acquiring the sensor to which the candidate motion gesture data belong to obtain a candidate sensor set.
The retraining module 204 is configured to retrain the trained support vector machine model according to the new motion gesture data collected by the candidate sensor set;
And the optimizing module 205 is configured to update the candidate sensor set by using a genetic algorithm, obtain the candidate sensor set with the best accuracy of the output result of the retrained support vector machine model and the least number of candidate sensor sets, and obtain the optimized sensor set.
In an embodiment of the present invention, as an optional embodiment, the apparatus further includes:
The system comprises a gesture monitoring module (not shown in the figure) for wearing each optimized sensor in the optimized sensor set at a corresponding part of a monitored object to acquire gesture data of the moving object, performing high-pass filtering and normalization on the gesture data of the moving object, grouping the gesture data of the normalized moving object to obtain grouping data, inputting the grouping data into a retraining support vector machine, and obtaining a label value corresponding to the gesture data of the normalized moving object.
In an embodiment of the present invention, as another optional embodiment, the apparatus further includes:
And the data alignment module aligns the gesture data of the moving object acquired by the optimizing sensors at the two sides by utilizing a dynamic time warping algorithm if the optimizing sensors are arranged at the left side and the right side of the same part.
In the embodiment of the present invention, as an optional embodiment, grouping normalized gesture data of a moving object to obtain grouped data includes:
Dividing the optimized sensors into three groups of upper limbs, lower limbs and a trunk, wherein the trunk comprises an upper chest, a crotch, the upper limbs comprise large arms, small arms and the lower limbs comprise thighs, calves and insteps;
and classifying normalized moving object posture data corresponding to each grouped sensor into one group.
Fig. 4 shows a schematic diagram of a sensor structure according to an embodiment of the present invention. As shown in fig. 4, the sensor includes an IMU module 41, a clip 42, an elastic strap 43, wherein,
The elastic band 43 comprises an elastic band (not shown) woven by polyester yarns and elastic bands, and plastic buckles at two ends of the elastic band;
In the embodiment of the invention, the length of the elastic binding band can be adjusted through the plastic buckle, the elastic band has certain elasticity, and the IMU module 41 can be firmly placed on the body to prevent the IMU module from sliding.
The clips 42 are metal clips for use in combination with plastic clips of the elastic straps 43 to secure the IMU module 41 to the elastic straps of the elastic straps 43;
the IMU module 41 is a sensor circuit encapsulated by a plastic housing that is attached to the clip 42.
In an embodiment of the present invention, the IMU module 41 includes a data storage module, a bluetooth module, a central processing unit, and an inertial sensor module (not shown). Wherein,
The inertial sensing module is used for acquiring motion gesture data and comprises a triaxial gyroscope, a triaxial accelerometer and a triaxial magnetometer.
In the embodiment of the invention, after the inertial sensing module collects motion gesture data, a triaxial gesture angle is obtained through classical Kalman (Kalman) complementary filtering and gesture resolving.
In the embodiment of the invention, the central processing unit is an STM32 chip and is used for generating control instructions to control the working processes of all modules, and the data analysis module is used for processing data.
The data analysis module is used for carrying out data preprocessing and calculating parameters, and obtaining tag values through a convolutional neural network (CNN, convolutional Neural Network) network.
And the Bluetooth module is used for data transmission among different sensors, is in wireless connection with an external terminal and transmits data for analysis.
And the data storage module is used for storing the data and the tag value.
In the embodiment of the invention, the acceleration data collected by the triaxial gyroscope and the acceleration data collected by the triaxial accelerometer are subjected to Kalman complementary filtering to obtain a pitch angle and a roll angle, and the pitch angle and the roll angle are analyzed with geomagnetic field data collected by the triaxial magnetometer to obtain a triaxial attitude angle.
In the embodiment of the invention, a stroke is taken as an example for specific explanation. The exercise symptoms of the cerebral apoplexy patient comprise abnormal dorsiflexion range of the foot step, hand hurdle exercise, abrupt extension of the knee in the support phase, stiff legs and circle exercise. In the embodiment of the invention, after the sensor layout based on the sensor layout shown in fig. 2 is optimized, the obtained sensor layout schematic diagram is shown in fig. 5, and fig. 5 shows the sensor layout schematic diagram based on cerebral apoplexy optimization provided by the embodiment of the invention. Seven optimal sensor layout positions shown in figures 1-7 are obtained and are respectively positioned at the upper chest, the lower arm of the affected side, the thigh on the two sides, the lower leg on the two sides and the instep on the affected side. By arranging the sensors at the positions, the acquired motion gesture data has stronger distinguishing degree between healthy people and patients, the motion gesture tracking can be well carried out, the number of the arranged sensors is small, the acquired motion gesture data quantity can be effectively reduced, and therefore the motion gesture detection efficiency is improved. Further, the sensor is wearable, small in size and good in portability, does not influence daily life of a wearer, and can be used for a long time. In addition, the number of sensors is optimized, the wearing is more convenient and simple, the discomfort is light, the complexity is reduced, the cost is reduced, and the endurance time is prolonged. In addition, based on the label value output by the neural network, the rehabilitation effect can be determined through the comparison of the front label value and the rear label value, and a reference is provided for the adjustment of the rehabilitation scheme.
As shown in fig. 6, an embodiment of the present application provides a computer device 600 for performing the method for optimizing sensor layout based on motor-based diseases in fig. 1, where the device includes a memory 601, a processor 602 connected to the memory 601 through a bus, and a computer program stored on the memory 601 and capable of running on the processor 602, where the steps of the method for optimizing sensor layout based on motor-based diseases are implemented when the processor 602 executes the computer program.
Specifically, the above-mentioned memory 601 and processor 602 can be general-purpose memories and processors, and are not particularly limited herein, and the above-mentioned method of optimizing sensor layout based on motor-nerve-like diseases can be performed when the processor 602 runs a computer program stored in the memory 601.
Corresponding to the method for optimizing sensor layout based on motor-based diseases in fig. 1, the embodiment of the application further provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, performs the steps of the method for optimizing sensor layout based on motor-based diseases.
In particular, the storage medium can be a general-purpose storage medium, such as a mobile magnetic disk, a hard disk, etc., and the computer program on the storage medium can execute the method for optimizing sensor layout based on motor neuro diseases when being executed.
In the embodiments provided herein, it should be understood that the disclosed systems and methods may be implemented in other ways. The system embodiments described above are merely illustrative, e.g., the division of the elements is merely a logical functional division, and there may be additional divisions in actual implementation, and e.g., multiple elements 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 with each other may be through some communication interface, system or unit indirect coupling or communication connection, which may be in electrical, mechanical or other form.
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.
In addition, each functional unit in the embodiments provided in the present application 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.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. The storage medium includes a U disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, an optical disk, or other various media capable of storing program codes.
It should be noted that like reference numerals and letters refer to like items in the following figures, and thus, once an item is defined in one figure, no further definition or explanation of that in the following figures is necessary, and furthermore, the terms "first," "second," "third," etc. are used merely to distinguish between descriptions and are not to be construed as indicating or implying relative importance.
It should be noted that the foregoing embodiments are merely illustrative embodiments of the present application, and not restrictive, and the scope of the application is not limited to the embodiments, and although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that any modification, variation or substitution of some of the technical features of the embodiments may be made within the technical scope of the present application disclosed in the present application, and the spirit, the scope and the scope of the technical aspects of the embodiments do not deviate from the spirit and scope of the technical aspects of the embodiments. Are intended to be encompassed within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A method for optimizing sensor layout based on motor nerve diseases, comprising the steps of:
Arranging a sensor at a concerned position corresponding to the target motor nerve disease, and collecting the motion gesture data of the concerned position;
Training a support vector machine model according to the acquired motion gesture data, and acquiring the sensitivity of each sensor corresponding to the motion gesture data by using the trained support vector machine model;
Acquiring a candidate sensor set according to the sensitivity of each sensor;
retraining the trained support vector machine model according to the new motion gesture data acquired by the candidate sensor set;
and updating the candidate sensor set by using a genetic algorithm, and obtaining the candidate sensor set with the optimal accuracy of the output result of the retrained support vector machine model and the minimum number of the candidate sensor sets to obtain an optimized sensor set.
2. The method of claim 1, wherein training a support vector machine model based on the collected motion pose data comprises:
Filtering the collected motion gesture data by using a high-pass filter to obtain angle data taking 0 as a reference;
For each action, acquiring peak-to-peak data of each angle data according to the angle data corresponding to each sensor of the action;
And carrying out normalization processing on the peak-to-peak value data to obtain normalized data in the range of 0-1, and training a support vector machine model by using the normalized data.
3. The method of claim 2, wherein the acquiring the sensitivity of each sensor corresponding to the motion gesture data using the trained support vector machine model comprises:
Sequentially changing the value of the motion gesture data in the motion gesture data set aiming at each motion gesture data in the motion gesture data set input into the support vector machine model to obtain the motion gesture data set with each changed value;
inputting the motion gesture data set of each change value into a trained support vector machine model to obtain a tag value of the motion gesture data set of the change value;
And calculating the difference value between the maximum value and the minimum value in the label value according to the motion gesture data of each change value, and calculating the ratio of the difference value to the maximum value to obtain the sensitivity of the motion gesture data.
4. A method according to claim 3, wherein said obtaining a candidate set of sensors based on the sensitivity of each sensor comprises:
sequencing the sensitivity of each obtained motion gesture data according to the size from large to small, and selecting candidate motion gesture data from the sequencing;
and acquiring a sensor to which the candidate motion gesture data belong to, and obtaining a candidate sensor set.
5. The method according to any one of claims 1 to 4, further comprising:
Wearing each optimizing sensor in the optimizing sensor set at a corresponding part of the monitoring object to acquire gesture data of the moving object;
performing high-pass filtering and normalization on the gesture data of the moving object;
Grouping the normalized moving object posture data to obtain grouping data, inputting the grouping data into a retraining support vector machine, and obtaining a label value corresponding to the normalized moving object posture data.
6. The method of claim 5, wherein after the acquisition of the moving object pose data, the method further comprises, prior to high pass filtering the moving object pose data:
If the left side and the right side of the same part are both provided with the optimizing sensors, the dynamic time warping algorithm is utilized to align the gesture data of the moving object collected by the optimizing sensors at the two sides.
7. The method of claim 5, wherein grouping the normalized moving object pose data to obtain grouped data comprises:
Dividing the optimized sensors into three groups of upper limbs, lower limbs and a trunk, wherein the trunk comprises an upper chest, a crotch, the upper limbs comprise large arms, small arms and the lower limbs comprise thighs, calves and insteps;
and classifying normalized moving object posture data corresponding to each grouped sensor into one group.
8. A motor nerve disease based sensor layout optimizing device, comprising:
The data acquisition module is used for arranging a sensor at a concerned position corresponding to the target motor nerve disease and acquiring the motion gesture data of the concerned position;
the sensitivity acquisition module is used for training a support vector machine model according to the acquired motion gesture data and acquiring the sensitivity of each sensor corresponding to the motion gesture data by utilizing the trained support vector machine model;
The primary screening module is used for acquiring a candidate sensor set according to the sensitivity of each sensor;
The retraining module is used for retraining the trained support vector machine model according to the new motion gesture data acquired by the candidate sensor set;
and the optimization module is used for updating the candidate sensor set by utilizing a genetic algorithm, obtaining the candidate sensor set with the optimal accuracy of the output result of the retrained support vector machine model and the minimum number of the candidate sensor sets, and obtaining the optimized sensor set.
9. A computer device comprising a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory in communication via the bus when the computer device is in operation, the machine-readable instructions when executed by the processor performing the steps of the method for optimizing sensor placement based on motor neurological disorders of any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that it has stored thereon a computer program which, when being executed by a processor, performs the steps of the method for optimizing sensor layout based on motor-nerve-like diseases according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111651576.0A CN114330561B (en) | 2021-12-30 | 2021-12-30 | A method and device for optimizing sensor deployment based on motor neuron diseases |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111651576.0A CN114330561B (en) | 2021-12-30 | 2021-12-30 | A method and device for optimizing sensor deployment based on motor neuron diseases |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114330561A CN114330561A (en) | 2022-04-12 |
| CN114330561B true CN114330561B (en) | 2025-03-11 |
Family
ID=81018323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111651576.0A Active CN114330561B (en) | 2021-12-30 | 2021-12-30 | A method and device for optimizing sensor deployment based on motor neuron diseases |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114330561B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117009826B (en) * | 2023-08-30 | 2024-04-02 | 北京长木谷医疗科技股份有限公司 | A method and device for adaptive determination of acetabular cup angle based on deep learning |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111368970A (en) * | 2020-02-17 | 2020-07-03 | 哈尔滨工业大学 | A sensor optimization method based on deep reinforcement learning |
| CN112784490A (en) * | 2021-01-25 | 2021-05-11 | 西安交通大学 | Sensor position optimization method of helicopter health management system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7259778B2 (en) * | 2003-07-01 | 2007-08-21 | L-3 Communications Corporation | Method and apparatus for placing sensors using 3D models |
| CN112818927A (en) * | 2021-02-26 | 2021-05-18 | 上海交通大学 | Real-time classification method and system for human body lower limb movement modes |
| CN113094853B (en) * | 2021-04-01 | 2024-07-02 | 上海天麦能源科技有限公司 | Optimized layout method and system for gas pipe network leakage positioning |
-
2021
- 2021-12-30 CN CN202111651576.0A patent/CN114330561B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111368970A (en) * | 2020-02-17 | 2020-07-03 | 哈尔滨工业大学 | A sensor optimization method based on deep reinforcement learning |
| CN112784490A (en) * | 2021-01-25 | 2021-05-11 | 西安交通大学 | Sensor position optimization method of helicopter health management system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114330561A (en) | 2022-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kristoffersson et al. | A systematic review of wearable sensors for monitoring physical activity | |
| Ravi et al. | A deep learning approach to on-node sensor data analytics for mobile or wearable devices | |
| Sztyler et al. | Online personalization of cross-subjects based activity recognition models on wearable devices | |
| Zubair et al. | Human activity recognition using wearable accelerometer sensors | |
| Yurtman et al. | Automated evaluation of physical therapy exercises using multi-template dynamic time warping on wearable sensor signals | |
| CN109875565A (en) | An automatic assessment method for upper limb motor function in stroke based on deep learning | |
| Jalloul et al. | Activity recognition using complex network analysis | |
| Ravı et al. | A deep learning approach to on-node sensor data analytics for mobile or wearable devices | |
| Giuberti et al. | Automatic UPDRS evaluation in the sit-to-stand task of Parkinsonians: Kinematic analysis and comparative outlook on the leg agility task | |
| García-de-Villa et al. | Inertial sensors for human motion analysis: A comprehensive review | |
| KR102302719B1 (en) | Apparatus and method for classification of gait type by performing neural network analysis for various detection information | |
| Nguyen et al. | IMU-based spectrogram approach with deep convolutional neural networks for gait classification | |
| Wang et al. | A selection framework of sensor combination feature subset for human motion phase segmentation | |
| Ponciano et al. | Smartphone-based automatic measurement of the results of the Timed-Up and Go test | |
| CN114330561B (en) | A method and device for optimizing sensor deployment based on motor neuron diseases | |
| Schuldhaus | Human activity recognition in daily life and sports using inertial sensors | |
| Cheng et al. | Periodic physical activity information segmentation, counting and recognition from video | |
| Sengupta et al. | A survey of wearable sensors and machine learning algorithms for automated stroke rehabilitation | |
| KR102280291B1 (en) | Apparatus and method for identify patients with parkinson's disease and patients with podarthritis by performing neural network analysis by various detection information | |
| KR102194313B1 (en) | Apparatus and method for identifying individuals by performing neural network analysis for various detection information | |
| Ali et al. | Detection and analysis of transitional activity in manifold space | |
| Yin et al. | Neural networks for pathological gait classification using wearable motion sensors | |
| KR20230041908A (en) | Apparatus and method for monitoring upper extremity rehabilitation based on deep learning | |
| Bagheri et al. | Comprehensive and Cognitive Approaches in Rehabilitation Design using IMUs for STEM Branch Providing Novel IMU-Based System Design via Machine Learning Algorithms | |
| KR20210084846A (en) | Method and system for artificial intelligence-based brain disease diagnosis using human dynamic characteristics information |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |