CN210077662U - Flexible gait monitoring device - Google Patents
Flexible gait monitoring device Download PDFInfo
- Publication number
- CN210077662U CN210077662U CN201821836259.XU CN201821836259U CN210077662U CN 210077662 U CN210077662 U CN 210077662U CN 201821836259 U CN201821836259 U CN 201821836259U CN 210077662 U CN210077662 U CN 210077662U
- Authority
- CN
- China
- Prior art keywords
- neural network
- sensing array
- pressure sensing
- flexible
- comparison module
- 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
- 230000005021 gait Effects 0.000 title claims abstract description 25
- 238000012806 monitoring device Methods 0.000 title claims abstract description 10
- 210000003127 knee Anatomy 0.000 claims abstract description 36
- 238000013528 artificial neural network Methods 0.000 claims abstract description 30
- 238000006243 chemical reaction Methods 0.000 claims abstract description 22
- 238000012544 monitoring process Methods 0.000 claims abstract description 10
- 230000036541 health Effects 0.000 claims abstract description 9
- 238000004458 analytical method Methods 0.000 abstract description 4
- 238000011160 research Methods 0.000 abstract description 4
- 230000003862 health status Effects 0.000 abstract description 2
- 210000002683 foot Anatomy 0.000 description 31
- 238000012549 training Methods 0.000 description 25
- 210000002569 neuron Anatomy 0.000 description 24
- 238000000034 method Methods 0.000 description 5
- 230000036544 posture Effects 0.000 description 3
- 230000005236 sound signal Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 210000003205 muscle Anatomy 0.000 description 2
- 206010027336 Menstruation delayed Diseases 0.000 description 1
- 210000001361 achilles tendon Anatomy 0.000 description 1
- 210000003423 ankle Anatomy 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000037237 body shape Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 210000004744 fore-foot Anatomy 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 210000000474 heel Anatomy 0.000 description 1
- 210000001624 hip Anatomy 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 230000004630 mental health Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- -1 periodo Chemical group 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
Images
Abstract
The utility model discloses a flexible gait monitoring device, which is characterized in that the monitoring device is formed by connecting a flexible pressure sensing array with a gating circuit, an analog-digital conversion circuit, a neural network trainer and a characteristic comparison module in series, the flexible pressure sensing array collects a plurality of groups of pressure signals of knees and soles of a human body during walking and accesses the gating circuit into the analog-digital conversion circuit, and output signals of the flexible pressure sensing array are transmitted to the neural network trainer through a serial port for learning; the neural network trainer inputs the trained neural network into the characteristic comparison module, and compares and identifies the characteristic comparison module with a pressure signal input by the flexible pressure sensing array in real time, so that the intelligent monitoring of human body form health is realized. Compared with the prior art, the utility model the rate of accuracy that has the prediction is high, simple structure, and convenient to use realizes the healthy effective intelligent monitoring of human form, carries out reliable analysis and research for human form and health status and provides reliable, effectual data support.
Description
Technical Field
The utility model belongs to the technical field of the intelligent monitoring technique and specifically relates to a flexible gait monitoring devices that artifical intelligent computing learnt based on FPGA.
Background
The gait is the behavior characteristic of human walking, and relates to the muscle and joint coordinated movement of feet, ankles, knees, hips, trunk, neck, shoulders and arms, and the human morphology and the human health state can be predicted by analyzing and researching the gait. The good walking habit helps people to keep physical and mental health, and irregular walking postures can aggravate pressure stimulation to triceps surae muscles and achilles tendons after years of years, thereby bringing harm to the body. Along with the rapid development of wearable technology, the gait monitoring device is miniaturized and wearable, plays an important role in the aspects of adolescent exercise posture formation, postoperative rehabilitation of patients and the like, and greatly improves the coverage rate of exercise foot posture correction training.
The patent application number is CN201610519761.7, the patent name is a gait parameter acquisition method and equipment, and discloses a gait data acquisition method, which comprises the following steps: acquiring sound signal curves of a left foot and a right foot; extracting the peak position of the sound signal curve for representing the node foot touchdown sound and the peak position of the non-node foot touchdown sound according to a peak detection algorithm, and then calculating the step distance Lsd of each step as VSound( t2-t1) (ii) a Wherein, VSoundSpeed of sound propagation in air, t1And t2Respectively, the time when the gait data acquisition devices fixed on different feet acquire the sound generated by the same single-foot touchdown, wherein t2To capture the time of non-nodal foot contact sound, t1To adoptThe time of the node foot touching the ground is collected, and the daily walking of the person is monitored by collecting walking sound signals.
The gait can be subdivided into eight stages, namely an initial landing stage, a support reaction stage, a midpoint support stage, a support later stage, a swing earlier stage, a swing early stage, a swing middle stage and a swing later stage. In different dynamic stages, the knee and foot of the human body have different shapes, and the knee bends 0 at the initial landing stageoHeel strike; knee flexion 0 in support reaction periodo-20oThe sole touches the ground, the gravity center gradually moves to the center of the foot, and the sole is parallel; mid-point support knee flexion 20o-0oThe foot center supports and the heel gradually lifts off; late support knee flexion 0oWhen the heel leaves the ground, the sole touches the ground; ante-swing knee flexion 0o-40oThe foot gradually lifts off the ground; the knee flexion degree in early swing stage is 40oBecomes 60oThe foot is off the ground; the knee flexion degree in the middle period of swing is 60oBecomes 30oThe foot is off the ground; knee flexion at end of swing 30o-0oThe knee and foot of the human body have different shapes in the eight gait phases, and the generated pressure is different.
In the prior art, the influence of feet on the gait is mostly concerned, and the influence of knees on the gait is not considered, so that the gait analysis and research reliability is poor, and the human body shape and the human body health state prediction is greatly influenced.
SUMMERY OF THE UTILITY MODEL
The utility model aims at the not enough and a flexible gait monitoring devices who designs of prior art, adopt flexible pressure sensing array collection human knee and plantar pressure signal when walking, multiunit pressure data are defeated to artificial intelligence learning system after circuit module handles, artificial intelligence trains and learns neural network, utilize the neural network that the training is successful to discern walking each stage to the human body, realize the healthy effective intelligent monitoring of human form, carry out reliable analysis and research for human form and health condition and provide reliably, effectual data support.
The purpose of the utility model is realized like this: a flexible gait monitoring device is characterized in that the monitoring device consists of a flexible pressure sensing array, a gating circuit, an analog-to-digital conversion circuit, a neural network trainer and a characteristic comparison module, wherein the flexible pressure sensing array is used for collecting a plurality of groups of pressure signals of knees and soles of a human body during walking and accessing the pressure signals into the analog-to-digital conversion circuit through the gating circuit, and output signals of the flexible pressure sensing array are transmitted to the neural network trainer for learning through a serial port; the neural network trainer inputs the trained neural network into the characteristic comparison module, and compares and identifies the characteristic comparison module with a pressure signal input by the flexible pressure sensing array in real time, so that the intelligent monitoring of human body form health is realized.
The flexible pressure sensing array is attached to the knee and the sole and used for collecting pressure data of eight gait stages, namely an initial landing stage, a support reaction stage, a midpoint support stage, a support later stage, a swing early stage, a swing middle stage and a swing later stage.
The neural network trainer is a three-layer network structure consisting of an input layer neuron, a hidden layer neuron and an output layer neuron, the input layer neuron accesses an output signal of the analog-to-digital conversion circuit into the hidden layer neuron, the hidden layer neuron enters the output layer neuron, the output of the input layer neuron is compared by a training parameter comparison unit, and a training result is output when the training requirement is met; otherwise, the training result is returned to the input layer neuron after the weight is corrected and enters the hidden layer neuron and the output layer neuron again through the input layer neuron, and the training of the multiple-cycle learning is carried out until the learning result meets the set training parameters.
Compared with the prior art, the utility model the rate of accuracy that has the prediction is high, simple structure, and convenient to use realizes the healthy effective intelligent monitoring of human form, carries out reliable analysis and research for human form and health status and provides reliable, effectual data support.
Drawings
FIG. 1 is a schematic structural view of the present invention;
FIG. 2 is a schematic diagram of a neural network trainer;
FIG. 3 is a schematic view of the knee and foot stress processes during a gait phase;
fig. 4 is an embodiment.
Detailed Description
Referring to the attached drawing 1, the utility model is composed of a flexible pressure sensing array 1, a gating circuit 2, an analog-to-digital conversion circuit 3, a neural network trainer 4 and a characteristic comparison module 5, the flexible pressure sensing array 1 collects a plurality of groups of pressure signals of knees and soles of a human body during walking and accesses the analog-to-digital conversion circuit 3 through the gating circuit 2, and output signals of the flexible pressure sensing array are transmitted to the neural network trainer 4 through a serial port for learning; the neural network trainer 4 inputs the trained neural network into the characteristic comparison module 5, and compares and identifies the characteristic comparison module with the pressure signal output by the analog-to-digital conversion circuit 3 in real time, so that the intelligent monitoring of human body form health is realized.
Referring to fig. 2, the neural network trainer 4 is a three-layer network structure composed of an input layer neuron 41, an implicit layer neuron 42 and an output layer neuron 43, the input layer neuron 41 accesses an output signal of the analog-to-digital conversion circuit 22 to the implicit layer neuron 42, the output signal is input to the output layer neuron 43 from the implicit layer neuron 42, the output signal is compared by a training parameter comparison unit 44, and if the requirement of a set training parameter is met, the training is completed; if the set training parameter requirement is not met, the training is finished after the training is adjusted by the correction weight 45, the training returns to the hidden layer neuron 42, the hidden layer neuron 42 enters the output layer neuron 43 again, and the training of the multiple-cycle learning is carried out in this way until the preset training parameter requirement is met. The trained neural network is transmitted to the characteristic comparison module 5 from the neuron 43 of the output layer, a plurality of groups of pressure signals acquired by the flexible pressure sensing array 1 in real time are sequentially processed by the gating circuit 2 and the analog-to-digital conversion circuit 3 and then transmitted to the trained neural network, and the gait phase is identified after the characteristics of the neural network are compared, so that the intelligent monitoring of the human body morphological health is realized.
Referring to fig. 3, when a person walks, the gait can be subdivided into: an initial touchdown period (heel stress, foot sole and knee non-stress) → a support reaction period (heel, knee and forefoot stress) → a mid-point support period (heel, foot sole and knee stress) → a support late period (knee and heel non-stress, foot sole stress) → an early swing period (knee stress, heel and foot sole non-stress) → an early swing period (knee stress, foot and foot sole non-stress) → a mid-swing period (knee stress, foot and foot sole non-stress) → a late swing period (knee stress, foot and foot sole stress) → eight gait periods. The eight gaits cause different pressure distributions on the foot and the knee, the knee straightens with flexion, the plantar point of force varies between the heel and the sole, and the flexion of the knee causes a variation in the pressure value sensed by the flexible pressure sensor 1 at the knee. When the sole is stressed on the heel, the stress of the heel part of the flexible pressure sensing array 1 is obvious, and when the sole is stressed on the sole, the stress of the sole part of the flexible pressure sensing array 1 is obvious. Signals collected by the flexible pressure sensing array 1 attached to the knee and the sole are processed by the gating circuit 2, then are input to the analog-to-digital conversion circuit 3, are converted into digital signals, and are then transmitted to the neural network trainer 4, the neural network is trained based on multiple groups of pressure data, and the successfully trained neural network is used for identifying the walking stage of the human body.
The present invention is further illustrated by the following specific examples.
Example 1
Referring to fig. 4, the flexible pressure sensing array 1 is attached to the knee and sole of the subject 6, and pressure data of eight gait phases, i.e., initial landing phase, support reaction phase, midpoint support phase, support late phase, pre-swing phase, early swing phase, mid-swing phase and post-swing phase, are collected. The flexible pressure sensing array 1 uses 64 rows of sensing arrays and 64 columns of sensing arrays, and a plurality of sensing units work simultaneously to acquire pressure values of feet and knees. The flexible pressure sensing array 1 transmits the acquired pressure signals to the gating module 2, the gating module 2 adopts a CD4067 chip, output signals of the gating module are connected to the analog-to-digital conversion circuit 3, and the analog-to-digital conversion circuit 3 converts analog signals of the data of eight walking stages of the testee 6 into digital signals and then transmits the digital signals to the neural network trainer 4 for learning. 3600 sets of data are collected in each walking stage, and the training parameters in the training parameter unit 44 are set as: training a minimum error: 0.001; maximum number of steps allowed for training: 5000 steps; learning rate: 0.05; the training results are displayed every 10 steps. After the training result meets the parameter requirement set in the training parameter unit 44, the training is finished, and the successfully trained neural network is output to the feature comparison module 5 from the output layer neuron 43. When the testee 6 walks again, the flexible pressure sensing array 1 processes the knee and foot pressure data collected in real time through the gating module 2 and the analog-to-digital conversion circuit 3, and then the knee and foot pressure data are input into the trained neural network through the characteristic comparison module 5 to identify the walking stage of the human body, so that the intelligent monitoring of the human body form health is realized.
The present invention is further described above, but not limited to this patent, and all equivalent implementations of the present invention are intended to be encompassed by the scope of the claims of this patent.
Claims (1)
1. A flexible gait monitoring device is characterized by comprising a flexible pressure sensing array (1), a gating circuit (2), an analog-to-digital conversion circuit (3), a neural network trainer (4) and a characteristic comparison module (5), wherein the flexible pressure sensing array (1) is used for collecting a plurality of groups of pressure signals of knees and soles of a human body during walking and accessing the pressure signals into the analog-to-digital conversion circuit (3) through the gating circuit (2), and output signals of the flexible pressure sensing array are transmitted to the neural network trainer (4) through serial ports for learning; the neural network trainer (4) inputs the trained neural network into the characteristic comparison module (5), and compares and identifies the characteristic comparison module with a voltage signal output by the analog-to-digital conversion circuit (3) in real time, so that intelligent monitoring of human body form health is realized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201821836259.XU CN210077662U (en) | 2018-11-08 | 2018-11-08 | Flexible gait monitoring device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201821836259.XU CN210077662U (en) | 2018-11-08 | 2018-11-08 | Flexible gait monitoring device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN210077662U true CN210077662U (en) | 2020-02-18 |
Family
ID=69469555
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201821836259.XU Active CN210077662U (en) | 2018-11-08 | 2018-11-08 | Flexible gait monitoring device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN210077662U (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109431510A (en) * | 2018-11-08 | 2019-03-08 | 华东师范大学 | A kind of flexible gait monitoring device calculated based on artificial intelligence |
| CN113598753A (en) * | 2021-07-14 | 2021-11-05 | 华中科技大学 | Wearable distributed flexible pressure sensing arm ring |
-
2018
- 2018-11-08 CN CN201821836259.XU patent/CN210077662U/en active Active
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109431510A (en) * | 2018-11-08 | 2019-03-08 | 华东师范大学 | A kind of flexible gait monitoring device calculated based on artificial intelligence |
| CN113598753A (en) * | 2021-07-14 | 2021-11-05 | 华中科技大学 | Wearable distributed flexible pressure sensing arm ring |
| CN113598753B (en) * | 2021-07-14 | 2022-04-08 | 华中科技大学 | Wearable distributed flexible pressure sensing arm ring |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Martinez-Hernandez et al. | Simultaneous Bayesian recognition of locomotion and gait phases with wearable sensors | |
| CN108720841A (en) | Wearable lower extremity movement correction system based on cloud detection | |
| CN109794042A (en) | A kind of body gait based on cloud platform and lower limb harmony rehabilitation training platform | |
| CN106308809A (en) | Method for recognizing gait of thigh amputation subject | |
| CN210077662U (en) | Flexible gait monitoring device | |
| KR101492480B1 (en) | System for Analyzing Surface Electromyographic based on Gait Phases | |
| CN109446972B (en) | Gait recognition model establishing method, recognition method and device based on electromyographic signals | |
| CN109350071A (en) | A kind of flexible gait monitoring method calculated based on artificial intelligence | |
| CN110339024A (en) | Lower limb exoskeleton robot and its real-time gait switching method and storage device | |
| CN105286804A (en) | Wearable knee-crawling movement physiological parameter detection device | |
| CN109567812A (en) | Gait analysis system based on Intelligent insole | |
| CN107357419B (en) | Action recognition system and method based on common shrinkage rate | |
| CN109805935A (en) | A kind of intelligent waistband based on artificial intelligence hierarchical layered motion recognition method | |
| KR101492478B1 (en) | Locomotion Mode Pattern Recognition System using Fusion Sensor | |
| Wang et al. | Research on gait recognition based on lower limb emg signal | |
| Mishra et al. | EMG analysis of lower limb muscles for developing robotic exoskeleton orthotic device | |
| Chia et al. | A multi-channel biomimetic neuroprosthesis to support treadmill gait training in stroke patients | |
| CN109431510A (en) | A kind of flexible gait monitoring device calculated based on artificial intelligence | |
| Giuffrida et al. | Upper-extremity stroke therapy task discrimination using motion sensors and electromyography | |
| CN209574702U (en) | A kind of lower limb exoskeleton formula gait analysis system | |
| Adil et al. | Extreme learning machine based sEMG for drop-foot after stroke detection | |
| Gouwanda et al. | Application of hybrid multi-resolution wavelet decomposition method in detecting human walking gait events | |
| Shen et al. | Recognition of phases in sit-to-stand motion by Neural Network Ensemble (NNE) for power assist robot | |
| CN112331299A (en) | Posture correction method | |
| CN110693501A (en) | Wireless walking gait detection system based on multi-sensor fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231031 Address after: Room 2002, No. 9, Xinghewan Phase III, Lane 3988, Duhui Road, Minhang District, Shanghai, 201108 Patentee after: Shanghai Yingshitu Technology Co.,Ltd. Address before: 200241 No. 500, Dongchuan Road, Shanghai, Minhang District Patentee before: EAST CHINA NORMAL University |