CN116754114A - Analysis method and system for ground contact and ground separation conditions of walking stick - Google Patents
Analysis method and system for ground contact and ground separation conditions of walking stick Download PDFInfo
- Publication number
- CN116754114A CN116754114A CN202310706314.2A CN202310706314A CN116754114A CN 116754114 A CN116754114 A CN 116754114A CN 202310706314 A CN202310706314 A CN 202310706314A CN 116754114 A CN116754114 A CN 116754114A
- Authority
- CN
- China
- Prior art keywords
- pressure
- point
- crutch
- initial
- acceleration
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000926 separation method Methods 0.000 title claims abstract description 25
- 238000004458 analytical method Methods 0.000 title claims abstract description 18
- 241001166076 Diapheromera femorata Species 0.000 title description 19
- 238000005070 sampling Methods 0.000 claims abstract description 79
- 230000005021 gait Effects 0.000 claims abstract description 74
- 238000012545 processing Methods 0.000 claims abstract description 21
- 230000001133 acceleration Effects 0.000 claims description 147
- 238000000034 method Methods 0.000 claims description 42
- 238000012549 training Methods 0.000 claims description 20
- 239000004568 cement Substances 0.000 claims description 11
- 230000008034 disappearance Effects 0.000 claims description 10
- 230000001174 ascending effect Effects 0.000 claims description 9
- 238000010801 machine learning Methods 0.000 claims description 7
- 230000005484 gravity Effects 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000035945 sensitivity Effects 0.000 claims description 6
- 230000000295 complement effect Effects 0.000 claims description 3
- 238000010845 search algorithm Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 230000033001 locomotion Effects 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 208000001132 Osteoporosis Diseases 0.000 description 1
- 208000027089 Parkinsonian disease Diseases 0.000 description 1
- 206010034010 Parkinsonism Diseases 0.000 description 1
- 208000006011 Stroke Diseases 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000002493 climbing effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000001771 impaired effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000037230 mobility Effects 0.000 description 1
- 230000003287 optical effect Effects 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 an analysis method of the ground contact and ground separation condition of a crutch, which comprises the steps of acquiring pressure signals in four directions, which are acquired by a four-way pressure signal sensor arranged in a crutch base, of each sampling period in a sampling total period; processing the pressure signals in four directions acquired in each sampling period; and analyzing the pressure signals in four directions processed in all the sampling periods on the same time axis based on a peak searching algorithm of the threshold value to obtain an initial touchdown direction and an end touchdown direction, and further combining peak time points corresponding to the initial touchdown direction and the end touchdown direction on the whole time axis to determine gait periods contained in the sampling total periods and initial touchdown points and end contact points corresponding to each gait period. The invention also provides an analysis system for the ground contact and ground separation conditions of the crutch. By implementing the invention, the ground contact and ground separation conditions of the crutch can be accurately analyzed.
Description
Technical Field
The invention relates to the technical field of computer intelligent recognition, in particular to a method and a system for analyzing the ground contact and ground separation conditions of a crutch.
Background
With the aggravation of the aging phenomenon of the population of the society, the number of the aged population is continuously increased, and the situation of walking assisted by using the crutch is more and more common. The crutch gait can reflect the health condition of the crutch and symptoms of a plurality of diseases, such as osteoporosis and the like. In addition, the crutch can also be used for preventing the elderly from falling down and monitoring the rehabilitation progress of patients suffering from stroke and parkinsonism. Therefore, it is important to monitor walking gait information of the aged and the mobility impaired.
With the rapid development of the Internet of things and artificial intelligence technology, a plurality of methods related to walking stick gait monitoring and recognition are emerging in recent years. Some researchers have proposed methods of capturing images of the crutch using cameras, sensing crutch motion with millimeter wave radar, and installing a large range of pressure signal sensor arrays on the walking field to monitor and analyze the gait of the crutch. However, these methods need to be implemented in a fixed field, and it is difficult to achieve gait monitoring at any time and any place.
In order to realize the monitoring of walking stick gait at any time and any place, many researchers propose to mount inertial sensors on the body of a walking stick user or directly on a crutch, monitor and analyze the walking stick gait by measuring acceleration and angular velocity, and simultaneously, mount pressure sensors on the handle of the crutch to analyze the pressure information of the hand to the handle to analyze the walking stick gait. The method can analyze walking stick gait to a certain extent, but can not accurately distinguish the occurrence of ground contact and ground separation events of the walking stick, and is easy to misjudge when the walking stick is swung in the air.
Disclosure of Invention
The technical problem to be solved by the embodiment of the invention is to provide a method and a system for analyzing the ground contact and ground separation conditions of a crutch, which can accurately analyze the ground contact and ground separation conditions of the crutch.
In order to solve the technical problems, the embodiment of the invention provides a method for analyzing the ground contact and ground separation conditions of a crutch, which comprises the following steps:
in the sampling total period, acquiring pressure signals in four directions, which are acquired by a four-way pressure signal sensor arranged in a crutch base, of each sampling period;
processing the pressure signals in four directions acquired in each sampling period;
and analyzing the pressure signals in four directions processed in each sampling period on the same time axis based on a peak searching algorithm of a threshold value to obtain an initial touchdown direction and an end touchdown direction, and determining gait periods contained in the sampling total period and initial touchdown points and end contact points corresponding to each gait period by combining pressure occurrence time points corresponding to the initial touchdown direction and pressure disappearance time points corresponding to the end touchdown direction.
The pressure signals in four directions acquired in each sampling period are calculated through a formula (1) to obtain corresponding voltage values;
pressure＝(ad/(2 12 -1))×V ref (1)；
The pressure is a voltage value corresponding to the pressure signal in each direction; AD is the AD sampling result of the pressure signal sensor in each direction, i.e., raw data; v (V) ref Is the external reference voltage of the pressure signal sensor in each direction, which is a fixed value.
The pressure signals in four directions acquired in each sampling period are processed through a formula (2);
y[n]＝b 0 x[n]+b 1 x[n-1]+b 2 x[n-2]+b 3 x[n-3]+b 4 x[n-4]-
a 1 y[n-1]-a 2 y[n-2]-a 3 y[n-3]-a 4 y[n-5](2)；
wherein ,b0 ,b 1 ,b 2 ,b 3 Is the forward coefficient of the filter; a, a 1 ,a 2 ,a 3 ,a 4 Is the feedback coefficient of the filter; x [ n ]]And y [ n ]]A sequence of discrete time domains of the input and output signals, respectively.
The peak value searching algorithm based on the threshold value analyzes the pressure signals in four directions processed in each sampling period on the same time axis to obtain an initial touchdown direction and an end touchdown direction, and determines a gait period contained in the sampling total period and an initial touchdown point and an end contact point corresponding to each gait period by combining a pressure occurrence time point corresponding to the initial touchdown direction and a pressure disappearance time point corresponding to the end touchdown direction, wherein the specific steps include:
respectively carrying out peak value searching on the pressure signals in four directions processed in each sampling period based on a peak value searching algorithm of a threshold value; the peak searching algorithm based on the threshold value specifically comprises the steps of determining a pressure signal in one direction of four directions as a current searched data signal, searching extremum points of the current searched data signal to obtain all peaks and valleys, and if the fact that the peaks and the valleys are adjacent to each other in the current searched data signal and the numerical value difference between the peaks and the adjacent valleys is larger than a preset threshold value is detected, storing the peaks as true peak points, otherwise, storing the peaks as false peak points and discarding the peaks;
On the same time axis, analyzing peak points stored after peak value searching is carried out on pressure signals in four directions respectively to obtain an initial touchdown direction and an end touchdown direction, and determining gait cycles contained in the sampling total period and initial touchdown points and end contact points corresponding to each gait cycle by combining pressure occurrence time points corresponding to the initial touchdown direction and pressure disappearance time points corresponding to the end touchdown direction, wherein the method specifically comprises the following steps: traversing the whole time axis from zero, finding the time when the first peak point appears and the pressure direction corresponding to the peak point, taking the time as the initial ground contact direction of the crutch, continuing traversing backwards by taking the time of the point as the start time until the peak point of pressure in the other three directions appears is found, and recording the time when the last pressure peak point appears and the corresponding pressure direction, and taking the time as the end ground contact direction of the crutch; if the pressure in the initial touchdown direction appears again and the pressures in the other three directions do not all appear, the touchdown direction is directly used as the touchdown ending direction;
traversing forward on the one path of signal in each initial contact direction by taking each initial contact direction and the corresponding peak time point as references until the pressure becomes zero, recording the time of the point when the pressure becomes zero, and recording the time as the initial contact point of the gait cycle; and traversing backward on the one path of signal in each end touch direction by taking each end touch direction and the corresponding peak time as references, until the pressure becomes zero, recording the time when the pressure becomes zero, and recording the time as the end contact point of the gait cycle; and (3) until the whole time axis is traversed, searching of all initial contact points and ending contact points in the sampling total period can be completed.
Wherein the method further comprises:
in the sampling total period, acquiring acceleration signals and angular velocity signals acquired by an inertial sensor arranged in a crutch base in each sampling period, and processing the acceleration signals and the angular velocity signals to obtain pitch angles, roll angles, x-axis accelerations, z-axis accelerations and y-axis angular velocities processed in all the sampling periods;
according to the determined initial touchdown point and the determined ending contact point corresponding to each gait cycle, pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to each initial touchdown point and each ending contact point are obtained, and peak value size of four-way pressure, pitch angle peak value size and ratio of the pitch angle peak value size to the initial touchdown point, ratio of the roll angle peak value size to the initial touchdown point, ratio of the x-axis acceleration peak value size to the x-axis acceleration size to the initial touchdown point, ratio of the z-axis acceleration peak value size to the z-axis acceleration size corresponding to the initial touchdown point and y-axis angular velocity peak value in each gait cycle are further determined; simultaneously importing the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity corresponding to each initial touchdown point and each end touchdown point corresponding to each gait cycle, and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the ratio of the roll angle peak value to the roll angle value corresponding to the initial touchdown point, the ratio of the x-axis acceleration peak value to the x-axis acceleration value corresponding to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration peak value to the initial touchdown point and the y-axis angular velocity peak value in each gait cycle into a trained walking scene recognition model to obtain a current walking scene; wherein the current crutch scene is one of a cement flat land, a grassland, an ascending stair and a descending stair.
The acceleration and the angular velocity acquired in each sampling period can be calculated by a formula (3);
where i=x, y, z.The original output of the triaxial acceleration under the body coordinate system is given in g; />Is an acceleration signal generated by an inertial sensor; sens (Sens) acc The sensitivity corresponding to the acceleration range; />Is the output of the triaxial angular velocity under the body coordinate system, and the unit is rad/s; />Is an angular velocity signal generated by an inertial sensor; sens (Sens) ω Is the sensitivity corresponding to the angular velocity range; sens (Sens) acc With Sens ω Can be obtained by querying an inertial sensor manual.
The pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity processed in each sampling period are obtained by executing the following steps:
(7.1) estimating zero points and proportionality coefficients of the triaxial acceleration of the inertial sensor by a recursive least square ellipsoid hypothesis calibration algorithm:
initializing c 0 ＝[0 1 0 1 0 -g 2] and P0 ＝1000×I 6 Wherein g is the size of the earth gravity field, I 6 Is a sixth-order identity matrix;
rotating the inertial sensor along three axes of the sensor, collecting acceleration data and building up at each cycle and /> wherein />The original output is the tri-axial acceleration;
And (3) respectively calculating:P n ＝(I 6 -Ky n )P n-1 and cn ＝c n-1 +K(z n -y n c n-1 ) Repeating the steps until c n Converging;
calculating the sensor zero point b x ＝c n (1),Scaling factor and />
Using the estimated zero and scaling factor, the acceleration output of the calibrated sensor can be calculatedWherein i=x, y, z; />Is the calibration output of the triaxial accelerometer; />Is the original output of the triaxial accelerometer; b i Is the calculated sensor zero point; s is(s) fi Is the calculated scaling factor;
and (7.2) fusing acceleration and angular velocity signals through a complementary filter, and calculating attitude information, wherein the method specifically comprises the following steps:
calculating the magnitude of the earth gravity field in the volume coordinate system through a formula (4):
wherein ,is the estimated gravitational field in the volumetric coordinate system; k (k) g Is a parameter that adjusts the filter cut-off frequency; />Is the calibration output of the triaxial accelerometer; /> Is the output of a triaxial gyroscope; t (T) a Is the period of the filter;
based on the estimated gravitational field magnitude, the magnitude of the attitude angle is calculated using equations (5) and (6):
wherein θ is the pitch angle; phi is the roll angle;
the three-axis acceleration in the body coordinate system is converted to the three-axis acceleration in the horizontal coordinate system by the attitude angle information using formulas (7) and (8):
wherein ,the acceleration is represented by a horizontal coordinate system; / >A rotation matrix from a body coordinate system to a horizontal coordinate system; the method comprises the steps of carrying out a first treatment on the surface of the />
The crutch scene recognition model is obtained by training a lightGBM machine learning algorithm; in the super-parameters of the lightGBM algorithm, the learning rate is 0.001, the training times are 50000 times, and the number of the training times is 1000 times in advance.
The embodiment of the invention also provides an analysis system for the ground contact and ground separation conditions of the crutch, which comprises the following steps:
the pressure signal acquisition unit is used for acquiring pressure signals in four directions acquired by a four-way pressure signal sensor arranged in the crutch base in each sampling period;
the pressure signal processing unit is used for processing the pressure signals in four directions acquired in each sampling period;
the touchdown and touchdown analysis unit is used for analyzing the pressure signals in four directions processed in each sampling period on the same time axis based on a peak value searching algorithm of a threshold value so as to obtain an initial touchdown direction and an end touchdown direction, and determining gait periods contained in the sampling total period and initial touchdown points and end contact points corresponding to each gait period by combining pressure occurrence time points corresponding to the initial touchdown direction and pressure disappearance time points corresponding to the end touchdown direction.
Wherein, still include:
the acceleration signal and angular velocity signal acquisition and processing unit is used for acquiring and processing acceleration signals and angular velocity signals acquired by an inertial sensor arranged in the crutch base in each sampling period so as to obtain pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity processed in each sampling period;
the ground contact point and off-ground acceleration signal and angular velocity signal determining unit is used for obtaining pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to each initial ground contact point and each end contact point according to the determined initial ground contact point and end contact point corresponding to each gait cycle;
the crutch scene recognition unit is used for recognizing the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity corresponding to each initial touchdown point and each ending touchdown point respectively; and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the roll angle peak value and the ratio of the roll angle value to the initial touchdown point, the x-axis acceleration peak value and the ratio of the x-axis acceleration value to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration value to the initial touchdown point, and the y-axis angular velocity peak value in each gait cycle are also provided. Simultaneously importing the data into a trained crutch scene recognition model to obtain a current crutch scene; wherein the current crutch scene is one of a cement flat land, a grassland, an ascending stair and a descending stair.
The embodiment of the invention has the following beneficial effects:
1. the invention analyzes the pressure signals in four directions of each sampling period of the pressure sensor based on a peak value searching algorithm of the threshold value to rapidly determine the gait period contained in the sampling total period and the initial touchdown point and the ending contact point corresponding to each gait period, thereby accurately analyzing the situation of the walking stick touchdown and the ground separation;
2. based on the pressure signal of the pressure sensor, the gait data of the walking stick is acquired by combining the acceleration signal and the angular velocity signal of the inertial sensor, so that the defect of identifying the walking stick state caused by a single sensor can be avoided, and the accuracy of gait analysis is further improved;
3. according to the invention, the machine learning lightGBM algorithm is utilized to treat the leaning crutch scene classification recognition of the intelligent crutch as a multi-value classification problem, so that not only can the leaning crutch scene be rapidly recognized, but also the recognition precision can be effectively improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are required in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that it is within the scope of the invention to one skilled in the art to obtain other drawings from these drawings without inventive faculty.
FIG. 1 is a flow chart of a method for analyzing the ground contact and lift conditions of a crutch according to an embodiment of the present invention;
FIG. 2 is a schematic diagram showing the distribution of acceleration meters before and after calibration in an analysis method of the ground contact and lift-off conditions of a crutch according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating the installation of an intelligent crutch pressure sensor and an inertial sensor in an application scenario of a method for analyzing the situation of touching and leaving the ground of a crutch according to an embodiment of the present invention;
FIG. 4 is a pressure diagram generated by four pressure sensors in the crutch in the process of walking in the application scenario of the analysis method of the ground contact and ground separation situation of the crutch provided by the embodiment of the invention;
FIG. 5 is a schematic diagram of the cycle of crutch and crutch movement and the distribution of IC and EC points thereof in an application scenario of the method for analyzing the situation of touching and leaving the ground of a crutch provided by the embodiment of the invention;
FIG. 6 is a graph of the result of finding the pressure peak point in the application scenario of the analysis method for the ground contact and lift conditions of the crutch according to the embodiment of the present invention;
fig. 7 is a schematic diagram of distribution of pitch angles corresponding to IC points and EC points in an application scenario of a method for analyzing ground contact and ground separation conditions of a crutch according to an embodiment of the present invention;
fig. 8 is a schematic diagram of the distribution of the Y-axis angular velocity corresponding to the IC point and the EC point in the application scenario of the analysis method of the ground contact and ground separation situation of the crutch according to the embodiment of the present invention;
Fig. 9 is a schematic structural diagram of an analysis system for ground contact and ground separation of a crutch according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings, for the purpose of making the objects, technical solutions and advantages of the present invention more apparent.
As shown in fig. 1, in an embodiment of the present invention, a method for analyzing the ground contact and ground separation condition of a crutch is provided, including the following steps:
step S1, acquiring pressure signals in four directions, which are acquired by a four-way pressure sensor arranged in a crutch base, in each sampling period;
the specific process is that the pressure sensor is set to sample at the frequency of 10Hz, so that pressure signals in four directions, which are acquired by the four-way pressure signal sensor, are obtained in each sampling period. The four-way pressure signal sensor is a flexible film pressure sensor and is provided with four pressure sensing units, and the four pressure sensing units are arranged on the front, back, left and right directions of the base of the crutch.
S2, processing pressure signals in four directions acquired in each sampling period;
firstly, calculating the pressure signals in four directions acquired in each sampling period through a formula (1) to obtain corresponding voltage values;
pressure＝(ad/(2 12 -1))×V ref (1)；
The pressure is a voltage value corresponding to the pressure signal in each direction; ad is the pressure sensor in each directionAD sampling result of the signal, namely the original data; v (V) ref Is the external reference voltage of the pressure signal sensor in each direction, which is a fixed value.
Secondly, processing the pressure signals in four directions acquired in each sampling period through a formula (2);
y[n]＝b 0 x[n]+b 1 x[n-1]+b 2 x[n-2]+b 3 x[n-3]+b 4 x[n-4]-
a 1 y[n-1]-a 2 y[n-2]-a 3 y[n-3]-a 4 y[n-5](2)；
wherein ,b0 ,b 1 ,b 2 ,b 3 Is the forward coefficient of the filter; a, a 1 ,a 2 ,a 3 ,a 4 Is the feedback coefficient of the filter; x [ n ]]And y [ n ]]A sequence of discrete time domains of the input and output signals, respectively. It should be noted that b 0 ,b 1 ,b 2 ,b 3 and a1 ,a 1 ,a 1 ,a 1 May be calculated based on the desired cut-off frequency and passband/stopband attenuation.
And S3, analyzing the pressure signals in four directions processed in each sampling period on the same time axis based on a peak searching algorithm of a threshold value to obtain an initial touchdown direction and an end touchdown direction, and determining gait periods contained in the sampling total period and initial touchdown points and end contact points corresponding to each gait period by combining the pressure occurrence time corresponding to the initial touchdown direction and the pressure disappearance time corresponding to the end touchdown direction.
The specific process is that the walking stick mainly has two stages of movement: a ground touching stage and a dead space stage. There are two important crutch events: initial Contact (IC) and End Contact (EC). Therefore, the walking gait features of the crutch are required to be extracted according to the movement cycle law when the crutch is leaned, and the walking gait features are specifically as follows:
Firstly, respectively carrying out peak searching on pressure signals in four directions processed in each sampling period based on a peak searching algorithm of a threshold value; the peak searching algorithm based on the threshold value specifically comprises the steps of determining a pressure signal in one direction of four directions as a current searched data signal, searching extremum points of the current searched data signal to obtain all peaks and valleys, and if the fact that the peaks and the valleys are adjacent to each other in the current searched data signal and the numerical value difference between the peaks and the adjacent valleys is larger than a preset threshold value is detected, storing the peaks as true peak points, otherwise, storing the peaks as false peak points and discarding the peaks.
Secondly, on the same time axis, respectively analyzing peak points stored after peak value searching of pressure signals in four directions to obtain an initial touchdown direction and an end touchdown direction, and determining gait cycles contained in a sampling total period and initial touchdown points and end contact points corresponding to each gait cycle by combining pressure occurrence time corresponding to the initial touchdown direction and pressure disappearance time corresponding to the end touchdown direction, wherein the method specifically comprises the following steps: traversing the whole time axis from zero, finding the time when the first peak point appears and the pressure direction corresponding to the peak point, taking the time as the initial ground contact direction of the crutch, continuing traversing backwards by taking the time of the point as the start time until the peak point of pressure in the other three directions appears is found, and recording the time when the last pressure peak point appears and the corresponding pressure direction, and taking the time as the end ground contact direction of the crutch; if the pressure in the initial touchdown direction appears again and the pressures in the other three directions do not all appear, the touchdown direction is directly used as the touchdown ending direction;
Finally, taking each initial contact direction and the corresponding peak time point as a reference, traversing forward on the path of signal in each initial contact direction until the pressure becomes zero, recording the time of the point when the pressure point becomes zero, and recording the time as the initial contact point of the gait cycle; and traversing backward on the one path of signal in each end touch direction by taking each end touch direction and the corresponding peak time as references, until the pressure becomes zero, recording the time when the pressure becomes zero, and recording the time as the end contact point of the gait cycle; the searching of the initial contact point and the ending contact point in the sampling period can be completed.
In the embodiment of the invention, because the crutch scene comprises but is not limited to a cement land, a grassland, an ascending stair, a descending stair and the like, the crutch scene of the user can be accurately identified by combining the crutch gesture signals (such as acceleration signals and angular velocity signals) on the basis of the pressure signals, the crutch state identification defect caused by a single sensor can be avoided, the gait analysis accuracy is further improved, the crutch scene can be rapidly identified, and the identification accuracy is effectively improved.
At this time, the method further includes:
step S41, acquiring acceleration signals and angular velocity signals acquired by an inertial sensor arranged in a crutch base in each sampling period, and processing the acceleration signals and the angular velocity signals to obtain pitch angles, roll angles, x-axis accelerations, z-axis accelerations and y-axis angular velocities processed in all the sampling periods;
the method comprises the specific process that the inertial sensor and the pressure sensor are sampled at the frequency of 10Hz, so that an acceleration signal and an angular velocity signal acquired by the inertial sensor in each sampling period are obtained. The inertial sensor is a six-axis inertial measurement unit and comprises a triaxial accelerometer and a triaxial angular velocity meter, and the installation position of the inertial sensor is arranged in a small circuit board at the bottom of the crutch.
Step S42, according to the determined initial touchdown point and the determined ending contact point corresponding to each gait cycle, obtaining a pitch angle, a roll angle, an x-axis acceleration, a z-axis acceleration and a y-axis angular velocity corresponding to each initial touchdown point and each ending contact point respectively;
firstly, the acceleration and the angular velocity acquired for each sampling period can be calculated by a formula (3);
where i=x, y, z. Is a body coordinate systemThe original output of the lower triaxial acceleration is in g; />Is an acceleration signal generated by an inertial sensor; sens (Sens) acc The sensitivity corresponding to the acceleration range; />Is the output of the triaxial angular velocity under the body coordinate system, and the unit is rad/s; />Is an angular velocity signal generated by an inertial sensor; sens (Sens) ω Is the sensitivity corresponding to the angular velocity range; sens (Sens) acc With Sens ω Can be obtained by querying an inertial sensor manual.
Secondly, the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity processed in each sampling period are obtained by executing the following steps:
(7.1) estimating zero points and proportionality coefficients of the triaxial acceleration of the inertial sensor by a recursive least square ellipsoid hypothesis calibration algorithm:
initializing c 0 ＝[0 1 0 1 0-g 2] and P0 ＝1000×I 6 Wherein g is the size of the earth gravity field, I 6 Is a sixth-order identity matrix;
rotating the inertial sensor along three axes of the sensor, collecting acceleration data and building up at each cycle and /> wherein />The original output is the tri-axial acceleration;
and (3) respectively calculating: and cn ＝c n-1 +K(z n -y n c n-1 ) Repeating the steps until c n Converging;
calculating the sensor zero point b x ＝c n (1),Scaling factor and />
Using the estimated zero and scaling factor, the acceleration output of the calibrated sensor can be calculated Where i=x, y, z. The method comprises the steps of carrying out a first treatment on the surface of the />Is the calibration output of the triaxial accelerometer; />Is the original output of the triaxial accelerometer; b i Is the calculated sensor zero point; s is(s) fi =calculated scaling factor; in one example, as shown in FIG. 2, a schematic diagram of the acceleration magnitude before and after accelerometer calibration is provided.
And (7.2) fusing acceleration and angular velocity signals through a complementary filter, and calculating attitude information, wherein the method specifically comprises the following steps:
calculating the magnitude of the earth gravity field in the volume coordinate system through a formula (4):
wherein ,is the estimated gravitational field in the volumetric coordinate system; k (k) g Is a parameter that adjusts the filter cut-off frequency; />Is the calibration output of the triaxial accelerometer; /> Is the output of a triaxial gyroscope; t (T) a Is the period of the filter;
based on the estimated gravitational field magnitude, the magnitude of the attitude angle is calculated using equations (5) and (6):
wherein θ is the pitch angle; phi is the roll angle;
the three-axis acceleration in the body coordinate system is converted to the three-axis acceleration in the northeast ground coordinate system NED by using formulas (7) and (8) through the attitude angle information:
wherein ,the acceleration is represented by a horizontal coordinate system; />A rotation matrix from a body coordinate system to a horizontal coordinate system; ψ=0; />
Step S43, extracting pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to the initial touchdown point and the final touchdown point in each gait cycle; and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the roll angle peak value and the ratio of the roll angle value to the initial touchdown point, the x-axis acceleration peak value and the ratio of the x-axis acceleration value to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration value to the initial touchdown point, and the y-axis angular velocity peak value in each gait cycle are also provided. Simultaneously importing the data into a trained crutch scene recognition model to obtain a current crutch scene; wherein the current crutch scene is one of a cement flat land, a grassland, an ascending stair and a descending stair.
Firstly, a crutch scene recognition model is predefined and training is carried out. The leaning crutch scene recognition model is obtained by training a lightGBM machine learning algorithm; at this time, in the super-parameters of the lightGBM algorithm, the learning rate is 0.001, the training time is 50000 times, and the number of rounds is stopped 1000 times in advance.
The input of the walking stick scene recognition model is pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to the initial touchdown point and the final touchdown point of each gait cycle; and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the roll angle peak value and the ratio of the roll angle value to the initial touchdown point, the x-axis acceleration peak value and the ratio of the x-axis acceleration value to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration value to the initial touchdown point, and the y-axis angular velocity peak value in each gait cycle are also provided.
The output is one of the scene types of the current walking stick, namely the cement level land, the grassland, the ascending stair and the descending stair. It should be noted that, the data set division, the training sample and the test sample used in the training process of the crutch scene recognition model, and the training mode of the crutch scene recognition model all belong to common technical means in the field, and are not described in detail herein.
Secondly, the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity corresponding to the initial touchdown point and the final touchdown point in each gait cycle are respectively carried out; and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the roll angle peak value and the ratio of the roll angle value to the initial touchdown point, the x-axis acceleration peak value and the ratio of the x-axis acceleration value to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration value to the initial touchdown point, and the y-axis angular velocity peak value in each gait cycle are also provided. And meanwhile, the current crutch scene can be obtained by importing the trained crutch scene recognition model.
As shown in fig. 3 to 8, an application scenario of the method for analyzing the ground contact and ground separation situation of the crutch provided in the embodiment of the present invention is further described, which specifically includes:
the pressure sensor and the inertial sensor are installed in the intelligent crutch as shown in fig. 3.
Inviting 16 volunteers to participate in the crutch scene experiment to acquire training data, wherein the crutch scene comprises: cement level land, grassland, ascending stairs and descending stairs; to ensure the authenticity of the data, volunteers are required to walk by walking on the crutch according to the state that the volunteers self consider to be most comfortable and natural, and the collected data are summarized and divided into four groups: data set 1 contains data for all 16 volunteers walking normally on a cement level land, each walking 20 passes around a cement land of about 70m length; the data set 2 contains data of all 16 volunteers performing stair climbing actions on stairs, and each person walks on stairs with the height of 4 stairs; the data set 3 contains data of all 16 volunteers who go downstairs on stairs, and each person goes 20 times to stairs with the height of 4 stairs; the data set 4 contains data for all 16 volunteers walking normally on grasslands, each walking 20 passes of grasslands approximately 70m in length.
When each person leans on the crutch, firstly, the pressure signals (shown in figure 4) in four directions obtained when the person leans on the crutch are periodically obtained through the pressure sensor, and the acceleration signals and the angular velocity signals are periodically obtained through the inertial sensor; then, processing the pressure signals, the acceleration signals and the angular velocity signals of all sampling periods; then, a peak search algorithm based on the threshold value is performed on all the pressure signals on the same time axis, wherein the forward direction and the corresponding peak point search result thereof are obtained (as shown in fig. 5). Further, based on the leaning principle (shown in fig. 6), the initial touchdown point IC and the end touchdown point EC are determined according to the initial touchdown direction and the end touchdown direction. Meanwhile, pitch angles corresponding to the initial touchdown point IC and the end touchdown point EC are calculated according to the processed acceleration signal and the processed angular velocity signal (as shown in fig. 7). Finally, the roll angle, the X-axis angular velocity, the Z-axis angular velocity, and the Y-axis angular velocity of the IC point and the EC point are calculated (as shown in fig. 8).
Considering that each gait cycle corresponds to one crutch, 9738 crutch is available in the data set 1; 8113 overall turns in dataset 2; together, data set 3 has 8234 turns; there were 7802 turns in dataset 4, totaling 33887 turns.
To verify the validity of the model, the 4 data sets need to be first labeled. Subsequently, the 4 data sets were pooled together and randomly split into a training set containing 27109 data and a test set containing 6778 data for classification training and testing at a ratio of 8:2. The training set is used for training the model, and the testing set is used for testing the effectiveness and accuracy of the model.
Training by using a lightGBM machine learning algorithm based on all gait characteristics of the training set; wherein in the super parameters of the lightGBM algorithm: the learning rate is 0.001, the training times are 50000 times, the number of the training turns is 1000 times in advance, and the rest parameters are defaults. And inputting all gait characteristics of the test set into the trained model to enable the model to output a result, and comparing the result with the label marked in the previous step to obtain the accuracy rate of classification and identification of the model, as shown in the table 1.
As can be seen from Table 1, the intelligent crutch formed by the pressure sensor at the bottom of the crutch and the inertial sensor is adopted for classifying and identifying the crutch situation, so that the problem of poor crutch event identification capability caused by independently using the inertial sensor or the pressure sensor at the handle can be effectively avoided, and the accuracy rate of classifying and identifying the crutch situation is greatly improved.
TABLE 1
The embodiment of the invention uses python 3.8 to realize the walking gait scene classification method based on the lightGBM machine learning algorithm, and all experiments are carried out on an Inteli5-7400CPU machine.
As shown in fig. 9, in an embodiment of the present invention, a system for analyzing ground contact and ground separation of a crutch is provided, including:
the pressure signal acquisition unit 110 is used for acquiring pressure signals in four directions acquired by a four-way pressure signal sensor arranged in the crutch base in each sampling period;
a pressure signal processing unit 120, configured to process the pressure signals in four directions acquired in each sampling period;
the touchdown and touchdown analysis unit 130 is configured to analyze the pressure signals in four directions processed in each sampling period on the same time axis based on a peak searching algorithm of the threshold value, so as to obtain an initial touchdown direction and an end touchdown direction, and determine a gait period contained in the sampling total period and an initial touchdown point and an end contact point corresponding to each gait period in combination with a pressure occurrence time corresponding to the initial touchdown direction and a pressure disappearance time corresponding to the end touchdown direction.
Wherein, still include:
the acceleration signal and angular velocity signal acquisition and processing unit is used for acquiring and processing acceleration signals and angular velocity signals acquired by an inertial sensor arranged in the crutch base in each sampling period in the sampling total period so as to obtain pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity after processing in all the sampling periods;
the ground contact point and off-ground acceleration signal and angular velocity signal determining unit is used for obtaining pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to each initial ground contact point and each end contact point according to the determined initial ground contact point and end contact point corresponding to each gait cycle;
the crutch scene recognition unit is used for recognizing the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity corresponding to each initial touchdown point and each ending touchdown point respectively; and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the roll angle peak value and the ratio of the roll angle value to the initial touchdown point, the x-axis acceleration peak value and the ratio of the x-axis acceleration value to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration value to the initial touchdown point, and the y-axis angular velocity peak value in each gait cycle are also provided. Simultaneously importing the data into a trained crutch scene recognition model to obtain a current crutch scene; wherein the current crutch scene is one of a cement flat land, a grassland, an ascending stair and a descending stair.
The embodiment of the invention has the following beneficial effects:
1. the invention analyzes the pressure signals in four directions of each sampling period of the pressure sensor based on a peak value searching algorithm of the threshold value to rapidly determine the gait period contained in the sampling total period and the initial touchdown point and the ending contact point corresponding to each gait period, thereby accurately analyzing the situation of the walking stick touchdown and the ground separation;
2. based on the pressure signal of the pressure sensor, the gait data of the walking stick is acquired by combining the acceleration signal and the angular velocity signal of the inertial sensor, so that the defect of identifying the walking stick state caused by a single sensor can be avoided, and the accuracy of gait analysis is further improved;
3. according to the invention, the machine learning lightGBM algorithm is utilized to treat the leaning crutch scene classification recognition of the intelligent crutch as a multi-value classification problem, so that not only can the leaning crutch scene be rapidly recognized, but also the recognition precision can be effectively improved.
It should be noted that, in the above system embodiment, each unit included is only divided according to the functional logic, but not limited to the above division, so long as the corresponding function can be implemented; in addition, the specific names of the functional units are also only for distinguishing from each other, and are not used to limit the protection scope of the present invention.
Those of ordinary skill in the art will appreciate that all or a portion of the steps in implementing the methods of the above embodiments may be implemented by a program to instruct related hardware, where the program may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc.
The foregoing disclosure is illustrative of the present invention and is not to be construed as limiting the scope of the invention, which is defined by the appended claims.
Claims (10)
1. The method for analyzing the ground contact and ground separation conditions of the crutch is characterized by comprising the following steps:
in the sampling total period, acquiring pressure signals in four directions, which are acquired by a four-way pressure signal sensor arranged in a crutch base, of each sampling period;
processing the pressure signals in four directions acquired in each sampling period;
and analyzing the pressure signals in four directions processed in each sampling period on the same time axis based on a peak searching algorithm of a threshold value to obtain an initial touchdown direction and an end touchdown direction, and determining gait periods contained in the sampling total period and initial touchdown points and end contact points corresponding to each gait period by combining pressure occurrence time points corresponding to the initial touchdown direction and pressure disappearance time points corresponding to the end touchdown direction.
2. The method for analyzing the ground contact and ground separation conditions of the crutch according to claim 1, wherein the pressure signals in four directions obtained in each sampling period are calculated by a formula (1) to obtain corresponding voltage values;
pressure＝(ad/(2 12 -1))×V ref (1)；
the pressure is a voltage value corresponding to the pressure signal in each direction; AD is the AD sampling result of the pressure signal sensor signal in each direction, i.e. raw data; v (V) ref Is the external reference voltage of the pressure signal sensor in each direction, which is a fixed value.
3. The method for analyzing the ground contact and separation conditions of the crutch according to claim 2, wherein the pressure signals in four directions acquired in each sampling period are processed by the formula (2);
y[n]＝b 0 x[n]+b 1 x[n-1]+b 2 x[n-2]+b 3 x[n-3]+b 4 x[n-4]-
a 1 y[n-1]-a 2 y[n-2]-a 3 y[n-3]-a 4 y[n-5](2)；
wherein ,b0 ,b 1 ,b 2 ,b 3 Is the forward coefficient of the filter; a, a 1 ,a 2 ,a 3 ,a 4 Is the feedback coefficient of the filter; x [ n ]]And y [ n ]]A sequence of discrete time domains of the input and output signals, respectively.
4. The method for analyzing the ground contact and lift-off condition of a crutch according to claim 3, wherein the step of analyzing the pressure signals in four directions processed in each sampling period on the same time axis by the peak search algorithm based on the threshold value to obtain an initial ground contact direction and an end ground contact direction, and determining the gait cycle contained in the sampling period and the initial ground contact point and the end contact point corresponding to each gait cycle by combining the pressure occurrence time corresponding to the initial ground contact direction and the pressure disappearance time corresponding to the end ground contact direction comprises the following steps:
Respectively carrying out peak value searching on the pressure signals in four directions processed in each sampling period based on a peak value searching algorithm of a threshold value; the peak searching algorithm based on the threshold value specifically comprises the steps of determining a pressure signal in one direction of four directions as a current searched data signal, searching extremum points of the current searched data signal to obtain all peaks and valleys, and if the fact that the peaks and the valleys are adjacent to each other in the current searched data signal and the numerical value difference between the peaks and the adjacent valleys is larger than a preset threshold value is detected, storing the peaks as true peak points, otherwise, storing the peaks as false peak points and discarding the peaks;
on the same time axis, analyzing peak points stored after peak value searching is carried out on pressure signals in four directions respectively to obtain an initial touchdown direction and an end touchdown direction, and determining gait cycles contained in the sampling total period and initial touchdown points and end contact points corresponding to each gait cycle by combining peak time points corresponding to the initial touchdown direction and the end touchdown direction on the whole time axis, wherein the steps comprise: traversing the whole time axis from zero, finding the time when the first peak point appears and the pressure direction corresponding to the peak point, taking the time as the initial ground contact direction of the crutch, continuing traversing backwards by taking the time of the point as the start time until the peak point of pressure in the other three directions appears is found, and recording the time when the last pressure peak point appears and the corresponding pressure direction, and taking the time as the end ground contact direction of the crutch; if the pressure in the initial touchdown direction appears again and the pressures in the other three directions do not all appear, the touchdown direction is directly used as the touchdown ending direction;
Traversing forward on the one path of signal in each initial contact direction by taking each initial contact direction and the corresponding peak time point as references until the pressure becomes zero, recording the time of the point when the pressure becomes zero, and recording the time as the initial contact point of the gait cycle; and traversing backward on the one path of signal in each end touch direction by taking each end touch direction and the corresponding peak time as references, until the pressure becomes zero, recording the time when the pressure becomes zero, and recording the time as the end contact point of the gait cycle; and (3) until the whole time axis is traversed, searching of all initial contact points and ending contact points in the sampling total period can be completed.
5. The method of analyzing ground contact and lift of a crutch of claim 4, the method further comprising:
in the sampling total period, acquiring acceleration signals and angular velocity signals acquired by an inertial sensor arranged in a crutch base in each sampling period, and processing the acceleration signals and the angular velocity signals to obtain pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity processed in each sampling period;
According to the determined initial touchdown point and the determined ending contact point corresponding to each gait cycle, pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to each initial touchdown point and each ending contact point are obtained, and peak value size of four-way pressure, pitch angle peak value size and ratio of the pitch angle peak value size to the initial touchdown point, ratio of the roll angle peak value size to the initial touchdown point, ratio of the x-axis acceleration peak value size to the x-axis acceleration size to the initial touchdown point, ratio of the z-axis acceleration peak value size to the z-axis acceleration size corresponding to the initial touchdown point and y-axis angular velocity peak value in each gait cycle are further determined; simultaneously importing the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity corresponding to each initial touchdown point and each end touchdown point corresponding to each gait cycle, and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the ratio of the roll angle peak value to the roll angle value corresponding to the initial touchdown point, the ratio of the x-axis acceleration peak value to the x-axis acceleration value corresponding to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration peak value to the initial touchdown point and the y-axis angular velocity peak value in each gait cycle into a trained walking scene recognition model to obtain a current walking scene; wherein the current crutch scene is one of a cement flat land, a grassland, an ascending stair and a descending stair.
6. The method for analyzing the ground contact and separation of a crutch according to claim 5, wherein the acceleration and angular velocity obtained in each sampling period can be calculated by the formula (3);
where i=x, y, z.The original output of the triaxial acceleration under the body coordinate system is given in g; />Is an acceleration signal generated by an inertial sensor; sens (Sens) acc The sensitivity corresponding to the acceleration range; />Is the output of the triaxial angular velocity under the body coordinate system, and the unit is rad/s; />Is an angular velocity signal generated by an inertial sensor; sens (Sens) ω Is the sensitivity corresponding to the angular velocity range; sens (Sens) acc With Sens ω Can be obtained by querying an inertial sensor manual.
7. The method for analyzing the ground contact and lift conditions of the crutch according to claim 6, wherein the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity processed in each sampling period are obtained by executing the following steps:
(7.1) estimating zero points and proportionality coefficients of the triaxial acceleration of the inertial sensor by a recursive least square ellipsoid hypothesis calibration algorithm:
initializing c 0 ＝[0 1 0 1 0 -g 2] and P0 ＝1000×I 6 Wherein g is the size of the earth gravity field, I 6 Is a sixth-order identity matrix;
Rotating the inertial sensor along three axes of the sensor, collecting acceleration data and building up at each cycle and /> wherein />The original output is the tri-axial acceleration;
and (3) respectively calculating:P n ＝(I 6 -Ky n )P n-1 and cn ＝c n-1 +K(z n -y n c n-1 ) Repeating the steps until c n Converging;
calculating the sensor zero point b x ＝c n (1),Scaling factor and />
Using the estimated zero and scaling factor, the acceleration output of the calibrated sensor can be calculatedWhere i=x, y, z. The method comprises the steps of carrying out a first treatment on the surface of the />Is the calibration output of the triaxial accelerometer; />Is the original output of the triaxial accelerometer; b i Is the calculated sensor zero point; s is(s) fi Is the calculated scaling factor;
and (7.2) fusing acceleration and angular velocity signals through a complementary filter, and calculating attitude information, wherein the method specifically comprises the following steps:
calculating the magnitude of the earth gravity field in the volume coordinate system through a formula (4):
wherein ,is the estimated gravitational field in the volumetric coordinate system; k (k) g Is a parameter that adjusts the filter cut-off frequency; />Is the calibration output of the triaxial accelerometer; /> Is the output of a triaxial gyroscope; t (T) a Is the period of the filter;
based on the estimated gravitational field magnitude, the magnitude of the attitude angle is calculated using equations (5) and (6):
wherein θ is the pitch angle; phi is the roll angle;
The three-axis acceleration in the body coordinate system is converted to the three-axis acceleration in the horizontal coordinate system by the attitude angle information using formulas (7) and (8):
wherein ,the acceleration is represented by a horizontal coordinate system; />A rotation matrix from a body coordinate system to a horizontal coordinate system; the method comprises the steps of carrying out a first treatment on the surface of the />
8. The method for analyzing the ground contact and lift-off condition of a crutch according to claim 6, wherein the crutch scene recognition model is trained by using a lightGBM machine learning algorithm; in the super-parameters of the lightGBM algorithm, the learning rate is 0.001, the training times are 50000 times, and the number of the training times is 1000 times in advance.
9. An analysis system for ground contact and ground separation conditions of a crutch, which is characterized by comprising:
the pressure signal acquisition unit is used for acquiring pressure signals in four directions acquired by a four-way pressure signal sensor arranged in the crutch base in each sampling period;
the pressure signal processing unit is used for processing the pressure signals in four directions acquired in each sampling period;
the touchdown and touchdown analysis unit is used for analyzing the processed pressure signals in four directions on the same time axis based on a peak value searching algorithm of the threshold value to obtain an initial touchdown direction and an end touchdown direction, and determining an initial touchdown point and an end contact point corresponding to each gait cycle by combining a pressure occurrence time point corresponding to the initial touchdown direction and a pressure disappearance time point corresponding to the end touchdown direction.
10. The crutch ground contact and lift analysis system of claim 9, further comprising:
the acceleration signal and angular velocity signal acquisition and processing unit is used for acquiring and processing acceleration signals and angular velocity signals acquired by an inertial sensor arranged in the crutch base in each sampling period so as to obtain pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity processed in each sampling period;
the ground contact point and off-ground acceleration signal and angular velocity signal determining unit is used for obtaining pitch angle, roll angle, x-axis acceleration, z-axis acceleration and y-axis angular velocity corresponding to each initial ground contact point and each end contact point according to the determined initial ground contact point and end contact point corresponding to each gait cycle;
the crutch scene recognition unit is used for recognizing the pitch angle, the roll angle, the x-axis acceleration, the z-axis acceleration and the y-axis angular velocity corresponding to each initial touchdown point and each ending touchdown point respectively; and the peak value of four-way pressure, the pitch angle peak value and the ratio of the pitch angle peak value to the pitch angle value corresponding to the initial touchdown point, the roll angle peak value and the ratio of the roll angle value to the initial touchdown point, the x-axis acceleration peak value and the ratio of the x-axis acceleration value to the initial touchdown point, the z-axis acceleration peak value and the ratio of the z-axis acceleration value to the initial touchdown point, and the y-axis angular velocity peak value in each gait cycle are also provided. Simultaneously importing the data into a trained crutch scene recognition model to obtain a current crutch scene; wherein the current crutch scene is one of a cement flat land, a grassland, an ascending stair and a descending stair.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310706314.2A CN116754114A (en) | 2023-06-14 | 2023-06-14 | Analysis method and system for ground contact and ground separation conditions of walking stick |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310706314.2A CN116754114A (en) | 2023-06-14 | 2023-06-14 | Analysis method and system for ground contact and ground separation conditions of walking stick |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116754114A true CN116754114A (en) | 2023-09-15 |
Family
ID=87954694
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310706314.2A Pending CN116754114A (en) | 2023-06-14 | 2023-06-14 | Analysis method and system for ground contact and ground separation conditions of walking stick |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116754114A (en) |
-
2023
- 2023-06-14 CN CN202310706314.2A patent/CN116754114A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11039760B2 (en) | Detection of walking in measurements of the movement of a user | |
| CN108549856B (en) | Human body action and road condition identification method | |
| CN110553643B (en) | Pedestrian self-adaptive zero-speed updating point selection method based on neural network | |
| CN104864873B (en) | A kind of method that utilization human body motion feature aids in Orientation on map | |
| CN112464738B (en) | Improved naive Bayes algorithm user behavior identification method based on mobile phone sensor | |
| CN108958482B (en) | Similarity action recognition device and method based on convolutional neural network | |
| CN114469073B (en) | Gait analysis and abnormality detection method based on wearable sensor | |
| CN110659677A (en) | Human body falling detection method based on movable sensor combination equipment | |
| CN111178155A (en) | Gait feature extraction and gait recognition method based on inertial sensor | |
| 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 | |
| Fida et al. | Real time event-based segmentation to classify locomotion activities through a single inertial sensor | |
| Paraschiv-Ionescu et al. | Real-world speed estimation using single trunk IMU: methodological challenges for impaired gait patterns | |
| CN113768471B (en) | Parkinson disease auxiliary diagnosis system based on gait analysis | |
| Falbriard et al. | Running speed estimation using shoe-worn inertial sensors: Direct integration, linear, and personalized model | |
| US20220072373A1 (en) | Wearable system and method for sensing and monitoring a swimming activity of a user | |
| KR102194313B1 (en) | Apparatus and method for identifying individuals by performing neural network analysis for various detection information | |
| CN116754114A (en) | Analysis method and system for ground contact and ground separation conditions of walking stick | |
| Howedi et al. | Accelerometer-based human fall detection using fuzzy entropy | |
| CN112762934A (en) | Lower limb movement direction prediction device and method | |
| CN208876547U (en) | A kind of gait analysis device based on IMU inertial sensor | |
| CN114601455A (en) | Motion recognition method based on two-stage neural network | |
| Wang et al. | Human motion phase segmentation based on three new features | |
| CN111337050A (en) | Zero-speed judgment condition and step counting method based on multi-condition fusion | |
| Han et al. | Motion mode recognition in multi-storey buildings based on the naive Bayes method | |
| Lin et al. | Gait Parameters Analysis Based on Leg-and-shoe-mounted IMU and Deep Learning |
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 | ||
| CB03 | Change of inventor or designer information |
Inventor after: Xu Yu Inventor after: Weng Zhixiang Inventor after: Ye Sheng Inventor after: Tang Zhenzhou Inventor before: Xu Yu Inventor before: Wen Zhixiang Inventor before: Ye Sheng Inventor before: Tang Zhenzhou |
|
| CB03 | Change of inventor or designer information |