Summary of the invention
Technical problem to be solved by this invention is to provide a kind of limb motion to detect critic network system and method thereof, utilize the present invention for user, train according to the target trajectory curve in system template storehouse, or self limb motion is detected and assessed.
For realizing object of the present invention, the invention provides a kind of limb motion and detect critic network system, comprising: CPU, and be connected to the sensing network system of at least two sensing units compositions of CPU.
Sensing unit attaches on limbs, for detection of and record limbs real time kinematics parameter, and be transferred to CPU;
CPU, for receiving the limbs real time kinematics parameter of sensing unit transmission, and processes and assesses the kinestate of limbs according to the limbs real time kinematics parameter receiving.
More preferably, limb motion of the present invention detects critic network system, and sensing unit comprises: inertial sensor and sensing communication module.
Inertial sensor, for according to the detection instruction of CPU, detects and records limbs real time kinematics parameter.
Sensing communication module, for receiving and transmit the detection instruction of CPU, and is sent to CPU by inertial sensor according to the limbs real time kinematics parameter that detects command detection record.
Detect the beginning recording instruction that instruction comprises limb motion, and limb motion number of times or time parameter instruction.
More preferably, limb motion of the present invention detects critic network system, and inertial sensor is XYZ tri-axle inertial sensors.
More preferably, limb motion of the present invention detects critic network system, and CPU comprises: data communication module and data processing module.
Data communication module, the limbs real time kinematics parameter sending for receiving sensing communication module; Send the detection instruction of CPU to sensing unit; Also for the data processed result that sends data processing module to remote terminal, and receiving remote instruction.
Data processing module, be used for sending limb motion control instruction, and according to the limb motion parameter receiving, judge the type of sports that limbs carry out, analyze the limb motion cycle, obtain the real time kinematics geometric locus of limb motion, and calculate the quantity of same limb motion type, and obtain the assessment result of limb motion state according to the quantity assessment of the real time execution geometric locus of limb motion and same limb motion type.
More preferably, limb motion of the present invention detects critic network system, and CPU also comprises data memory module and display module.
Data memory module, for store limb motion type default and that show and described limb motion type space geometric locus data on display module, and the limb motion parameter detecting, and the quantity of the limbs real time kinematics geometric locus going out according to limb motion calculation of parameter and same limb motion type, assessment result.
Display module, for the control instruction of sending according to data processing module, the limb motion type of storing in reading out data memory module the target trajectory curve that shows limb motion type, and according to data processing module, limbs real time kinematics parameter is received and evaluation process after fructufy time show times of exercise, the limbs real time kinematics geometric locus of limbs real time kinematics parameter, same limb motion type and assessment result.
More preferably, limb motion of the present invention detects critic network system, kinematic parameter comprises: X and the Y-axis angular speed of the limbs of each sensing unit record, speed and the linear acceleration etc. of X and Y direction or X, Y and Z axis angular speed, the speed of X, Y and Z-direction and linear acceleration.
The present invention also provides a kind of limb motion to detect appraisal procedure, comprises the following steps:
Steps A, attaches at least two sensing units on limbs, detects and record limb motion parameter, and is transferred to CPU;
Step B, CPU receives described sensing unit and transmits the limb motion parameter of returning, and processes and assess the kinestate of limbs according to the limb motion parameter receiving.
More preferably, limb motion of the present invention detects appraisal procedure, and steps A comprises the following steps:
Steps A 1, attaches at least two sensing units on the limbs that will move;
Steps A 2, number of times or the time of setting limb motion;
Steps A 3, the inertial sensor in sensing unit detects and records the real time kinematics parameter of limbs;
Steps A 4, completes the limb motion number of times of setting or after the time, the user's who records limbs real time kinematics parameter is sent to CPU by the sensing communication module in sensing unit.
More preferably, limb motion of the present invention detects appraisal procedure, further comprising the steps of between steps A 2 and A3:
Steps A 21, the data processing module of CPU sends demonstration control instruction to the display module of CPU, and the default limb motion type of storing in display module reading out data memory module and described limb motion type space geometric locus data also show;
Steps A 22, CPU sends detection record instruction to sensing unit, and the space tracking curve of the limb motion type that limbs show according to steps A 21 moves.
More preferably, limb motion of the present invention detects appraisal procedure, and steps A 21 comprises the following steps:
Steps A 211, selects at least one limb motion type the multiple limb motion type showing from the display module of CPU;
Steps A 212: on the display module of CPU, the selected target trajectory curve of limb motion type space geometric locus data show, target trajectory curve is to have the level and smooth sinusoidal wave signal of periodicity.
More preferably, limb motion of the present invention detects appraisal procedure, and step B comprises the following steps:
Step B1, the data communication module of CPU receives the real time kinematics parameter of limb motion, and sends real time kinematics parameter to the data processing module of CPU and the data memory module of CPU;
Step B2, data memory module stores real time kinematics parameter; Data processing module carries out the processing of limb motion parameter and assesses the kinestate of limbs.
More preferably, limb motion of the present invention detects appraisal procedure, in step B2, carries out the processing of limb motion parameter and assesses the kinestate of limbs, comprises the following steps:
Step B21, judges the type of sports that limbs carry out;
Step B22, analyzes the limb motion cycle, obtains the real time kinematics geometric locus of limb motion;
Step B23, calculates the quantity of same limb motion type;
Step B24, assesses the assessment result that obtains limb motion state according to the quantity of the real time kinematics geometric locus of limb motion and same limb motion type.
More preferably, limb motion of the present invention detects appraisal procedure, and in step B21, assessment obtains the assessment result of limb motion state, comprises the steps:
Step B211, according to receive at least 6 real-time limb motion parameters, respectively with default limb motion type sample template base in the kinematic parameter of each type of sports template samples carry out computing cross-correlation, obtain multiple cross correlation results;
Step B212, carries out K-arest neighbors classified counting to each cross correlation results, draws the distance of real-time limb motion parameter to each template samples;
Step B213, reads the labelling of K template samples minimum in described distance, and obtains the labelling of real-time limb motion test sample book according to the labelling of this K template samples, thereby judges the limb motion type under current limb motion.
More preferably, limb motion of the present invention detects appraisal procedure, and step B23 comprises the steps:
Step B231, by the amplitude normalization to 0 of at least 6 real-time limb motion parameters that detect and record to 1;
Step B232, by Fourier transformation or wavelet analysis periodicity analysis method, calculates the periodicity of the actual motion of real-time limb motion;
Step B233, according to the amplitude of user's actual motion, the cycle is drawn real time kinematics geometric locus;
Step B234, the contrast geometric locus of actual motion and the deviation of target trajectory curve, and the quantity of limb motion, complete the assessment of the kinestate of limb motion, obtains assessment result.
More preferably, limb motion of the present invention detects appraisal procedure, and K value minimum is 11.
More preferably, limb motion of the present invention detects appraisal procedure, and in step B234, contrasting the computational methods that the geometric locus of actual motion and the deviation of target trajectory curve adopt is correlation coefficient process;
The coordinate ordered series of numbers T={{x of target trajectory curve
1, y
1, z
1, { x
2, y
2, z
2..., { x
n, y
n, z
n, the coordinate ordered series of numbers S={{x ' of actual motion geometric locus
1, y '
1, z '
1, x '
2, y '
2, z '
2..., x '
n, y '
n, z '
n, n measures number of samples, and correlation coefficient r is:
The value of correlation coefficient r between-1 and+1 between ,-1≤r≤+ 1;
| r|=1, represent that T ordered series of numbers and S ordered series of numbers are complete linear correlation, be functional relationship, show that actual motion geometric locus and target trajectory curve now match;
R=0, represents that T ordered series of numbers and S ordered series of numbers, without linear dependency relation, show that actual motion geometric locus and target trajectory curve now do not match completely;
| r| > 0, represents that T ordered series of numbers is relevant to S ordered series of numbers; | r| rises at 1 o'clock from 0, shows that the quality of hands-on campaign is improving gradually.
More preferably, limb motion of the present invention detects appraisal procedure, and in step B234, contrasting the computational methods that the geometric locus of actual motion and the deviation of target trajectory curve adopt is mean error quadratic method;
The coordinate ordered series of numbers T={{x of target trajectory curve
1, y
1, z
1, { x
2, y
2, z
2..., { x
n, y
n, z
n, the coordinate ordered series of numbers S={{x ' of actual motion geometric locus
1, y '
1, z '
1, x '
2, y '
2, z '
2..., x '
n, y '
n, z '
n, n measures number of samples, and coefficient of correlation r is:
The numerical value of r is between 0 and 1;
R=0, indicates to occur without actual motion;
R=1, represents that two movement locus match, and training moving-mass is very high;
R rises at 1 o'clock from 0, shows that the quality of hands-on campaign is improving gradually.
Beneficial effect of the present invention: utilize limb motion of the present invention to detect critic network system and method, target trajectory curve in the comparable lighting system template base of user is trained, also can self limb motion be detected and be assessed, cost be low, also saves time.
Detailed description of the invention
In order to make object, technical scheme and the advantage of limb motion detection critic network system of the present invention and method thereof clearer, below in conjunction with concrete drawings and the specific embodiments, limb motion detection critic network system of the present invention and method thereof are further elaborated.
Fig. 1 is the structural representation of the embodiment of the present invention, and as shown in Figure 1, limb motion detects critic network system, comprising: CPU 20, and be connected at least two sensing units 10 of CPU 20.
Sensing unit 10 attaches on limbs, for detection of and record limbs real time kinematics parameter, and be transferred to CPU.
Sensing unit 10 comprises: inertial sensor 101 and sensing communication module 102.
Inertial sensor 101, for according to the detection instruction of CPU 20, detects and records limbs real time kinematics parameter.
More preferably, inertial sensor 101 is XYZ tri-axle inertial sensors.
In the present embodiment, inertial sensor 101 is XYZ tri-axle inertial sensors.
Sensing communication module 102, for receiving and transmit the detection instruction of CPU 20, and is sent to CPU 20 by inertial sensor 101 according to the limbs real time kinematics parameter that detects command detection record.
In the present embodiment, as a kind of embodiment, described sensing communication module 102 can be a 802.15.4 radio receiving transmitting module, is realized by XBee module, also can be realized by the CC2530 chip of Texas Instrument or other suitable ZigBee module/chip.
Detect the beginning recording instruction that instruction comprises limb motion, and limb motion number of times or time parameter instruction.
CPU 20, the limbs real time kinematics parameter of transmitting for receiving sensing unit 10, and process and assess the kinestate of limbs according to the limbs real time kinematics parameter receiving.
CPU 20 comprises: data communication module 201, data processing module 202, data memory module 203 and display module 204.
Data communication module 201, the limbs real time kinematics parameter sending for receiving sensing communication module 102; Send the detection instruction of CPU 20 to sensing unit 10; Also for the data processed result that sends data processing module 202 to remote terminal, and receiving remote instruction.
Data processing module 202, be used for sending limb motion control instruction, and according to the limb motion parameter receiving, judge the type of sports that limbs carry out, analyze the limb motion cycle, obtain the real time kinematics geometric locus of limb motion, and calculate the quantity of same limb motion type, and obtain the assessment result of limb motion state according to the quantity assessment of the real time execution geometric locus of limb motion and same limb motion type.
Data memory module 203, for store limb motion type and limb motion type space geometric locus data default and that show on display module 204, and the limb motion parameter detecting, and the quantity of the limbs real time kinematics geometric locus going out according to limb motion calculation of parameter and same limb motion type, assessment result.
Display module 204, for the control instruction of sending according to data processing module 202, the limb motion type of storage show the target trajectory curve of limb motion type in reading out data memory module 203, and according to data processing module 203, limbs real time kinematics parameter is received and evaluation process after fructufy time show times of exercise, the limbs real time kinematics geometric locus of limbs real time kinematics parameter, same limb motion type and assessment result.
In the present embodiment, as a kind of embodiment, described data communication module 201 can be a 802.15.4 radio receiving transmitting module, is realized by XBee module, also can be realized by the CC2530 chip of Texas Instrument or other suitable ZigBee module/chip.Data processing module 202, data memory module 203 and display module 204 are realized by a computer, and data processing module 202 is computer CPU, and data memory module 203 is computer hard disc, and display module 204 is computer monitor.CPU also can be realized by a smart mobile phone or a panel computer.
The limb motion of the present embodiment detects appraisal procedure, comprises the following steps:
Steps A, attaches at least two sensing units 10 on limbs, detects and record limb motion parameter, and is transferred to CPU 20;
In the present embodiment, select three sensing units 10, attached and user's wrist respectively, on forearm and postbrachium.
Step B, CPU 20 receives the limb motion parameter that described sensing unit 10 transmission are returned, and processes and assess the kinestate of limbs according to the limb motion parameter receiving.
The limb motion parameter of passing back in the present embodiment comprises: the attached sensing unit place limbs that three sensing units record are respectively in the angular speed of X, Y and Z axis, the speed of X, Y and Z-direction and linear acceleration.
More preferably, steps A comprises the following steps:
Steps A 1, attaches at least two sensing units 10 on the limbs that will move;
In the present embodiment, select three sensing units 10, attach to respectively user's wrist, on forearm and postbrachium.
Steps A 2, number of times or the time of setting limb motion;
In the present embodiment, set user completed 5 upper limb sides and lifts in 10 seconds.
Steps A 3, the inertial sensor 101 in sensing unit 10 detects and records the real time kinematics parameter of limbs;
In the present embodiment, attach to respectively user's wrist, the inertial sensor 101 in the sensing unit 10 on forearm and postbrachium detects respectively and recording user wrist, and forearm and postbrachium be at the angular speed of X, Y and Z axis, the speed of X, Y and Z-direction and linear acceleration.
Steps A 4, completes the limb motion number of times of setting or after the time, the user's who records limbs real time kinematics parameter is sent to CPU 20 by the sensing communication module 102 in sensing unit 10.
After user completed 5 upper limb sides and lifts in 10 seconds in the present embodiment, attach to user's wrist, the user's who records limbs real time kinematics parameter is sent to CPU 20 by the sensing communication module 102 in 3 sensing units 10 on forearm and postbrachium.
More preferably, further comprising the steps of between steps A 2 and A3:
Steps A 21, the data processing module 202 of CPU 20 sends demonstration control instruction to the display module 204 of CPU 20, the default limb motion type of storage and limb motion type space geometric locus data demonstration in display module 204 reading out data memory modules 203;
In the present embodiment, display module 204 will show that limb motion type is the space tracking curve that upper limb side is lifted and upper limb side is lifted.
Steps A 22, CPU 20 sends detection record instruction to sensing unit 10, and the space tracking curve of the limb motion type that limbs show according to steps A 21 moves.
In the present embodiment, CPU 20 sends detection record instruction to sensing unit 10, and the space tracking curve that the upper limb side that user shows according to steps A 21 is lifted moves.
More preferably, steps A 21 comprises the following steps:
Steps A 211, selects at least one limb motion type the multiple limb motion type showing from the display module 204 of CPU 20;
In the present embodiment, the multiple limb motion type that user shows from display module 204, select upper limb side to lift.
Steps A 212: on the display module 204 of CPU 20, the selected target trajectory curve of limb motion type space geometric locus data show, described target trajectory curve is illustrated with level and smooth sine wave or other smoothed curves similarly with periodicity with periodicity.
In the present embodiment, display module 204 shows the target trajectory curve that upper limb side is lifted.
More preferably, step B comprises the following steps:
Step B1, the data communication module 201 of CPU 20 receives the real time kinematics parameter of limb motion, and sends real time kinematics parameter to the data processing module 202 of CPU 20 and the data memory module 203 of CPU 20;
Step B2, data memory module 203 stores real time kinematics parameter; Data processing module 202 carries out the processing of limb motion parameter and assesses the kinestate of limbs.
More preferably, in step B2, carry out the processing of limb motion parameter and assess the kinestate of limbs, comprise the following steps:
Step B21, judges the type of sports that limbs carry out;
In the present embodiment, it is that upper limb side is lifted that data processing module 202 is judged the ongoing motion of user by K-arest neighbors classified counting.
Step B22, analyzes the limb motion cycle, obtains the real time kinematics geometric locus of limb motion;
In the present embodiment, data processing module 202, by Fourier transformation analytical method, calculates the periodicity of the actual motion of real-time limb motion.
Step B23, calculates the quantity of same limb motion type;
Step B24, assesses the assessment result that obtains limb motion state according to the quantity of the real time kinematics geometric locus of limb motion and same limb motion type.
In the present embodiment, data processing module 202 carries out upper limb side according to user and lifts the real time kinematics geometric locus of motion and complete quantity assessment that upper limb side lifts the assessment result that obtains limb motion state.
More preferably, in step B21, assessment obtains the assessment result of limb motion state, comprises the steps:
Step B211, according to receive at least 6 real-time limb motion parameters, respectively with default limb motion type sample template base in the kinematic parameter of each type of sports template samples carry out computing cross-correlation, obtain multiple cross correlation results;
In the present embodiment, CPU 20 receives 9 real-time upper extremity exercise parameters, respectively with default limb motion type sample template base in the upper limb side kinematic parameter of lifting template samples carry out computing cross-correlation, obtain multiple cross correlation results.
Step B212, carries out K-arest neighbors classified counting to each cross correlation results, draws the distance of real-time limb motion parameter to each template samples;
Step B213, reads the labelling of K template samples minimum in described distance, and obtains the labelling of real-time limb motion test sample book according to the labelling of this K template samples, thereby judges the limb motion type under current limb motion.
As shown in table 1, use different K values, the different classify accuracy that K-arest neighbors sorting algorithm calculates, K value is 3,5,7,9 o'clock, all there will be once erroneous judgement, in the time that K value is 11, no longer occurs erroneous judgement, rate of accuracy reached to 100%.In the present embodiment, K value is 11.
The different classify accuracy tables that table 1K-arest neighbors sorting algorithm calculates
K value |
3 |
5 |
7 |
9 |
11 |
Erroneous judgement number |
1 |
1 |
1 |
1 |
0 |
Accuracy rate (%) |
97.2 |
97.2 |
97.2 |
97.2 |
100 |
More preferably, step B23 comprises the steps:
Step B231, by the amplitude normalization to 0 of at least 6 real-time limb motion parameters that detect and record to 1;
In the present embodiment, by the amplitude normalization to 0 of 9 real-time limb motion parameters that detect and record to 1.
Step B232, by Fourier transformation or wavelet analysis periodicity analysis method, calculates the periodicity of the actual motion of real-time limb motion;
In the present embodiment, adopt the analytical method of Fourier transformation.
Step B233, according to the amplitude of user's actual motion, the cycle is drawn real time kinematics geometric locus;
Step B234, the contrast geometric locus of actual motion and the deviation of target trajectory curve, and the quantity of limb motion, complete the assessment of the kinestate of limb motion, obtains assessment result.(as shown in Figure 2) actual motion geometric locus and target trajectory curve show simultaneously, can intuitively contrast actual motion performance.
As a kind of embodiment, in the present embodiment, adopt the correlation coefficient process contrast geometric locus of actual motion and the deviation of target trajectory curve.
The coordinate ordered series of numbers T={{x of target trajectory curve
1, y
1, z
1, { x
2, y
2, z
2..., { x
n, y
n, z
n, the coordinate ordered series of numbers S={{x ' of actual motion geometric locus
1, y '
1, z '
1, x '
2, y '
2, z '
2..., x '
n, y '
n, z '
n, n measures number of samples, and correlation coefficient r is:
The value of correlation coefficient r between-1 and+1 between ,-1≤r≤+ 1;
| r|=1, represent that T ordered series of numbers and S ordered series of numbers are complete linear correlation, be functional relationship, show that actual motion geometric locus and target trajectory curve now match;
R=0, represents that T ordered series of numbers and S ordered series of numbers, without linear dependency relation, show that actual motion geometric locus and target trajectory curve now do not match completely;
| r| > 0, represents that T ordered series of numbers is relevant to S ordered series of numbers; | r| rises at 1 o'clock from 0, shows that the quality of hands-on campaign is improving gradually.
As a kind of embodiment, the contrast geometric locus of actual motion and the deviation of target trajectory curve also can adopt mean error quadratic method.
The coordinate ordered series of numbers T={{x of target trajectory curve
1, y
1, z
1, { x
2, y
2, z
2..., { x
n, y
n, z
n, the coordinate ordered series of numbers S={{x ' of actual motion geometric locus
1, y '
1, z '
1, x '
2, y '
2, z '
2..., x '
n, y '
n, z '
n, n measures number of samples, and coefficient of correlation r is:
The numerical value of r is between 0 and 1;
R=0, indicates to occur without actual motion;
R=1, represents that two movement locus match, and training moving-mass is very high;
R rises at 1 o'clock from 0, shows that the quality of hands-on campaign is improving gradually.
More preferably, in order to ensure the accuracy of the judgement to limb motion type, K value minimum is 11.
In the present embodiment, CPU 20 is carried out user upper limb side and lifts the real time kinematics parameter of motion, the geometric locus of actual motion and assessment result (also can be passed through GPRS/GSM by the Internet, WIFI, the means of communication such as 3G) be sent on hospital or rehabilitation center or rehabilitation expert's server or terminal unit, doctor or rehabilitation expert assess and provide rehabilitation suggestion according to above-mentioned information to user's current motion, then assessment and suggestion are sent to CPU 20, be stored in data memory module 203 and by display module 204 and present to user.
Doctor or rehabilitation expert can also by the Internet (also can pass through GPRS/GSM, WIFI, the means of communication such as 3G), from data memory module 203, inquiring user carries out the historical record of rehabilitation exercise.
Fig. 3 is the target trajectory curve chart of the second embodiment of the present invention, as shown in Figure 3, has listed four kinds of different classes of upper extremity exercise classifications measuring with two XYZ, tri-axle inertial sensors in figure.Each sports category is at X, Y, and Z direction has the two suite lines that provided by different XYZ tri-axle inertial sensors (black with Lycoperdon polymorphum Vitt).These curves have formed the feature of identifying these sports category.
The limb motion of the embodiment of the present invention detects critic network system and method, and different target trajectories has different paths, represents different training methodes and path.User is following target trajectory and is carrying out training, and as far as possible and target trajectory coincide, minimizing error.Mistake extent has shown the extent of damage of user's upper extremity function.So, can carry out rehabilitation training, also can check rehabilitation situation according to the assessment result of training.Limb motion of the present invention detects 9 kinematic parameters that critic network system records by inertia sensing unit inspection, while and pre-recorded template set are carried out match search or are carried out other the retrieving algorithm based on signal processing and artificial intelligence, conclude the classification of the ongoing motion of user, and and then calculate and other range deviation of this target class, also real-time definite user's upper limb is for the transient position of relatively static health simultaneously, and the position of these transient states can form again the actual motion track of upper extremity exercise; System calculates the periodicity of user's upper extremity exercise by the periodicity analysis of the track to the continuous motion of the upper limb of recording, again according to the periodicity of upper extremity exercise, determine that also recording user upper limb arm carries out the number of times of rehabilitation exercise, thereby judge the intensity that user carries out rehabilitation training.User can, according to the actual physical ability of oneself, carry out the repeating motion of some strength, i.e. the periodic movement of some.Also can, in the time that patient carries out clinical diagnosis to hospital, be set in advance by physiatrician, and can, in training course for the treatment of subsequently, adjust at any time.
The foregoing is only preferred example of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any amendment of making, be equal to replacement, improve, all should be included in protection scope of the present invention.