CN105919796A - Traditional Chinese medicine massage robot system and acupoint searching method thereof - Google Patents

Abstract

The invention discloses a traditional Chinese medicine massage robot system and an acupoint searching method thereof and belongs to the field of medical service massage robots. The traditional Chinese medicine massage robot system comprises a robot, a motion control cabinet and an operating platform, wherein the robot comprises an adjusting platform, a mechanical arm and a supporting frame; the adjusting platform comprises a liftable bed and a bed board, the supporting frame is arranged at the upper part of the bed and can move in the front and back direction of the bed; the mechanical arm is arranged on the supporting frame, a human-simulated massage hand is arranged at the tail end of the mechanical arm, the motion control cabinet is arranged at the bottom of the bed, and the operating platform is located on one side of the robot. Accordingly, various massage techniques including finger pressing, finger rubbing, palm pushing, palm pressing, palm rubbing, pinching, rapping and the like can be completed for a patient. The acupoint searching method of the traditional Chinese medicine massage robot system optimizes a BP (back-propagation) neural network with a genetic algorithm, and the precision for searching acupoint coordinates of the traditional Chinese medicine massage robot system is improved.

Description

Traditional Chinese medicine massage robot system and acupuncture point finding method thereof

Technical field

The present invention relates to medical services massage robot field, be specifically related to a kind of Traditional Chinese medicine massage robot System and acupuncture point finding method thereof.

Background technology

" 12 " period Chinese population aging accelerated development, aging population situation is more acute, Progressively present aging, aging, three new features of empty nesting accelerated development.Aging population Acceleration necessarily cause aged's health problem to highlight, advanced age, anergy and the treatment of ill the elderly Nursing problem, more should obtain the common concern of society.But owing to the elderly's medical services type industry is sent out The imperfection of exhibition, makes Corresponding Countermeasures and seems the most urgent.

Along with the continuous of intelligent robot technology was improved in recent years, service robot industry have also been obtained Quickly development.Succeeding in developing of various medical treatment auxiliary robot, can effectively alleviate medical treatment at present and protect Strong resource is nervous, the problem of social population's aging aggravation.Grinding of relevant robot assisted medical technology Study carefully, the most also have become as one of international robot field forward position study hotspot.And existing market On the massage machine product succeeded in developing be required for user from main modulation position, massage manipulation is single, The effect loosened and lessen fatigue can only be played, it is impossible to play treatment or improve the effect of disease.

Therefore, in order to improve the present situation of current Chinese traditional massage treatment, doctor of chiropractic from heavy muscle power Work frees, and in view of the serious social state of astogeny, be badly in need of research a kind of for The Traditional Chinese medicine massage robot system of the elderly's DD.

But, Traditional Chinese medicine massage robot system patient is treated, massage, when massaging, need Patient's Different Acupoint is carried out the treatment massage of different dynamics, the letter fed back by pressure sensor in real time The size of breath adjusting massage dynamics.If the acupuncture point that massage robot can accurately not search out patient is sat Mark, and find the inaccuracy of acupuncture point coordinate, not only make patient be subject to pain in various degree, and And do not reach result for the treatment of.So massage robot is that patient controls to the precise positioning of massaging point coordinate The important foundation treated, uses the BP neutral net of routine to carry out acupuncture point training study, when not only learning Between long, and precision is low.

Summary of the invention

The technical problem to be solved in the present invention be to provide a kind of collection electromechanical integration, machinery, computer, Control technology in one, can complete patient to refer to by, refer to rub, slap push away, slap rub by, the palm, pinch, This Chinese traditional massage machine is found in Traditional Chinese medicine massage robot system and the raising of the multiple massage manipulation such as knocking The acupuncture point finding method of device people's system acupuncture point coordinate precision.

For solving above-mentioned technical problem, the present invention provides technical scheme as follows:

A kind of Traditional Chinese medicine massage robot system, including robot, motion control cabinet and operating platform, institute State robot and include regulating platform, mechanical arm and bracing frame, wherein:

Described regulation platform includes liftable bed body, bed board, and support frame as described above is arranged at described bed body Top, and can move along direction before and after described bed body；

Described mechanical arm is arranged on support frame as described above, and the end of described mechanical arm is provided with realizing Wrist rotation, thumb folding and the palm refer to humanoid massage's hand of conversion, described mechanical arm and humanoid massage's hand On be provided with monitoring sensor；

Described motion control cabinet is arranged at the bottom of described bed body, and described operating platform is positioned at described machine The side of people, motion that described motion control cabinet is described mechanical arm and humanoid massage's hand provide power and The transmission of data message, completes the collection of each several part monitoring sensor information, transmits and control；

Described motion control cabinet interacts with operating platform, receives the motion that described operating platform sends Instruction, and upload each several part monitoring currency of sensor, described mechanical arm and humanoid massage in real time The running state information of hand.

Further, described mechanical arm includes Cartesian coordinate type horicontal motion mechanism and vertical mechanism, with And articulated type depth motion.

Further, described operating platform uses all touch screen structure.

Further, described bed body front end is provided with face, is provided with four rolling bottom described bed body.

The acupuncture point finding method of described Traditional Chinese medicine massage robot system, including:

Step 1: determine BP neural network structure；

Step 2: utilize BP neutral net described in genetic algorithm optimization；

Step 3: predict described BP neutral net；

Step 4: complete acupuncture point coordinate and find.

Further, described step 1 includes:

Step 11: determine the number of the input of acupuncture point coordinate, output parameter, to described input, output Parameter is trained, and determines the structure of described BP neutral net；

Step 12: determine the length that genetic algorithm is individual.

Further, described input parameter be the height of patient, body weight, big vertebra, chien chung shu, Jianjing, Dazhu and fei shu cave, described output parameter is patient's shen shu, Dachangshu and Yaoyangkuan(Du 3), described input, The coordinate of output parameter is X, Y coordinate.

Further, described step 2 includes:

Step 21: the initialization of population: individual to genetic algorithm uses real coding to initialize population, Each individuality of described population all comprises a network ownership value and threshold value；

Step 22: calculate the fitness value of described individuality, described fitness value is for using described BP god The square error that obtains through network training and, computing formula is:

$F=k\left({\mathrm{\Σ}}_{i=1}^n{\left(y_i-o_i\right)}^2\right)$

In formula: n is BP neutral net output node number；y_iExpectation for neutral net i-th node Output；o_iPrediction for i-th node exports；K is coefficient；

Step 23: select operation, described select operation for the selection strategy of fitness ratio, often each and every one The select probability p of body_iFor:

f_i=k/F_i

$p_i=\frac{f_i}{{\mathrm{\Σ}}_{i=1}^N{f}_i}$

In formula: F_iFitness value for individual i；K is coefficient；N is population at individual number；

Step 24: intersect and operate, the described operation employing real number interior extrapolation method that intersects, kth chromosome a_kWith the l chromosome a_lIntersection method of operating in j position is as follows:

$\left\{\begin{array}{c}{a}_{kj}=a_{kj}(1-b)+a_{ij}b\\ a_{ij}=a_{ij}(1-b)+a_{kj}b\end{array}\right.$

In formula: b is the random number that [0,1] is interval；

Step 25: mutation operation, chooses a of the individual jth gene of i-th_ijMake a variation, adopt As follows by mutation operation method:

$a_{ij}=\left\{\begin{array}{cc}{a}_{ij}+(a_{ij}-a_{max})*f\left(g\right)& r\&GreaterEqual;0.5\\ a_{ij}+(a_{min}-a_{ij})*f\left(g\right)& r<0.5\end{array}\right.$

F (g)=r₂(1-g/G_max)

In formula: a_maxFor gene a_ijThe upper bound；a_minFor gene a_ijLower bound；r₂It it is a random number； G is current iteration number of times；G_maxIt it is maximum evolution number of times；R is the random number that [0,1] is interval；

Step 26: continue to calculate the fitness value of each individuality in described population, until finding optimum Body.

Further, described step 3 includes:

Step 31: utilize described optimum individual to described BP neutral net initial weight and threshold value assignment；

Step 32: described weights and threshold value are carried out BP neural metwork training, terminates bar until meeting Part；

Step 33: predict the coordinate output at described acupuncture point, obtain result.

Further, described step 4 includes: Traditional Chinese medicine massage robot system obtains described output parameter Coordinate value, and be mapped to the coordinate value at other acupuncture points by the coordinate value of described output parameter, obtain Method is teaching or the visual identity of manipulator.

The method have the advantages that

Compared with prior art, the operating platform of the Traditional Chinese medicine massage robot system of the present invention can be in advance The information of typing patient also shows and the acupuncture point of storage patient and data of physiological index, it is achieved to the present invention The control of Traditional Chinese medicine massage robot so that patient can be completed by the Traditional Chinese medicine massage robot of the present invention Refer to by, refer to rub, slap push away, slap rub by, the palm, pinch, the multiple massage manipulation such as knocking, meanwhile, this The acupuncture point finding method of the Traditional Chinese medicine massage robot system of invention, utilizes genetic algorithm to compensate for routine The defect of BP neutral net, shortens training learning time, it is possible to increase the prediction essence of neutral net Degree, solves optimal solution, it is ensured that the acupuncture point of patient is sat by the Traditional Chinese medicine massage robot system of the present invention The prediction that mark is made, the Traditional Chinese medicine massage robot acupuncture point coordinate that improve the present invention finds conduct programming Precision.

Accompanying drawing explanation

Fig. 1 is the structural representation of the Traditional Chinese medicine massage robot system of the present invention；

Fig. 2 is the schematic flow sheet of the Traditional Chinese medicine massage robot system acupuncture point finding method of the present invention；

Fig. 3 is the predicated error curve map of the Chinese traditional massage machine system people acupuncture point finding method of the present invention, Wherein dotted line is training objective error, and number line is optimal objective error, and real broken line is for using the present invention The training of Traditional Chinese medicine massage robot system acupuncture point finding method during the change of error.

Detailed description of the invention

For making the technical problem to be solved in the present invention, technical scheme and advantage clearer, below will knot Conjunction drawings and the specific embodiments are described in detail.

On the one hand, the present invention provides a kind of Traditional Chinese medicine massage robot system, including robot 1, motion Switch board 40 and operating platform 50, robot 1 includes regulating platform 10, mechanical arm 20 and supporting Frame 14, wherein:

Regulation platform 10 includes liftable bed body 11, bed board 12, and bracing frame 14 is arranged at bed body 11 tops, and can move along direction before and after bed body 11；

Mechanical arm 20 is arranged on bracing frame 14, and the end of mechanical arm 20 is provided with realizing wrist Portion's rotation, thumb folding and the palm refer to humanoid massage's hand 30 of conversion, mechanical arm 20 and humanoid massage's hand Upper 30 are provided with monitoring sensor；

Motion control cabinet 40 is arranged at the bottom of bed body 11, and operating platform 50 is positioned at robot 1 Side, the motion that motion control cabinet 40 is mechanical arm 20 and humanoid massage's hand 30 provides power sum It is believed that the transmission of breath, complete the collection of each several part monitoring sensor information, transmit and control；

Motion control cabinet 40 also needs to interact with operating platform 50, receives operating platform 50 and sends Movement instruction, and upload the monitoring currency of sensor, mechanical arm 20 and humanoid massage in real time The information such as the running status of hand 30.

Humanoid massage's hand 30 manual imitation gimmick can carry out Chinese traditional massage treatment, it is possible to achieve wrist turns The conversion that dynamic, the folding of thumb, the palm refer to, so realize referring to pushing away by, the palm, slap rub etc. the multiple traditional Chinese medical science by Rub gimmick, and can cooperate and freely change between various massage manipulation, thus reach right Patient therapy and the effect of massage.

In massage processes, motion control cabinet 40 receives and parses through the order that operating platform 50 issues, Obtain the information such as acupuncture point coordinate and massage manipulation, then drive mechanical arm 20 to move, reach to specify cave Position coordinate points, and adjust humanoid massage's hand 30 to massaging attitude accordingly.Rule are performed after being adjusted in place Fixed massage manipulation, in massage processes, opens defence program simultaneously, if there is jerk, spacing The situations such as triggering, motion control cabinet 40 will stop current active by transmitting order to lower levels, and mechanical arm 20 is first gone up Move and then reset, after situation releases, continue massage flow process；When completing a massaging action, automatically Proceed to next massaging action, until motion control cabinet 40 receives end order.

Motion control cabinet 40 also needs to consider electromagnetic compatibility problem, makes all metal parts keep good Contact；The electric component of driver and other sensitivities should be remained above the mounting distance of 200mm；Dynamic Power cable uses shielded cable etc., provides the guarantee of security reliability for system.

The operating platform 50 of the Traditional Chinese medicine massage robot system of the present invention can the letter of typing patient in advance Cease and show and store acupuncture point and the data of physiological index of patient, it is achieved the Chinese traditional massage machine to the present invention The control of device people so that the Traditional Chinese medicine massage robot of the present invention can complete patient to refer to by, refer to rub, The palm pushes away, slap rubbed by, the palm, pinched, the multiple massage manipulation such as knocking.

Further, mechanical arm 20 can include Cartesian coordinate type horicontal motion mechanism and vertical mechanism, And articulated type depth motion.This structure can realize the flexibility of mechanical arm 20 motion, Mechanical arm 20 can be moved along the length direction of bed body 11, to ensure in the present invention Doctor's massage robot has enough by mount space.

As a modification of the present invention, operating platform 50 can use all touch screen 51 structure.This Planting structure makes the Traditional Chinese medicine massage robot system of the present invention have human-computer interaction function.

Preferably, bed body 11 front end can be provided with face (not shown), bottom bed body 11 preferably It is provided with four rolling 13.It is downward that face is easy to face when patient massages, and improves the comfort level of patient, And can substantially determine position of human body, it is simple to the tracing and positioning at acupuncture point；The setting of roller 13 can To facilitate the movement of the Traditional Chinese medicine massage robot of the present invention.

On the other hand, the acupuncture point finding method of the Traditional Chinese medicine massage robot system of the present invention, including:

Step 1: determine BP neural network structure；

Step 2: utilize genetic algorithm optimization BP neural network；

Step 3: predict described BP neutral net；

Step 4: complete acupuncture point coordinate and find.

The present invention provides the acupuncture point finding method of above-mentioned Traditional Chinese medicine massage robot system, utilizes genetic algorithm Compensate for the defect of ordinary BP nerve network, shorten training learning time, it is possible to increase nerve net The precision of prediction of network, solves optimal solution, it is ensured that the Traditional Chinese medicine massage robot system of the present invention is to trouble The prediction that the acupuncture point coordinate of person is made, the Traditional Chinese medicine massage robot acupuncture point coordinate that improve the present invention is found The precision of conduct programming.

As a modification of the present invention, step 1 may include that

Step 11: determine the number of the input of acupuncture point coordinate, output parameter, to described input, output Parameter is trained, and determines the structure of BP neutral net；

Step 12: determine the length that genetic algorithm is individual.

Preferably, input parameter is the height of patient, body weight, big vertebra, chien chung shu, Jianjing, dazhu With fei shu cave, output parameter is patient's shen shu, Dachangshu and Yaoyangkuan(Du 3), described input, output parameter Coordinate be X, Y coordinate.

As the another kind of improvement of the present invention, step 2 may include that

Step 21: the initialization of population: individual to genetic algorithm uses real coding to initialize population, Each individuality of population all comprises a network ownership value and threshold value；

Step 22: calculate individual fitness value, fitness value is preferably and uses BP neutral net instruction The square error got and, computing formula is:

$F=k\left({\mathrm{\&Sigma;}}_{i=1}^n{\left(y_i-o_i\right)}^2\right)$

In formula: n is BP neutral net output node number；y_iExpectation for neutral net i-th node Output；o_iPrediction for i-th node exports；K is coefficient.

Step 23: select operation, described select operation for the selection strategy of fitness ratio, often each and every one The select probability p of body_iFor:

f_i=k/F_i

$p_i=\frac{f_i}{{\mathrm{\&Sigma;}}_{i=1}^N{f}_i}$

In formula: F_iFitness value for individual i；K is coefficient；N is population at individual number.Individual Select probability p_iThe biggest, then the fit of this individuality is the highest, and selected chance is the biggest.

Step 24: intersect and operate, the described operation employing real number interior extrapolation method that intersects, kth chromosome a_kWith the l chromosome a_lIntersection method of operating in j position is as follows:

$\left\{\begin{array}{c}{a}_{kj}=a_{kj}(1-b)+a_{ij}b\\ a_{ij}=a_{ij}(1-b)+a_{kj}b\end{array}\right.$

In formula: b is the random number that [0,1] is interval.The purpose operated of intersecting is able to the next generation Producing new individuality, it is the most important means that genetic algorithm obtains new defect individual.

Step 25: mutation operation, chooses a of the individual jth gene of i-th_ijMake a variation, adopt As follows by mutation operation method:

$a_{ij}=\left\{\begin{array}{cc}{a}_{ij}+(a_{ij}-a_{max})*f\left(g\right)& r\&GreaterEqual;0.5\\ a_{ij}+(a_{min}-a_{ij})*f\left(g\right)& r<0.5\end{array}\right.$

F (g)=r₂(1-g/G_max)

In formula: a_maxFor gene a_ijThe upper bound；a_minFor gene a_ijLower bound；r₂It it is a random number； G is current iteration number of times；G_maxIt it is maximum evolution number of times；R is the random number that [0,1] is interval.Variation behaviour Work is to produce new a kind of individual method, is a requisite calculation step in genetic algorithm, protects Demonstrate,prove the validity of genetic algorithm, made genetic algorithm have the random searching ability of local, be to lose simultaneously Propagation algorithm keeps the diversity of population, prevents the convergence of non-adult form.

Step 26: continue to calculate the fitness value of each individuality in described population, until finding optimum Body.

Above-mentioned steps 21-26, by selection, intersection and mutation operation in genetic algorithm, selects adaptation The individuality that angle value is the highest, i.e. optimum individual, and obtain weights and the threshold value of this optimum individual.

Further, step 3 may include that

Step 31: utilize optimum individual to BP neutral net initial weight and threshold value assignment；

Step 32: weights and threshold value are carried out BP neural metwork training, until meeting termination condition； This termination condition can be intended minimal error.

Step 33: the coordinate output at prediction acupuncture point, obtains result.

It addition, after acupuncture point has been found, when using the Traditional Chinese medicine massage robot systematic treating of the present invention, Can i.e. map to the coordinate value at other acupuncture points by obtaining the coordinate value at least three acupuncture point, obtain Access method can be teaching or the visual identity of manipulator.Manipulator teaching is preferably by artificial dilatory Or remote control mode determines the coordinate value of at least three human acupoint, visual identity is stickup and is indicated to The coordinate value of few three human acupoints.The method avoids and be manually entered coordinate value and human acupoint The determination of initial point.

The specific embodiment of acupuncture point coordinate finding method is adopted by the Traditional Chinese medicine massage robot system of the present invention With three-layer neural network, by the height (cm) of patient, body weight (kg), big vertebra (X, Y coordinate, Lower with), chien chung shu, External Shoulder Shu, Jianjing, big, sky ancestor and Feiyu acupuncture point as input parameter, Coordinate is as follows:

Using shen shu, Dachangshu and Yaoyangkuan(Du 3) acupuncture point as output parameter, as follows:

Above parameter is trained, owing to these parameters have different magnitudes and dimension, therefore needs These data are normalized.In the embodiment of the present invention, owing to sample has 16 inputs Parameter, 6 output parameters, again due to hidden layer neutral net number n₂With input layer number n₁Between there is following approximation relation:

n₂=2 × n₁+1

So hidden layer neutral net number n in the present embodiment₂It is 33, set BP nerve net The structure of network is 16-33-6, and namely input layer has 16 nodes, and hidden layer has 33 nodes, Output layer has 6 nodes, has 16 × 33+33 × 6=726 weights, 33+6=39 threshold value, So the number of parameters of genetic algorithm optimization is 726+39=765 in the present embodiment.

Utilizing data below as the test data of sample, input parameter is as follows:

Output parameter is as follows:

Owing to these parameters have different magnitudes and dimension, it is therefore desirable to these data are carried out normalizing Change processes.In the present invention, hidden layer must transmit function and choose S type tan tansig simultaneously；Output Layer chooses S type logarithmic function logsig.The network structure of the present embodiment is 16-33-6, then weights and The number of threshold value is as follows:

The weights of the present embodiment and threshold value all use 10 binary system real codings, then individual two enter The code length of system is 7650.Wherein, input layer and hidden layer connect weights be encoded to front 5280 Position, hidden layer threshold coding is 5281～5610, and hidden layer is encoded to the connection weights of output layer 5611～7590, output layer threshold coding is 7591～7650.

As it is shown on figure 3, wherein dotted line is training objective error, number line is optimal objective error, real Broken line is to use error during the Traditional Chinese medicine massage robot system acupuncture point finding method training of the present invention Change, as can be seen from Figure 3 just can reach intended minimum mistake by the iterative steps of 7 steps Difference, it follows that by the Traditional Chinese medicine massage robot system acupuncture point finding method of the present invention can comparatively fast, Complete the searching of the acupuncture point coordinate of patient more accurately, improve the Traditional Chinese medicine massage robot system of the present invention The precision that the acupuncture point coordinate to patient of uniting is found.

The above is the preferred embodiment of the present invention, it is noted that general for the art For logical technical staff, on the premise of without departing from principle of the present invention, it is also possible to make some changing Entering and retouch, these improvements and modifications also should be regarded as protection scope of the present invention.

Claims (9)

1. a Traditional Chinese medicine massage robot system, it is characterised in that include robot, motion control cabinet And operating platform, described robot includes regulating platform, mechanical arm and bracing frame, wherein:

Described regulation platform includes liftable bed body, bed board, and support frame as described above is arranged at described bed body Top, and can move along direction before and after described bed body；

Described mechanical arm is arranged on support frame as described above, and the end of described mechanical arm is provided with realizing Wrist rotation, thumb folding and the palm refer to humanoid massage's hand of conversion, described mechanical arm and humanoid massage's hand On be provided with monitoring sensor；

Described motion control cabinet is arranged at the bottom of described bed body, and described operating platform is positioned at described machine The side of people, motion that described motion control cabinet is described mechanical arm and humanoid massage's hand provide power and The transmission of data message, completes the collection of each several part monitoring sensor information, transmits and control；

Described motion control cabinet interacts with operating platform, receives the motion that described operating platform sends Instruction, and upload each several part monitoring currency of sensor, described mechanical arm and humanoid massage in real time The running state information of hand.

Traditional Chinese medicine massage robot system the most according to claim 1, it is characterised in that described machine Mechanical arm includes Cartesian coordinate type horicontal motion mechanism and vertical mechanism, and articulated type depth motion.

Traditional Chinese medicine massage robot system the most according to claim 2, it is characterised in that described behaviour Make platform and use all touch screen structure.

Traditional Chinese medicine massage robot system the most according to claim 3, it is characterised in that described bed Body front end is provided with face, is provided with four rolling bottom described bed body.

5. find according to the acupuncture point of described Traditional Chinese medicine massage robot system arbitrary in Claims 1-4 Method, it is characterised in that including:

Step 1: determine BP neural network structure；

Step 2: utilize BP neutral net described in genetic algorithm optimization；

Step 3: predict described BP neutral net；

Step 4: complete acupuncture point coordinate and find.

Acupuncture point the most according to claim 5 finding method, it is characterised in that described step 1 Including:

Step 11: determine the number of the input of acupuncture point coordinate, output parameter, to described input, output Parameter is trained, and determines the structure of described BP neutral net；

Step 12: determine the length that genetic algorithm is individual.

Acupuncture point the most according to claim 6 finding method, it is characterised in that described input parameter For height, body weight, big vertebra, chien chung shu, Jianjing, dazhu and the fei shu cave of patient, described output is joined Number is patient's shen shu, Dachangshu and Yaoyangkuan(Du 3), and described input, the coordinate of output parameter are that X, Y sit Mark.

Acupuncture point the most according to claim 7 finding method, it is characterised in that described step 2 Including:

Step 21: the initialization of population: individual to genetic algorithm uses real coding to initialize population, Each individuality of described population all comprises a network ownership value and threshold value；

Step 22: calculate the fitness value of described individuality, described fitness value is for using described BP god The square error that obtains through network training and, computing formula is:

$F=k\left({\mathrm{\&Sigma;}}_{i=1}^n{\left(y_i-o_i\right)}^2\right)$

In formula: n is BP neutral net output node number；y_iExpectation for neutral net i-th node Output；o_iPrediction for i-th node exports；K is coefficient；

Step 23: select operation, described select operation for the selection strategy of fitness ratio, often each and every one The select probability p of body_iFor:

f_i=k/F_i

$p_i=\frac{f_i}{{\mathrm{\&Sigma;}}_{i=1}^N{f}_i}$

In formula: F_iFitness value for individual i；K is coefficient；N is population at individual number；

Step 24: intersect and operate, the described operation employing real number interior extrapolation method that intersects, kth chromosome a_kWith the l chromosome a_lIntersection method of operating in j position is as follows:

$\left\{\begin{array}{c}{a}_{kj}=a_{kj}(1-b)+a_{ij}b\\ a_{ij}=a_{ij}(1-b)+a_{kj}b\end{array}\right.$

In formula: b is the random number that [0,1] is interval；

Step 25: mutation operation, chooses a of the individual jth gene of i-th_ijMake a variation, adopt As follows by mutation operation method:

$a_{ij}=\left\{\begin{array}{cc}{a}_{ij}+(a_{ij}-a_{max})*f\left(g\right)& r\&GreaterEqual;0.5\\ a_{ij}+(a_{min}-a_{ij})*f\left(g\right)& r<0.5\end{array}\right.$

F (g)=r₂(1-g/G_max)

In formula: a_maxFor gene a_ijThe upper bound；a_minFor gene a_ijLower bound；r₂It it is a random number； G is current iteration number of times；G_maxIt it is maximum evolution number of times；R is the random number that [0,1] is interval；

Step 26: continue to calculate the fitness value of each individuality in described population, until finding optimum Body.

Acupuncture point the most according to claim 8 finding method, it is characterised in that described step 3 Including:

Step 31: utilize described optimum individual to described BP neutral net initial weight and threshold value assignment；

Step 32: described weights and threshold value are carried out BP neural metwork training, terminates bar until meeting Part；

Step 33: predict the coordinate output at described acupuncture point, obtain result.