1. An anti-drowning alarm system based on an artificial immune algorithm is characterized in that the anti-drowning alarm system is formed by remotely connecting an anti-drowning bracelet (8) and a background server (9) through a 4G module (7); prevent having inlayed in the drowned bracelet singlechip (4) and can carrying out drowned characteristic identification's heart rate measuring module (1) with this singlechip (4) connection, be used for measuring hydraulic pressure measuring module (2), be used for realizing GPS module (3) of location, LED module (5), bee calling organ module (6) and 4G module (7), wherein singlechip (4) can transmit backstage server (9) with data through 4G module (7), carry out alarm processing by backstage server (9).

2. The drowning prevention alarm system based on the artificial immune algorithm as claimed in claim 1, characterized in that the background server (9) performs alarm processing through mobile phone APP push (10), short message notification (11) and emergency call (12) dialing.

3. The drowning prevention alarm system based on the artificial immune algorithm as claimed in claim 1, wherein the single chip microcomputer (4) is AT89C51, the heart rate measurement module (1) is MAX30100, the pressure measurement module (2) is a diffused silicon pressure transmitter, and the information collected by the heart rate measurement module (1), the pressure measurement module (2) and the GPS module (3) is processed and judged by the single chip microcomputer (4) and then transmitted to the background server (9) through the 4G module (7) to give an alarm.

4. The drowning prevention alarm system based on the artificial immune algorithm as claimed in claim 1, wherein when drowning is detected, the GPS module (3) will perform data acquisition and then perform drowning characteristic recognition by the heart rate measuring module (1), and meanwhile, the LED module (5) will flash and the buzzer will alarm (6).

5. An artificial immune algorithm based drowning prevention alarm system according to claim 1, characterized in that it works as follows:

when a person wearing the drowning prevention bracelet (8) enters water, the pressure measurement module (2) and the heart rate measurement module (1) work simultaneously to perform real-time drowning detection and transmit collected information to the single chip microcomputer (4) to perform drowning identification on a wearer through a drowning identification algorithm, if the wearer drowns, the single chip microcomputer (4) identifies a drowning state, the single chip microcomputer controls the GPS module (3) to collect position information, the single chip microcomputer starts the LED module (5) and the buzzer module (6) to alarm, surrounding pedestrians are noticed, if the drowning state is not relieved after 5 seconds, the single chip microcomputer controls the 4G module (9) to transmit alarm information to the background server (9), and the background server (9) can automatically edit the positioning information into a short message notification (11) to send to an emergency contact person or push the short message (11) through a mobile phone APP (10) after receiving the alarm information, an emergency call (12) is also dialed and location information is also transmitted.

6. The drowning prevention alarm system based on the artificial immune algorithm as claimed in claim 1, wherein the way of the drowning characteristic recognition by the heart rate measurement module (1) is as follows: firstly, binary processing is carried out on heart rate data, and the heart rate within 10 seconds is judged to be smaller than 60 or larger than 180 and is recorded as 1, namely, the drowning characteristic; otherwise, a 0 is recorded, i.e. not drowned signature, yielding a 10 second heart rate binary antigen.

7. The drowning prevention alarm system based on the artificial immune algorithm as claimed in claim 6, wherein the single chip microcomputer (4) performs antigen basic feature classification learning according to the drowning feature identified by the heart rate measuring module (1), and specifically comprises the following steps:

step S71, an antibody detector generation stage, which respectively learns the characteristics of the autologous and allogeneic antibodies to form corresponding antibody libraries, wherein the definitions of the autologous and allogeneic antibodies are as follows:

autologous characteristics (drowning): the heart rate is less than 60 or the heart rate is more than 180 within 10 seconds, and accounts for 80 percent or more of the total heart rate;

foreign body characteristics (normal): the heart rate is more than 60 and the heart rate is less than 180 within 10 seconds, and accounts for 80 percent or more of the total heart rate;

step S72, randomly selecting and generating a corresponding antibody library after antigen classification is completed according to the autologous characteristics and the allogeneic characteristics;

step S73, calculating the affinity of the antigen and the antibody, and judging whether the affinity meets the condition, namely the affinity is more than 90 percent, generating a corresponding initial antibody library when the affinity meets the condition, learning the specific characteristics of the corresponding antigen, and ending; if not, generating a corresponding initial antibody library, and carrying out the next step;

step S74, selecting N antibodies with highest affinity from the initial antibody library Ab { N }, cloning and mutating the N antibodies to generate a new antibody library Ab { N }, and cloning and inhibiting the initial antibody library Ab { N } through the antibody library Ab { N };

step S75, performing population refreshing on the Ab { N } of the initial antibody library;

and step S76, returning to step S73 to perform iterative operation.

8. The drowning prevention alarm system based on artificial immune algorithm as claimed in claim 7, wherein the single chip microcomputer (4) performs antigen specific feature learning according to the recognized drowning feature and antigen basic feature classification learning, specifically as follows:

step S81, forming an antigen Ag by binary processing of heart rate data, classifying the antigen into an autologous state, namely drowning and a heterologous state, namely normal state;

step S82, generating an initial autoantibody library or an initial variant antibody library Ab { N } through antigen basic feature learning;

step S83, whether the recognition of the antigen and the antibody is finished or not is judged, the recognition is finished, otherwise, the affinity calculation of the antigen and the antibody is carried out, whether the affinity is more than 90% or not is judged, and if the affinity is more than 90%, the antigen Ag is added into an initial antibody library Ab { N }; if the affinity is less than 90%, executing the next step;

step S84, selecting N antibodies with highest affinity from Ab { N }, wherein Ab { N } is the selected antibodies.

Step S85, cloning Ab { n } to prevent the loss of genetic information; then, carrying out mutation on the operator by using a discrete coding algorithm according to the inverse ratio of the mutation rate and the affinity to generate C;

step S86, selecting n antibodies Ab x { n } with highest affinity from C to clone and inhibit Ab { n }, wherein the inhibition affinity is low, and the retention affinity is high; replacing the antibody with low affinity with the randomly generated antibody;

step S87, judging whether the iteration condition meets the condition that the affinity is more than 90%, if so, adding the antigen Ag into the initial antibody library Ab { N }, checking whether the identification of the residual antigen is finished, if so, finishing, and if not, continuing to learn; if not, the process returns to step S84 for iteration.

9. The system of claim 7, wherein the calculation of the affinity between antigen and antibody is performed by calculating the hamming distance between antigen and antibody as follows:

hamming distance between antigen-antibody:

wherein t is_ikIs the K-th position, t, in antigen i_jkIs the K-th position in the antibody j,the operation process is t_iAnd t_jPerforming AND or calculation, and then counting the number of 1 in the obtained binary character;

self-affinity:

wherein t is_WiTime to antigen abnormality of heart rate, t_WjAbnormal heart rate time for antibodies, t_allIs the total time of the antibody;

the affinity of the variant:

wherein t is_NiTime of normal heart rate of antigen, t_NjTime of normal heart rate of antibody, t_allIs the total time of the antibody.

10. The drowning prevention alarm system based on the artificial immune algorithm as claimed in claim 8, wherein the specific steps of the single chip microcomputer (4) for drowning identification are as follows:

step 1, preprocessing the antigen data;

step 2, classifying through basic characteristics of antigens;

step 3, classifying the antigen into self, and performing self library feature recognition algorithm operation, namely performing antigen basic feature learning and antigen specific feature learning on the antigen; the same as the same thing;

and 4, outputting a result after the identification is finished.

Background

In 11/17/2014, the reason why the information published by the world health organization died from drowning is one of ten major causes of death of children and young people, and 37.2 million people die from drowning all the year around the world. Over half of all drowned and died people are less than 25 years old in the world, the drowning rate of children under the years old is the highest, and the drowning rate of men is more than 2 times that of women; over 90% of drowning events occur in low and medium income countries. 2. Drowning accidents and traffic accidents become the biggest death killers of primary and secondary school students (188 deaths and 1266 deaths of students are caused by accidents and events 110 which occur together in primary and secondary schools nationwide from 1 month to 11 months in 2006 according to the report of Chinese education news network in 13 months in 2006, wherein the drowning accidents and the traffic accident deaths respectively account for 43.59 and 30.26 of deaths and live at the first and second places). Even if a rescuer swims in the swimming pool, the probability of drowning is also very high. The research result shows that: the lifeguards of the drowned person can be found to account for only 9% of the total number of the tested lifeguards within 10 seconds, the swim lifeguards of the drowned person can be found to account for 34% of the total number of all the tested lifeguards within 11-30 seconds, the lifeguards of the drowned person can be found to account for 16% of the total number of the tested lifeguards within 31-60 seconds, the lifeguards of the drowned person can be found to account for 18% of the total number of the tested lifeguards within 61-2 minutes, the lifeguards of the drowned person can be found to account for 10% of the total number of the tested lifeguards within 121-3 minutes, and the lifeguards of the drowned person can be found to account for 14% of the total number of the tested lifeguards within 3 minutes. Analysis of the research reports shows that the average time required for a drowner to be found by a swimming rescuer is 1 minute and 14 seconds, which is far more than half of the principle of 10/20 seconds of lifesaving.

At present, the common drowning identification is a sensor system detection, a video detection system and a radar detection system which are identified by three modes. The patent with publication number CN111243237A collects multiframe continuous echo signals of the millimeter wave radar; processing the multiple frames of continuous echo signals to obtain multiple frames of continuous images corresponding to the multiple frames of continuous echo signals; and inputting the multiple frames of continuous images into a preset model to obtain drowning identification results corresponding to each frame of image in the multiple frames of continuous images. Patent publication No. CN110853301A proposes a machine learning-based drowning prevention identification method for swimming pool. Including two methods of surface of water top discernment and surface of water below discernment, through at swimming pool bank and swimming pool wall installation camera, judge whether drowned through handling the video stream, solved the problem of the visual blind area in the swimming pool, the swimming pool safety circumstances can be looked over at any time to the technique with machine study carries out the analysis to the real-time video in the swimming pool, if the drowned circumstances has appeared in the judgement through the algorithm, just can be timely will drown information transmission give the rescuer.

The patent with publication number CN111243237A collects multiframe continuous echo signals of the millimeter wave radar; processing the multiple frames of continuous echo signals to obtain multiple frames of continuous images corresponding to the multiple frames of continuous echo signals; and inputting the multiple frames of continuous images into a preset model to obtain drowning identification results corresponding to each frame of image in the multiple frames of continuous images. However, it is not practical because it is expensive, has high limitation, has a large number of signal interference factors in the swimming pool, and requires a large amount of power to ensure reception of the echo.

Patent publication No. CN110853301A proposes a machine learning-based drowning prevention identification method for swimming pool. Including two methods of surface of water top discernment and surface of water below discernment, through at swimming pool bank and swimming pool wall installation camera, judge whether drowned through handling the video stream, solved the problem of the visual blind area in the swimming pool, the swimming pool safety circumstances can be looked over at any time to the technique with machine study carries out the analysis to the real-time video in the swimming pool, if the drowned circumstances has appeared in the judgement through the algorithm, just can be timely will drown information transmission give the rescuer. Similarly, the system has high cost and strong limitation, is only suitable for places with video monitoring, and is easy to generate misjudgment and missed judgment if the number of the swimmers is too large.

Most of the current sensor detection systems are identified by pressure sensors and heart rate sensors. And monitoring the water entry depth and the heartbeat pulse, and alarming when the heart rate exceeds or is lower than the safe heart rate or the water entry depth is lower than a safe threshold value. The technology can limit the swimming range of the swimmer, is simple to identify corresponding drowning characteristics, and is easy to cause misjudgment and missed judgment.

Disclosure of Invention

The invention aims to provide a drowning prevention alarm system based on an artificial immune algorithm, which can accurately judge the drowning characteristics by learning the heart rate data characteristics based on the artificial immune algorithm. The system is not influenced by geographical factors and the number of swimmers, and has low cost and high practicability.

In order to achieve the purpose, the technical scheme of the invention is as follows: an anti-drowning alarm system based on an artificial immune algorithm is formed by remotely connecting an anti-drowning bracelet (8) and a background server (9) through a 4G module (7); prevent having inlayed in the drowned bracelet singlechip (4) and can carrying out drowned characteristic identification's heart rate measuring module (1) with this singlechip (4) connection, be used for measuring hydraulic pressure measuring module (2), be used for realizing GPS module (3) of location, LED module (5), bee calling organ module (6) and 4G module (7), wherein singlechip (4) can transmit backstage server (9) with data through 4G module (7), carry out alarm processing by backstage server (9).

In one embodiment of the invention, the background server (9) carries out alarm processing through mobile phone APP pushing (10), short message notification (11) and emergency call dialing (12).

In one embodiment of the invention, the single chip microcomputer (4) is AT89C51, the heart rate measuring module (1) is MAX30100, the pressure measuring module (2) is a diffused silicon pressure transmitter, and information acquired by the heart rate measuring module (1), the pressure measuring module (2) and the GPS module (3) is processed and judged by the single chip microcomputer (4) and then transmitted to the background server (9) through the 4G module (7) to give an alarm.

In one embodiment of the invention, when drowning is detected, the GPS module (3) can carry out data acquisition and then carries out drowning characteristic identification by the heart rate measuring module (1), meanwhile, the LED module (5) can flash, and the buzzer can alarm (6).

In an embodiment of the present invention, the system operates as follows:

when a person wearing the drowning prevention bracelet (8) enters water, the pressure measurement module (2) and the heart rate measurement module (1) work simultaneously to perform real-time drowning detection and transmit collected information to the single chip microcomputer (4) to perform drowning identification on a wearer through a drowning identification algorithm, if the wearer drowns, the single chip microcomputer (4) identifies a drowning state, the single chip microcomputer controls the GPS module (3) to collect position information, the single chip microcomputer starts the LED module (5) and the buzzer module (6) to alarm, surrounding pedestrians are noticed, if the drowning state is not relieved after 5 seconds, the single chip microcomputer controls the 4G module (9) to transmit alarm information to the background server (9), and the background server (9) can automatically edit the positioning information into a short message notification (11) to send to an emergency contact person or push the short message (11) through a mobile phone APP (10) after receiving the alarm information, an emergency call (12) is also dialed and location information is also transmitted.

In an embodiment of the invention, the manner of the drowning characteristic identification performed by the heart rate measurement module (1) is as follows: firstly, binary processing is carried out on heart rate data, and the heart rate within 10 seconds is judged to be smaller than 60 or larger than 180 and is recorded as 1, namely, the drowning characteristic; otherwise, a 0 is recorded, i.e. not drowned signature, yielding a 10 second heart rate binary antigen.

In one embodiment of the invention, the singlechip (4) performs antigen basic feature classification learning according to drowning features identified by the heart rate measuring module (1), and the method specifically comprises the following steps:

step S71, an antibody detector generation stage, which respectively learns the characteristics of the autologous and allogeneic antibodies to form corresponding antibody libraries, wherein the definitions of the autologous and allogeneic antibodies are as follows:

autologous characteristics (drowning): the heart rate is less than 60 or the heart rate is more than 180 within 10 seconds, and accounts for 80 percent or more of the total heart rate;

foreign body characteristics (normal): the heart rate is more than 60 and the heart rate is less than 180 within 10 seconds, and accounts for 80 percent or more of the total heart rate;

step S72, randomly selecting and generating a corresponding antibody library after antigen classification is completed according to the autologous characteristics and the allogeneic characteristics;

step S73, calculating the affinity of the antigen and the antibody, and judging whether the affinity meets the condition, namely the affinity is more than 90 percent, generating a corresponding initial antibody library when the affinity meets the condition, learning the specific characteristics of the corresponding antigen, and ending; if not, generating a corresponding initial antibody library, and carrying out the next step;

step S74, selecting N antibodies with highest affinity from the initial antibody library Ab { N }, cloning and mutating the N antibodies to generate a new antibody library Ab { N }, and cloning and inhibiting the initial antibody library Ab { N } through the antibody library Ab { N };

step S75, performing population refreshing on the Ab { N } of the initial antibody library;

and step S76, returning to step S73 to perform iterative operation.

In one embodiment of the invention, the singlechip (4) performs antigen specific feature learning according to the recognized drowning feature and the antigen basic feature by classification learning, and the specific steps are as follows:

step S81, forming an antigen Ag by binary processing of heart rate data, classifying the antigen into an autologous state, namely drowning and a heterologous state, namely normal state;

step S82, generating an initial autoantibody library or an initial variant antibody library Ab { N } through antigen basic feature learning;

step S83, whether the recognition of the antigen and the antibody is finished or not is judged, the recognition is finished, otherwise, the affinity calculation of the antigen and the antibody is carried out, whether the affinity is more than 90% or not is judged, and if the affinity is more than 90%, the antigen Ag is added into an initial antibody library Ab { N }; if the affinity is less than 90%, executing the next step;

step S84, selecting N antibodies with highest affinity from Ab { N }, wherein Ab { N } is the selected antibodies.

Step S85, cloning Ab { n } to prevent the loss of genetic information; then, carrying out mutation on the operator by using a discrete coding algorithm according to the inverse ratio of the mutation rate and the affinity to generate C;

step S86, selecting n antibodies Ab x { n } with highest affinity from C to clone and inhibit Ab { n }, wherein the inhibition affinity is low, and the retention affinity is high; replacing the antibody with low affinity with the randomly generated antibody;

step S87, judging whether the iteration condition meets the condition that the affinity is more than 90%, if so, adding the antigen Ag into the initial antibody library Ab { N }, checking whether the identification of the residual antigen is finished, if so, finishing, and if not, continuing to learn; if not, the process returns to step S84 for iteration.

In one embodiment of the present invention, the calculation of the affinity between antigen and antibody is performed by calculating the hamming distance between antigen and antibody as follows:

hamming distance between antigen-antibody:

wherein t is_ikIs the K-th position, t, in antigen i_jkIs the K-th position in the antibody j,the operation process is t_iAnd t_jPerforming AND or calculation, and then counting the number of 1 in the obtained binary character;

self-affinity:

wherein t is_WiTime to antigen abnormality of heart rate, t_WjAbnormal heart rate time for antibodies, t_allIs the total time of the antibody;

the affinity of the variant:

wherein t is_NiTime of normal heart rate of antigen, t_NjTime of normal heart rate of antibody, t_allIs the total time of the antibody.

In one embodiment of the invention, the specific steps of the single chip microcomputer (4) for drowning identification are as follows:

step 1, preprocessing the antigen data;

step 2, classifying through basic characteristics of antigens;

step 3, classifying the antigen into self, and performing self library feature recognition algorithm operation, namely performing antigen basic feature learning and antigen specific feature learning on the antigen; the same as the same thing;

and 4, outputting a result after the identification is finished.

Compared with the prior art, the invention has the following beneficial effects:

because people basically have mobile phones at present, and the coverage of a 4g base station basically reaches 99%, the 4g module adopted by the invention for data transmission basically does not receive distance limitation, a receiver does not need to be built by the user, an alarm can be given at the first time of drowning, and meanwhile, an emergency contact person can be reminded by sending a short message and the attention of people around can be attracted by a buzzer and an LED lamp. The system device has the advantages of simple circuit structure, high safety, low equipment cost, long endurance time, high automation degree, remote automatic alarm and the like due to the adoption of the lithium battery for power supply.

The algorithm is based on a real-time clone selection algorithm in an artificial immune algorithm, and is mainly realized by maintaining a specific memory antibody set, selecting and cloning the antibody with the highest affinity, deleting the antibody with low affinity, mutating and reselecting a clone population of the antibody with high affinity, and carrying out iterative differentiation. The algorithm is different from the traditional clone selection algorithm, is dynamic in real time, learns while identifying the antigen and the antibody, responds to the dynamic change of the external antigen in real time, and the detector can be adjusted to a certain extent according to different antigen information. And the clone selection algorithm can extract features which are not easy to extract by other algorithms, so that accurate judgment is carried out.

Drawings

FIG. 1 is a block diagram of the system of the present invention.

FIG. 2 is a flowchart of the classification learning of the basic characteristics of the antigen according to the present invention.

FIG. 3 is a flow chart of the learning of antigen specific features of the present invention.

Fig. 4 is a flow chart of drowning recognition according to the present invention.

Detailed Description

The technical scheme of the invention is specifically explained below with reference to the accompanying drawings.

The invention relates to an anti-drowning alarm system based on an artificial immune algorithm, which is formed by remotely connecting an anti-drowning bracelet (8) and a background server (9) through a 4G module (7); prevent having inlayed in the drowned bracelet singlechip (4) and can carrying out drowned characteristic identification's heart rate measuring module (1) with this singlechip (4) connection, be used for measuring hydraulic pressure measuring module (2), be used for realizing GPS module (3) of location, LED module (5), bee calling organ module (6) and 4G module (7), wherein singlechip (4) can transmit backstage server (9) with data through 4G module (7), carry out alarm processing by backstage server (9). The background server (9) carries out alarm processing through mobile phone APP pushing (10), short message notification (11) and emergency call dialing (12). AT89C51 is selected for use in singlechip (4), MAX30100 is selected for use in heart rate measurement module (1), diffusion silicon pressure transmitter is selected for use in pressure measurement module (2), and the information that heart rate measurement module (1), pressure measurement module (2) and GPS module (3) gathered is handled through singlechip (4) and is judged, then transmits backstage server (9) through 4G module (7) and reports to the police.

The following is a specific implementation of the present invention.

As shown in figure 1, the drowning prevention bracelet (8) can be connected with a remote background server (9), can also be bound through a mobile phone APP, and can also be preset with an emergency contact person, when a person wearing the drowning prevention bracelet (8) enters water, the pressure measurement module (2) and the heart rate measurement module (1) work simultaneously to perform real-time drowning detection and transmit collected information to the single chip microcomputer (4) to perform drowning identification on a wearer through a drowning identification algorithm, if the drowning single chip microcomputer (4) of the wearer recognizes the drowning state, the single chip microcomputer controls the GPS module (3) to collect position information, and the single chip microcomputer starts the LED module (5) and the buzzer module (6) to alarm to arouse the attention of surrounding pedestrians, if the drowning state is not relieved after 5 seconds, the single chip microcomputer controls the 4G module (9) to transmit alarm information to the background server, after receiving the alarm information, the background server automatically edits the positioning information into a short message and sends the short message (11) to an emergency contact or a mobile phone APP for pushing. Emergency calls 110 and 120 are also dialed and location information is also sent. The system can effectively prevent the situation that the drowning person can not call for help, can provide safety guarantee for the swimmer by alarming and asking for help in time, and can also provide positioning information under the situation that no one is rescuing around the drowning person to avoid the drowning person from missing rescue time.

In this embodiment, the way that drowning characteristic identification was carried out to heart rate measurement module (1) is: firstly, binary processing is carried out on heart rate data, and the heart rate within 10 seconds is judged to be smaller than 60 or larger than 180 and is recorded as 1, namely, the drowning characteristic; otherwise, a score of 0, i.e., a non-drowning characteristic, yields a 10 second heart rate binary antigen, e.g., 1001101111.

In this embodiment, as shown in fig. 2, the single chip microcomputer (4) performs antigen basic feature classification learning according to the drowning feature identified by the heart rate measurement module (1), specifically as follows:

step S71, an antibody detector generation stage, which respectively learns the characteristics of the autologous and allogeneic antibodies to form corresponding antibody libraries, wherein the definitions of the autologous and allogeneic antibodies are as follows:

autologous characteristics (drowning): heart rate less than 60 or heart rate greater than 180 accounts for 80% or more of the total heart rate within 10 seconds (e.g., 1110101111);

foreign body characteristics (normal): heart rate greater than 60 and heart rate less than 180 accounts for 80% and more of the total heart rate within 10 seconds (e.g., 0000110000);

step S72, randomly selecting and generating a corresponding antibody library after antigen classification is completed according to the autologous characteristics and the allogeneic characteristics;

step S73, calculating the affinity of the antigen and the antibody, and judging whether the affinity meets the condition, namely the affinity is more than 90 percent, generating a corresponding initial antibody library when the affinity meets the condition, learning the specific characteristics of the corresponding antigen, and ending; if not, generating a corresponding initial antibody library, and carrying out the next step;

step S74, selecting N antibodies with highest affinity from the initial antibody library Ab { N }, cloning and mutating the N antibodies to generate a new antibody library Ab { N }, and cloning and inhibiting the initial antibody library Ab { N } through the antibody library Ab { N };

step S75, performing population refreshing on the Ab { N } of the initial antibody library;

and step S76, returning to step S73 to perform iterative operation.

In this embodiment, as shown in fig. 3, the single chip microcomputer (4) performs antigen-specific feature learning by classification learning according to the recognized drowning feature and the antigen basic feature, specifically as follows:

step S81, forming an antigen Ag by binary processing of heart rate data, classifying the antigen into an autologous state, namely drowning and a heterologous state, namely normal state;

step S82, generating an initial autoantibody library or an initial variant antibody library Ab { N } through antigen basic feature learning;

step S83, whether the recognition of the antigen and the antibody is finished or not is judged, the recognition is finished, otherwise, the affinity calculation of the antigen and the antibody is carried out, whether the affinity is more than 90% or not is judged, and if the affinity is more than 90%, the antigen Ag is added into an initial antibody library Ab { N }; if the affinity is less than 90%, executing the next step;

step S84, selecting N antibodies with highest affinity from Ab { N }, wherein Ab { N } is the selected antibodies.

Step S85, cloning Ab { n } to prevent the loss of genetic information; then, carrying out mutation on the operator by using a discrete coding algorithm according to the inverse ratio of the mutation rate and the affinity to generate C;

step S86, selecting n antibodies Ab x { n } with highest affinity from C to clone and inhibit Ab { n }, wherein the inhibition affinity is low, and the retention affinity is high; replacing the antibody with low affinity with the randomly generated antibody;

step S87, judging whether the iteration condition meets the condition that the affinity is more than 90%, if so, adding the antigen Ag into the initial antibody library Ab { N }, checking whether the identification of the residual antigen is finished, if so, finishing, and if not, continuing to learn; if not, the process returns to step S84 for iteration.

In this example, the calculation of the affinity of the antigen and the antibody is performed by calculating the hamming distance between antigen and antibody as follows:

hamming distance between antigen-antibody:

wherein t is_ikIs the K-th position, t, in antigen i_jkIs the K-th position in the antibody j,the operation process is t_iAnd t_jPerforming AND or calculation, and then counting the number of 1 in the obtained binary character;

self-affinity:

wherein t is_WiTime to antigen abnormality of heart rate, t_WjAbnormal heart rate time for antibodies, t_allIs the total time of the antibody;

the affinity of the variant:

wherein t is_NiTime of normal heart rate of antigen, t_NjTime of normal heart rate of antibody, t_allIs the total time of the antibody.

In this embodiment, as shown in fig. 4, the specific steps of the single chip microcomputer (4) for drowning recognition are as follows:

step 1, preprocessing the antigen data;

step 2, classifying through basic characteristics of antigens;

step 3, classifying the antigen into self, and performing self library feature recognition algorithm operation, namely performing antigen basic feature learning and antigen specific feature learning on the antigen; the same as the same thing;

and 4, outputting a result after the identification is finished.

The above are preferred embodiments of the present invention, and all changes made according to the technical scheme of the present invention that produce functional effects do not exceed the scope of the technical scheme of the present invention belong to the protection scope of the present invention.