CN115032904A - Indoor trajectory recording and intelligent analysis system for solitary old people based on fuzzy learning - Google Patents

Abstract

The invention provides a solitary old man indoor track recording and intelligent analysis system based on fuzzy learning, wherein a plurality of infrared sensors are arranged indoors, the infrared sensors are triggered by the motion of solitary old men, the infrared sensors send trigger signals to a main control device, the main control device records the current ID number of the infrared sensor, inquires the installation position bound by a database according to the ID number, stores the trigger time, and counts the trigger signals of the infrared sensors to generate activity condition data of the solitary old man and sends the activity condition data to a designated terminal. Through adopting comparatively cheap infrared sensor to comparatively economic mode has realized reading data to the indoor orbit record of solitary old man, adopts fuzzy algorithm's scheme, can improve information transfer's efficiency by a wide margin, makes the condition that children mastered solitary old man fast, ensures safety. The improved structure of the thermoelectric infrared sensor can avoid the interference of the normal data caused by the mistakenly collected pet action data.

Description

Indoor trajectory recording and intelligent analysis system for solitary old people based on fuzzy learning

Technical Field

The invention relates to the field of information management, in particular to an indoor trajectory recording and intelligent analysis system for solitary old people based on fuzzy learning.

Background

The development bulletin of the national aging cause of the year 2020 issued by the national health committee shows that as soon as 1 day zero of 11 months in 2020, 26402 million people of the aged population in the country 60 years and older account for 18.70% of the total population; 19064 million aged people in 65 years and over in China account for 13.50% of the total population; the nationwide elderly population fostering ratio is 19.70 percent, which is 7.80 percent higher than that in 2010. The aging of the population in China is mainly characterized by large scale of the population of the aged people, obviously accelerated aging process and the like. With the development of modern society, the family structure of people changes, and the active or passive living alone of the old is normalized gradually, so that the social phenomena of the empty-nest old and the solitary old in China are more prominent. The situation that the solitary old people are out of order due to unattended nursing and even are not known by people for many days at home is common. Therefore, how to timely detect abnormal behaviors such as illness and fall in the family environment of the empty nester and inform the family and the hospital at the first time to guarantee the safety of the empty nester is an important subject worth attention.

The life of the old people is generally very regular, but under special conditions such as malaise or illness, the life rules can be changed and abnormal. If the bedroom stays abnormally, the discomfort of the body is prompted; the times of entering and exiting the toilet are obviously increased, and gastrointestinal diseases are prompted; the complete disappearance of the movement track indicates that the solitary old people have serious diseases or even have no fortunate death, and the like. The products related to the old are relatively few in the current market, the product application is very limited, for example, the old can prevent losing a bracelet, although a positioning function is additionally arranged, children are difficult to know the indoor condition of the old practically, part of products adopt a camera device to detect the living condition of the old at any time, the method infringes the privacy of the old to a certain extent, and part of products reflect the living condition of the old by detecting the domestic water condition abroad, but the monitoring information of the method is limited, and the living condition of the old living alone cannot be reflected comprehensively. Although community workers care about solitary old people by installing love doorbells, asking for a visit to the door and the like, the community workers have limited hands and are difficult to know the health conditions of the solitary old people at any time. Patent document CN 113421370 a describes an indoor human body trajectory tracking method, system, and computer terminal, which implement indoor human body tracking through a relatively complex sensor, and have relatively high implementation cost and relatively difficult popularization. Similarly, the technique described in CN 205909831U is also available. Besides higher cost, more sensors are difficult to arrange, and data fusion is also difficult. Therefore, the indoor track recording and analysis of the solitary old people can be realized at lower cost, and the popularization is a technical problem.

Disclosure of Invention

The invention aims to provide an indoor trajectory recording and intelligent analysis system for solitary old people based on fuzzy learning, which can realize indoor trajectory recording and intelligent analysis for solitary old people at lower cost, has clear information, high data reliability, low arrangement cost and convenient popularization.

In order to solve the technical problems, the technical scheme adopted by the invention is as follows: the utility model provides an indoor orbit record of solitary old man and intelligent analysis system based on fuzzy study, be equipped with a plurality of infrared sensor indoor, infrared sensor is triggered by solitary old man's motion, infrared sensor sends triggering signal to master control set, current infrared sensor ID number of master control set record, and the mounted position who binds according to ID number inquiry database, and store and trigger the moment, master control set makes statistics of each infrared sensor's triggering signal, generate solitary old man's activity condition data, and with activity condition data transmission to appointed terminal.

In a preferred scheme, the infrared sensor comprises a correlation type infrared sensor, a reflection type infrared sensor and a thermoelectric type infrared sensor;

the correlation infrared sensors are arranged on two sides of the door;

the reflective infrared sensors are arranged on two sides of a door with a wall on one side;

the thermoelectric infrared sensor is arranged at a position with a larger indoor space.

In the preferred scheme, the correlation type infrared sensor and the reflection type infrared sensor are arranged on the inner side of a room, data of the correlation type infrared sensor and the reflection type infrared sensor are processed according to a switching signal, and whether the switching signal is an incoming signal or an outgoing signal is judged according to the position of a credible solitary old man.

In a preferred embodiment, the previous switching signal is checked for correctness on the basis of the signals received by the further infrared sensors which follow.

In a preferred scheme, the signal of the thermoelectric infrared sensor is a continuous signal, a plurality of signals generated in a time period are mixed with an initial signal by adopting fuzzy calculation, and the last signal is taken as an end signal to generate continuous data corresponding to the current thermoelectric infrared sensor.

In a preferred scheme, signals of the pyroelectric infrared sensor are verified by signals of the infrared sensors at other positions, and if a valid entry and exit record is generated at other positions, a time period corresponding to the entry and exit record is erased from the time period of the pyroelectric infrared sensor, that is, an initial signal and an end signal corresponding to the entry and exit record are generated in continuous data of the pyroelectric infrared sensor, wherein the end signal corresponds to the entry time in the entry and exit record, and the initial signal corresponds to the exit time in the entry and exit record.

In a preferred scheme, the pyroelectric infrared sensor is arranged at a position lower than the highest position of a human body and higher than the highest position of a pet, and a baffle is arranged below the pyroelectric infrared sensor to block infrared rays from the range below the highest position of the pet.

In the preferred scheme, the master control device is provided with a communicator, a storage, an arithmetic unit and a power supply, the communicator collects data of the infrared sensor by using a near field communication protocol, the near field communication protocol comprises a Bluetooth, WIFI, Zigbee or LoRa protocol, and the communicator is connected with the terminal through the Internet;

the memory is used for storing the data of the infrared sensor;

the arithmetic unit is used for calculating the data of the infrared sensor according to a fuzzy algorithm and generating the activity condition data of the elderly living alone.

In the preferred scheme, the activity data of the elderly living alone is chart data, the horizontal axis corresponds to time, the vertical axis corresponds to each room, and the activity track of the elderly living alone is reflected through the initial signal and the ending signal on the chart.

In the preferred scheme, training is carried out on the classifier by using a plurality of normal data groups for a plurality of days, and after the training is finished, a real-time dynamic activity track of the elderly living alone is generated;

and (4) importing the abnormal data group to train the marker, obtaining the state possibly corresponding to the abnormal condition, and prompting the abnormal condition in the activity track of the elderly living alone on the chart.

According to the system for recording the indoor track of the solitary old man and intelligently analyzing the indoor track of the solitary old man based on the fuzzy learning, the indoor track of the solitary old man is recorded in a relatively economic mode by adopting the relatively cheap infrared sensor, the data is read by adopting the scheme of the fuzzy algorithm, the efficiency of information transmission can be greatly improved, children can quickly master the situation of the solitary old man, and the safety is ensured. The improved structure of the thermoelectric infrared sensor can avoid the interference of the normal data caused by the mistakenly collected pet action data.

Drawings

The invention is further illustrated by the following examples in conjunction with the accompanying drawings:

fig. 1 is a schematic view of the indoor arrangement of the infrared sensor of the present invention.

FIG. 2 is a topology diagram of the system of the present invention.

FIG. 3 is a schematic diagram of the signal acquisition of the system of the present invention.

FIG. 4 is a graphical illustration of a normal data set according to the present invention.

FIG. 5 is a graphical illustration of an exception data set in accordance with the present invention.

In the figure: the system comprises an infrared sensor 1, a correlation infrared sensor 100, a reflection infrared sensor 200, a thermoelectric infrared sensor 300, a main control device 400, a restaurant 2, a washroom 3, a master bed 4, a guest bed 5, a living room 6 and a terminal 7.

Detailed Description

Example 1:

as in fig. 1-3, an indoor orbit record of solitary old man and intelligent analysis system based on fuzzy learning, be equipped with a plurality of infrared sensor indoor, infrared sensor 1 is triggered by solitary old man's motion, infrared sensor 1 sends triggering signal to master control set 400, current infrared sensor 1ID number of master control set 400 record, and the mounted position who binds according to ID number inquiry database, and store and trigger the moment, master control set 400 makes statistics of each infrared sensor 1's triggering signal, generate solitary old man's activity condition data, and with activity condition data transmission to appointed terminal 7.

In a preferred scheme, the infrared sensor 1 comprises a correlation type infrared sensor 100, a reflection type infrared sensor 200 and a thermoelectric type infrared sensor 300;

the correlation type infrared sensor 100 is installed at both sides of the door; such as U-LQ30-IN300DP OPP-2021; pu-lq30-in300dp, and the like.

The reflective infrared sensor 200 is installed at both sides of a door having a wall at one side; an advantage of the reflective infrared sensor 200 is that the specular reflector is easier to arrange on a wall. For example: PR18S-DM3DNO, and the like.

During the installation of the correlation infrared sensor 100 and the reflection infrared sensor 200, attention needs to be paid to avoiding interference of the door.

The terminal 7 in this example comprises a mobile phone, a tablet computer, a laptop computer or a desktop computer.

Preferably, the correlation infrared sensor 100 and the reflection infrared sensor 200 are installed at a position higher than the maximum height of the pet to prevent the collection of the pet's entrance and exit data.

In a preferred embodiment, the correlation infrared sensor 100 and the reflection infrared sensor 200 are installed inside a room, and data of the correlation infrared sensor 100 and the reflection infrared sensor 200 are processed according to a switching signal, and the switching signal is determined to be an incoming signal or an outgoing signal according to a position of an authentic elderly living alone, for example, a previously determined position, such as a bedroom.

The pyroelectric infrared sensor 300 is installed in a place where an indoor space is large, such as a living room, a terrace, or a garden. For example: MRT-311MRT-311 type sensors. And the opposite emission type infrared sensor 100 and the reflection type infrared sensor 200 are generally installed in a kitchen, a bedroom or a bathroom.

In a preferred embodiment, the previous switching signal is checked for correctness on the basis of the signals received by the subsequent other infrared sensors 1.

In a preferred embodiment, the signal of the pyroelectric infrared sensor 300 is a continuous signal, a plurality of signals generated in a time period are mixed with the initial signal by using fuzzy calculation, and the last signal is used as an end signal to generate continuous data corresponding to the current pyroelectric infrared sensor 300.

In a preferred embodiment, the signal of the pyroelectric infrared sensor 300 is verified by the signal of the infrared sensor 1 at another position, and if a valid entry/exit record is generated at another position, the time period corresponding to the entry/exit record is erased from the time period of the pyroelectric infrared sensor 300, that is, an initial signal and an end signal corresponding to the entry/exit record are generated in the continuous data of the pyroelectric infrared sensor 300, wherein the end signal corresponds to the entry time in the entry/exit record, and the initial signal corresponds to the exit time in the entry/exit record.

In a preferred embodiment, the pyroelectric infrared sensor 300 is installed at a position lower than the highest position of the human body and higher than the highest position of the pet, and a baffle is arranged below the pyroelectric infrared sensor 300 to block infrared rays from a range below the highest position of the pet. Preferably, the baffle comprises a plate extending horizontally forward for a distance, and a wave absorbing material is disposed on a side of the plate facing the pyroelectric infrared sensor 300 to block infrared waves from below and reflected infrared waves, so as to reduce false triggering.

In a preferred scheme, the main control device 400 is provided with a communicator, a memory, an arithmetic unit and a power supply, the communicator collects data of the infrared sensor 1 by using a near field communication protocol, the near field communication protocol comprises a bluetooth protocol, a WIFI protocol, a Zigbee protocol or a LoRa protocol, and the communicator is connected with the terminal 7 through the internet;

the memory is used for storing the data of the infrared sensor 1;

the arithmetic unit is used for calculating the data of the infrared sensor 1 according to a fuzzy algorithm and generating the activity data of the elderly living alone. The main chip of the master control device 400 preferably adopts an STM32 series chip or an ARM series chip, and the memory preferably adopts an sd card or a solid-state chip. The communicator preferably adopts a communication chip supporting Bluetooth, WIFI, Zigbee or LoRa and 4G or 5G.

In a preferred embodiment, as shown in fig. 4 and 5, the activity data of the elderly living alone is graph data, the horizontal axis corresponds to time, the vertical axis corresponds to each room, and the activity track of the elderly living alone is reflected by the initial signal and the end signal on the graph.

In the preferred scheme, a classifier is trained by using a plurality of normal data groups for a plurality of days, and after the training is finished, the activity track of the elderly living alone dynamically in real time is generated;

and (4) importing the abnormal data group to train the marker, obtaining the possibly corresponding state of the abnormal condition, and prompting the abnormal condition in the activity track of the elderly living alone on the chart.

Example 2:

a solitary old man, male, aged 85 years, living in cell a, cell B unit number 513. Has various chronic diseases and basically cannot go out. The basic life rule data of the old are as follows: 1. a bedroom: 10 pm (± 30 min) to 6 am (± 30 min), 12 pm 30(± 30 min) to 2 pm 30(± 30 min), 24 hour frequency: 2 times; 2. a toilet: duration 5-20 minutes/time, 24 hours frequency: 6-8 times; 3. a restaurant: duration 15-20 minutes/time, 24 hours frequency: 3-5 times; 4. a living room: duration 30-120 minutes/time, 24 hours frequency: 3-5 times.

The first step is as follows: and obtaining a normal activity track chart. After 7-day basic records of the old are recorded by the infrared sensor, a classifier is trained by a 7-day data set, and after the classifier training is finished, the input data set is imaged and is reflected on a chart by the size of x and y values of a two-dimensional data set to form a dynamic real-time track chart, so that a normal activity track chart of the old is obtained, as shown in fig. 3.

Preferably, an abnormal data set is established and imported to train the marker, so as to obtain a state possibly corresponding to an abnormal situation, and the abnormal situation in the activity track of the elderly living alone is prompted on a chart, as shown in fig. 4.

The second step is that: and (5) real-time monitoring and alarming. The system is compared with a chart monitoring range, information is reflected on the terminal displays of the old, the children and the children or community service personnel through an internet system through further microcomputer processing, and the APP early warning system gives different care measures.

The third step: and (5) alarming in abnormal conditions. On a certain day, the system alarms, community staff or old and children look at the terminal display, the system prompts that the old does not leave a bedroom at 9 am, the community staff immediately arrive at a cell A, a cell B, a cell C, the old is asked to generate heat and cough after entering the bedroom, the old has a history of chronic bronchitis, and the community staff immediately dials 120 emergency calls to send the old to a hospital for treatment.

Example 3:

a solitary old man, female, 92 years old, living in cell A, cell B, cell No. 612. Has various chronic diseases and is completely out of the home. The basic life rule data of the old are as follows: 1. a bedroom: 19 pm (± 30 minutes) to 6 am (± 30 minutes), 12 pm (± 30 minutes) to 2 pm (± 30 minutes), 24 hour frequency: 2 times; 2. a toilet: duration 10-30 minutes/time, 24 hours frequency: 6-8 times; 3. a restaurant: duration 25-30 minutes/time, 24 hours frequency: 4-5 times; 4. a living room: duration 60-120 minutes/time, 24 hours frequency: 3-5 times.

The first step is as follows: and obtaining a normal activity track chart. The invention is used for recording the 7-day basic record of the old man, the 7-day data set is used for training the classifier, after the training of the classifier is finished, the input data set is imaged and is reflected on the chart according to the size of the x and y values of the two-dimensional data set to form a dynamic real-time detection chart, and the normal activity track chart of the old man is obtained.

The second step is that: and monitoring and alarming in real time. And comparing the information with a chart monitoring range, reacting the information on the old, the children and the children or on a terminal display of community service personnel (APP early warning system) through an internet system by further microcomputer processing, and giving different care measures.

The third step: and (5) alarming in abnormal conditions. On a certain day, the system alarms, community workers or old and children check the APP early warning system of the terminal display to prompt the old to enter a toilet at 10 o ' clock at night, 20 o ' clock is still in the toilet at 11 o ' clock, the residence time is as long as 80 minutes, the community workers immediately arrive at the No. 612 of the first cell B to check, and the community workers knock for 10 minutes later to hear only indoor weak distress sounds. The old people can not fall down in a toilet after being immediately dialed for 110 alarm and 120 emergency call and checked by a policeman through breaking doors and entering a room, and the old people are immediately sent to a hospital for treatment by community staff.

Example 4:

a solitary old man, male, 98 years old, living in cell a, cell B, unit 406. Has various chronic diseases and has not been out of the home for 3 years. The basic life rule data of the old are as follows: 1. a bedroom: 9 pm 30(± 30 min) to 5 am (± 30 min), 12 pm 30(± 30 min) to 3 pm 30(± 30 min), 24 hour frequency: 2 times; 2. a toilet: duration 10-40 minutes/time, 24 hours frequency: 6-10 times; 3. a restaurant: duration 25-35 minutes/time, 24 hours frequency 7: 3-4 times; 4. a living room: duration 120-: 3-5 times.

The first step is as follows: and obtaining a normal activity track chart. The invention is used for recording the 7-day basic record of the old man, training a classifier by using a 7-day data set, and after the training of the classifier is finished, imaging the input data set, and reacting the input data set on a chart by using the x and y values of the two-dimensional data set to form a dynamic real-time detection chart so as to obtain a normal activity track chart of the old man.

The second step is that: and monitoring and alarming in real time. And comparing the information with a chart monitoring range, reacting the information on the old, children and women or on a terminal display of community service personnel through an internet system by further processing by a microcomputer, and giving different care measures.

The third step: and (5) alarming in abnormal conditions. On a certain day, the system alarms, community workers or old and children look up the terminal display APP early warning system, namely the system prompts that the old does not leave a bedroom at 9 o' clock 30 a day earlier, the community workers immediately arrive at the unit B406 of the cell A, and the user can knock the door for 20 minutes without any reaction indoors. The alarm 110 and the emergency call 120 are immediately dialed, and people and medical staff break the door and enter the room to check the alarm, so that the old people are found to be still on the bedroom bed but have no vital signs.

Example 5:

in a preferred scheme, logic for judging whether the activity track of the solitary old people is normal based on fuzzy learning does not have an accurate boundary between true and false, the transition of the fuzzy logic from true to false is gradual, and the process is described by a membership function.

In a preferred scheme, the basic idea of the core idea fuzzy theory for data processing is as follows: the degree of belongings is used instead of belongings or not. Let X be the possible number of times of entrance and exit or single stay time in a certain room in one day, and A represents the preferred scheme, training the classifier by using the normal data group of a plurality of days, and generating the real-time dynamic activity track of the elderly living alone and the membership function mu after the training is finished _A Is a function of mapping any element X in X to A, i.e.

μ _A ：X→[0，1]

x→μ _A (x)

μ _A (x) The degree of membership of x to the activity track of the aged is normal. Mu.s _A (x) The larger the value, the higher the degree to which x is subordinate to A. Membership μ of all X in X _A (x) The formed set is called a fuzzy set on X, and the expression method comprises the following steps: a

1. Zadeh notation

(i) The universe of discourse is discrete and limited in the number of elements:

or

A＝{μ _A (x ₁ )/x ₁ ，μ _A (x ₂ )/x ₂ ，…，μ _A (x _n )/x _n }

(ii) The universe of discourse is continuous, or infinite in the number of elements:

note that: the "/" herein does not represent a division number, but is used to indicate membership of data within a data set.

2. Sequence representation method

A＝{(μ _A (x ₁ )，x ₁ )，(μ _A (x ₂ )，x ₂ )…，(μ _A (x _n )，x _n )}

3. Vector representation

A＝{μ _A (x ₁ )，μ _A (x ₂ )，…，μ _A (x _n )}

Note that: when the vector is used for representation, the item with the membership function equal to 0 cannot be omitted; the entry of 0 in the Zadeh and ordinal representations may be omitted.

In a preferred scheme, the membership degrees of all elements in the fuzzy set collectively form a membership function of the fuzzy set, and the membership function is a quantitative description of a fuzzy concept. The membership function is determined according to experience or statistics, the membership function established in the invention is normal distribution, and the specific function construction mode is as follows:

in a preferred scheme, a classifier is trained by using normal data groups for multiple days, after the training is finished, a real-time dynamic activity track of the elderly living alone is generated, and the time range [ a, b ] (h) and the frequency range [ m, n ] (times) of single activity of each ID number are counted for chart data formed by activity track information.

Taking time as a domain of discourse, taking U as [ a,24] (unit is h) (note that the time is less than a, the domain of discourse is not included), and giving membership functions of 'normal activity track' T and 'abnormal activity track' F:

taking the number of times as a domain, taking U as [0,20] (the unit is the number of times), and giving membership functions of 'activity normal' T and 'activity track abnormity' F:

(Note: where the parameters λ and γ represent the sensitivity of the membership function to time changes, and need to be adjusted according to the ordinate ID number; u is the time counted from when the IR trigger 1 is triggered.)

In the preferred scheme, a fuzzy set aiming at the normal activity track of the old people can be obtained:

u is time: a. the _T1 ＝{μ _A (x ₁ )/x ₁ ，μ _A (x ₂ )/x ₂ ，…，μ _A (x _n )/x _n }

U is the number of times: a. the _T2 ＝{μ _A (x ₁ )/x ₁ ，μ _A (x ₂ )/x ₂ ，…，μ _A (x _n )/x _n }

And aiming at the fuzzy set of the activity track abnormity of the old:

u is time: a. the _F1 ＝{μ _A (x ₁ )/x ₁ ，μ _A (x ₂ )/x ₂ ，…，μ _A (x _n )/x _n }

U is the number of times: a. the _F2 ＝{μ _A (x ₁ )/x ₁ ，μ _A (x ₂ )/x ₂ ，…，μ _A (x _n )/x _n }

Importing abnormal data group to train the marker, obtaining the possible corresponding state of the abnormal condition, comparing different abnormal conditions with A _F1 And A _F2 And matching the membership degrees, and obtaining the most possible corresponding abnormal condition of each membership degree according to the Bayesian network.

In a preferred scheme, when the infrared sensor 1 is triggered, the complete number n of times that the infrared sensor 1 with the ID number is triggered in the current day is read first, and the number n +1 for A is obtained _T1 If the membership degree of (A) is 1, the feedback is that the activity track is normal, and if the membership degree of (A) is less than 1, the activity track is normal _F1 The membership degree of the old people is matched with the most probable abnormal situation and prompts the abnormal situation in the activity track of the old people living alone on the chart. At the same time, the timing is started after the entering signal is triggered, if the total time length t is corresponding to A when the ending signal is received _T2 Is 1, the feedback is that the active track is normal, if the time increases to t, the end signal is still not received, and t is A _T2 If the membership degree of (b) is less than 1, according to t updated in real time and the sum of t and A _T2 The membership degree of the old people is matched with the most probable abnormal situation, and the abnormal situation in the activity track of the old people living alone is marked on the chart for prompting.

The above-described embodiments are merely preferred embodiments of the present invention, and should not be construed as limiting the present invention, and the scope of the present invention is defined by the claims, and equivalents including technical features described in the claims. I.e., equivalent alterations and modifications within the scope hereof, are also intended to be within the scope of the invention.

Claims (10)

1. The utility model provides an indoor orbit of solitary old man is taken notes and intelligent analysis system based on fuzzy learning, characterized by: the infrared sensors (1) are triggered by the movement of the solitary old people, the infrared sensors (1) send trigger signals to the main control device (400), the main control device (400) records the current ID number of the infrared sensors (1), inquires the installation position bound by the database according to the ID number and stores the trigger time; the main control device (400) counts the trigger signals of the infrared sensors (1), generates activity condition data of the elderly living alone, and sends the activity condition data to the designated terminal (7).

2. The system of claim 1 for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning, which is characterized in that: the infrared sensor (1) comprises a correlation type infrared sensor (100), a reflection type infrared sensor (200) and a thermoelectric type infrared sensor (300);

the correlation type infrared sensors (100) are arranged on two sides of the door;

the reflective infrared sensor (200) is installed on both sides of a door with a wall on one side;

the pyroelectric infrared sensor (300) is installed at a position where the indoor space is large.

3. The system for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning according to claim 2, wherein: the correlation type infrared sensor (100) and the reflection type infrared sensor (200) are installed on the inner side of a room, data of the correlation type infrared sensor (100) and the reflection type infrared sensor (200) are processed according to a switching signal, and whether the switching signal is an incoming signal or an outgoing signal is judged according to the position of a credible solitary old man.

4. The system of claim 3 for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning, which is characterized in that: and checking whether the previous switching signal is correct according to the signals received by the other subsequent infrared sensors (1).

5. The system of claim 2 for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning, which is characterized in that: the signal of the thermoelectric infrared sensor (300) is a continuous signal, a plurality of signals generated in a time period are mixed with the initial signal by adopting fuzzy calculation, the last signal is used as an end signal, and continuous data corresponding to the current thermoelectric infrared sensor (300) are generated.

6. The system of claim 5 for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning, which is characterized in that: checking the signal of the pyroelectric infrared sensor (300) by using the signal of the infrared sensor (1) at other positions, and if a valid access record is generated at other positions, erasing the time period corresponding to the access record from the time period of the pyroelectric infrared sensor (300), namely generating an initial signal and an end signal corresponding to the access record in the continuous data of the pyroelectric infrared sensor (300), wherein the end signal corresponds to the entry time in the access record, and the initial signal corresponds to the exit time in the access record.

7. The system of claim 5 for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning, which is characterized in that: the pyroelectric infrared sensor (300) is arranged at a position which is lower than the highest position of a human body and higher than the highest position of a pet, and a baffle plate is arranged below the pyroelectric infrared sensor (300) to block infrared rays from a range below the highest position of the pet.

8. The system for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning as claimed in any one of claims 1 to 7, wherein: the main control device (400) is provided with a communicator, a storage, an arithmetic unit and a power supply, the communicator collects data of the infrared sensor (1) by a near-field communication protocol, the near-field communication protocol comprises a Bluetooth, WIFI, Zigbee or LoRa protocol, and the communicator is connected with the terminal (7) through the Internet;

the memory is used for storing data of the infrared sensor (1);

the arithmetic unit is used for calculating the data of the infrared sensor (1) according to a fuzzy algorithm and generating the activity data of the elderly living alone.

9. The system of claim 8, which is used for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning, and is characterized in that: the activity data of the elderly living alone are chart data, the horizontal axis corresponds to time, the vertical axis corresponds to each room, and the activity track of the elderly living alone is reflected through an initial signal and an ending signal on the chart.

10. The system for indoor trajectory recording and intelligent analysis of solitary old people based on fuzzy learning according to claim 9, wherein: training the classifier by using a plurality of normal data groups for a plurality of days, and generating real-time dynamic activity tracks of the elderly living alone after training is completed;

and (4) importing the abnormal data group to train the marker, obtaining the state possibly corresponding to the abnormal condition, and prompting the abnormal condition in the activity track of the elderly living alone on the chart.

Images