Passive tracking and positioning method

Abstract

The invention discloses a passive tracking and positioning method, which comprises the following steps of: carrying or wearing an electronic tag by a user, reading the electronic tag by using reading equipment, obtaining ID information in the electronic tag, and transmitting the time for reading the electronic tag, the ID information and the duration for transmitting a signal of the electronic tag to a reader to a server through a network; and the server matches the corresponding user by using the ID information, calculates the distance between the electronic tag and the reading equipment or the position of the electronic tag relative to the reading equipment by reading the position of the reader and the duration of transmitting the signal of the electronic tag to the reader, and records the distance and the position. Subsequently, road stiffness information is obtained by reading the time of the electronic tag and corresponding position information, path tracking is completed, and the path and the position of the user can be calculated in combination with the map, the building structure chart and the position of the reading device on the map and the building structure chart, the moving speed of the user, the historical position and the path of the user and the like.

Description

Technical Field

[0001] The invention relates to positioning and location tracking technology, and in particular to a passive tracking and positioning method. Background Art

[0002] Active positioning technology is highly accurate and fast. Most of them use internal devices to actively transmit or receive signals to obtain positioning, such as GPS positioning, Beidou positioning, etc. Passive positioning and tracking technology is mainly used in the Internet of Things, especially in warehousing, logistics systems, and urban asset management systems. It mainly uses electronic tags to fix on the target object, and then reads the electronic tag information through the corresponding reader to identify the target object and locates the target object in combination with the corresponding positioning technology (such as RFID positioning technology). This type of technology has high positioning accuracy, but has high requirements for the system or equipment, and is currently not suitable for outdoor positioning.

[0003] Active positioning technology requires power support, so it needs to be charged, and the structure of the wearable or mobile terminal is relatively complex, so it is not suitable for scenarios where there is no power supply for a long time and it needs to be worn for a long time. With the development of Internet of Things technology, especially the popularization of 5G technology, as well as the construction of various outdoor passive base stations and RFID satellite systems (such as ocean shipping, polar exploration, wildlife protection, etc.), passive positioning technology also has room for realization. At the same time, passive positioning technology is manifested as a thin electronic tag on the wearable or mobile terminal, and does not require power and has a simple structure, which has advantages that active positioning cannot match.

[0004] However, there is currently no passive positioning technology suitable for outdoor use, which has become a technical problem that needs to be solved. Summary of the invention

[0005] In view of the above-mentioned defects in the prior art, the technical problem to be solved by the present invention is to provide a passive tracking and positioning method, which can realize passive positioning and path tracking.

[0006] To achieve the above object, the present invention provides a passive tracking and positioning method, including an electronic tag carried or worn by a user, the electronic tag has the user's ID information built in, and a subsequent system associates the user with the ID information as the user's identity identification mark;

[0007] The reading device reads the information of the electronic tag and uploads the read information to the server or designated device for storage through the network;

[0008] The database stores maps or building structure diagrams in advance, and the electronic tags can be located by combining the position of the reading device in the stored maps or buildings. The steps include:

[0009] S100, obtaining the user's walking speed V, the number of times the user passes N1, the number of people passing the path and position N2, the path relationship R between possible paths, the user's travel purpose P, and the possible path length L;

[0010] S200, parameter weight setting:

[0011] S210, weight (α) of walking speed (V);

[0012] S220, weight (β1) of the number of user passes (N1);

[0013] S230, weight (β2) of the number of people (N2) passing through the path and location;

[0014] S240, weight (γ) of path relationship (R);

[0015] S250, weight (δ) of the travel purpose (P);

[0016] S260, weight (ε) of path length (L);

[0017] S300, position estimation, the walking speed, the number of user passes, the number of people on the path, the path relationship, the purpose of the pass, and the path length are scored by the corresponding scoring functions. The scoring functions of the walking speed, the number of user passes, the number of people on the path, the path relationship, the purpose of the pass, and the path length are respectively denoted as V Score(pj) 、N 1Score(pj) 、N 2Score(pj) , R Score(pj) , P Score(pj) , L Score(pj) ; Calculate the likelihood score S of each location point based on the relationship between each parameter and weight core (p j ):

[0018] S core (p j )=α*V Score(pj) +β1*N 1Score(pj) +β2*N 2Score(pj) +γ*R Score(pj) +δ*P Score(pj) +ε*L Score(pj)

[0019] The highest scoring locations are selected as the final possible user location set, and the user's path is inferred based on the predicted location, the last located location, the walking speed, and the navigable path.

[0020] The beneficial effects of the present invention are:

[0021] The present invention uses a user to carry or wear an electronic tag, uses a reading device to read the electronic tag and obtain the ID information in the electronic tag, and then transmits the time of reading the electronic tag, the ID information, and the duration of the electronic tag signal transmission to the reader to the server through the network. The server uses the ID information to match the corresponding user, and uses the position of the reader and the duration of the electronic tag signal transmission to the reader to calculate the distance between the electronic tag and the reading device or the position of the electronic tag relative to the reading device, and records it. Subsequently, the road information is obtained by reading the time of the electronic tag and the corresponding position information to complete the path tracking, and the user's path and position can be calculated in combination with the map, the building structure diagram, and the position of the reading device on the map and the building structure diagram, combined with the user's moving speed, the user's historical position and path, etc. At the same time, it can be combined with other positioning devices to achieve more accurate positioning and path calculation to meet different positioning requirements. The present invention is particularly suitable for passive positioning monitoring of special groups, such as children, the elderly, and people with mental disorders.

[0022] At present, the reading distance of passive electronic tags is not as good as that of active electronic tags, but the coverage range of general 5G-Α base stations is about 300 meters, so in cities with dense 5G-Α base stations, the present invention can achieve approximate positioning. At the same time, there are a large number of 5G-Α base stations and reading devices in environments such as subways, shopping malls, and supermarkets, so high-precision positioning can be achieved indoors. In the wild, the adaptability of the present invention will be much worse, because it mainly relies on reading equipment to read information and then upload it, but with the increasing development of communication technology and the increasing maturity of Internet of Things technology, it is only a matter of time before this problem is solved. Of course, in some scenarios, positioning can be achieved by setting up a large number of reading devices outdoors, such as outdoor personnel location monitoring systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a schematic block diagram of the system structure of the present invention;

[0024] Figure 2 is a schematic diagram of the structure of the electronic tag 100;

[0025] Figure 3 The improved structure of the electronic tag 100 is shown in FIG. Figure 1 ;

[0026] Figure 4 The improved structure of the electronic tag 100 is shown in FIG. Figure 2 ;

[0027] Figure 5 This is the installation diagram of the electronic tag 100 Figure 1 ;

[0028] Figure 6 This is a schematic diagram of the anti-tampering installation structure of the electronic tag 100 Figure 1 ;

[0029] Figure 7 This is a schematic diagram of the anti-tampering installation structure of the electronic tag 100 Figure 2 ;

[0030] Figure 8 is a schematic diagram of the structure of the electronic tag 100 after adding a heat insulation film 150;

[0031] Fig. 9 is a schematic diagram of the structure of the electronic tag 100 after the adhesive layer 160 is added;

[0032] Fig.10 It is a schematic diagram of the structure of the inner electronic tag;

[0033] Fig.11 It is a schematic diagram of the structure of the outer electronic tag;

[0034] Fig.12 It is a schematic diagram of the inner electronic tag and the outer electronic tag cooperating to detect body temperature, sweating, and water ingress;

[0035] Fig.13 It is a schematic block diagram of the operation process of the passive tracking and positioning method. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0037] See also Figure 1 , Figure 2 The passive tracking and positioning system of this embodiment includes:

[0038] The electronic tag is carried or worn by the user and has the user's ID information built in. The subsequent system associates the user with the ID information as the user's identity identification mark. In this embodiment, the electronic tag can be packaged into a film or other portable form that is easy to fix on the user's body or clothing, so that the user can fix the electronic tag on clothes, pants, shoes, etc. This is obviously lighter and more portable than the traditional active positioning method, and will hardly add burden and discomfort to the user. In this embodiment, the electronic tag 100 can be fixed on the user's clothes, pants, shoes and other wearable items by sewing, pasting, etc., or even directly pasted on the user's body.

[0039] Reading device, used to read the information of the electronic tag and upload the read information to the server or designated device through the network for storage, so as to facilitate subsequent call and access;

[0040] Network, used to connect reading devices, servers, management terminals and other devices to the network;

[0041] The server is used to store the information uploaded by the reading device, receive instructions and access requests from the management end, store the database, and perform subsequent location and path calculations;

[0042] The management terminal is used to access the server through the network to obtain the data of the server or send instructions to the server. The management terminal of this embodiment can be a PC terminal, a mobile phone terminal, etc. The PC terminal can be a web page terminal or a specially designed software, and the mobile phone terminal is mostly manifested as an API, a mini program, etc.

[0043] The reading device is a device that can read the electronic tag information, including an electronic tag reader (such as an RFID reader), a base station for reading passive tags, a satellite system for reading passive tags, etc. In this embodiment, there are many devices that can read electronic tags, such as gates and gates for reading electronic tags in shopping malls and supermarkets, subway gates, 5G-Α base stations, RFID base stations, ETC reading devices, vehicle annual inspection electronic tag reading devices, etc. The satellite system is a system in which multiple base stations communicate with satellites, and satellites communicate with servers. This system has been used in passive positioning solutions such as ocean shipping, polar exploration, and wildlife protection.

[0044] The reading device is fixed, so once the reading device reads the electronic tag information, it means that the electronic tag is around it. At the same time, the distance between the electronic tag and the reading device can be calculated by using the signal transmission time between the electronic tag and the reading device. At this time, it can be determined that the user is within a certain distance around the reading device. When there are two or more reading devices, the triangulation positioning method and RFID positioning method can be used to further locate the electronic tag. At this time, the specific location of the electronic tag and the reading device can be calculated using the position of the reading device and the distance between the electronic tag. Of course, these can be calculated by the software built into the server, and you only need to output the results and record them later. When the reading device reads the electronic tag signal, it also records the reading time, that is, the time when the electronic tag signal is obtained. In a certain period of time, multiple readers read the information of the electronic tag, and the server sorts it out in the order of reading the information to obtain the path information to achieve path tracking.

[0045] Maps or building structure diagrams can be stored in advance in the database, and the position of the reading device in the stored map or building is fixed. Therefore, once the reading device reads the tag, the position range or position of the electronic tag in the map or building can be roughly determined. At the same time, the server stores the position and path of each positioning to form a historical path or position, so as to facilitate the subsequent position and path calculation.

[0046] In the process of calculating the location and path, we first combine the map or building structure diagram to determine the routes that can be passed near the reading device, and then combine the historical path and the number of times each route is passed at the location. The higher the number, the greater the probability of appearing at the next location. At the same time, combined with the reading time, the time difference when calculating the location and the moving speed, we can further calculate the approximate location on each path; if there is a reading device at the calculated location, but this reading device does not read the signal of the electronic tag, it means that the probability of the user passing through this route is low. In this way, we can finally get a rough idea of ​​the user's walking path and location.

[0047] Of course, further positioning can be achieved by combining with other positioning devices, such as GPS devices and Beidou devices bound to user IDs, and using the positioning information of other positioning devices to further calculate the user's movement path and location. This method can achieve high-precision indoor navigation and positioning.

[0048] In some embodiments, it is possible to consider associating the ID of the electronic tag with the user's bus account and e-wallet account, so that after the reader reads the ID information of the electronic tag, it can be matched with the user's bus account and e-wallet account, and the fee can be deducted according to the authorization. This design is similar to the current technology of opening Alipay to deduct fees after reading the NFC of the mobile phone through the reader, and this technology can even be directly adopted. On the one hand, this design greatly facilitates the use of users, especially for some elderly people, who can deduct fees by directly reading the electronic tag, which is very convenient to use; on the other hand, uploading reading information through the reader can assist in locating the user, especially for public transportation, and the user's location can be relatively accurately calculated through the reader's reading route, running speed, etc. Of course, the ID of the electronic tag can also be associated with the access control and attendance equipment, so that such equipment can be used to read the tag to complete positioning and information upload.

[0049] See also Figure 2 , which is a structural diagram of one type of electronic tag, including a base layer 110, a protective layer 120, a functional part 200 is arranged or fixed on the base layer 110, and the base layer 110 located outside the functional part 200 is respectively fixed and sealed with the protective layer 120, so as to achieve waterproof and dustproof. The base layer 110 and the protective layer 120 are both made of flexible materials.

[0050] The functional part 200 is the functional part of the electronic tag, and generally includes a chip, a coupling element, an antenna, etc., as long as it can realize the function of the electronic tag.

[0051] The antenna and circuit of this embodiment may be printed on the base layer 110 or the protective layer 120 by conductive silver paste, conductive ink, etc., and then the chip mounted on the base layer 110 or the protective layer 120. This is actually the existing printed electronic label.

[0052] The functional part 200 of this embodiment can also be an independently manufactured component, which is then encapsulated in the base layer 110 and the protective layer 120. This embodiment is preferably this way, because the printing method is not resistant to bending and has slightly poor performance. Specifically, the functional part 200 can use a conductive material that is resistant to bending and low temperature to make antennas and circuits, and then install and fix the chip. In this embodiment, nano silver wire conductive film and graphene-based conductive film can be selected to make antennas, circuits, etc. of the functional part 200, such as punching or cutting antennas and circuits on the corresponding film, and then installing and fixing components such as chips on the film to obtain the functional part 200, and then encapsulate the functional part 200 in the base layer 110 and the protective layer 120. Nano silver wire conductive film and graphene-based conductive film have excellent conductivity and mechanical strength, and can maintain stable conductive properties in low temperature environments (minus 20 degrees Celsius and below), so they can meet the use environment of high-frequency bending and low temperature. In this case, the functional part 200 may be firstly bonded and fixed to the base layer 110 or the protective layer 120 , and then the base layer 110 and the protective layer 120 may be sealed and fixed.

[0053] The base layer 110 and the protective layer 120 can also be made of bending-resistant and low-temperature-resistant materials, such as polyimide film, cold-resistant PVC film, etc. At the same time, the base layer 110 and the protective layer 120 can be sealed and fixed by hot pressing, adhesive sealing, ultrasonic welding, etc.

[0054] Polyimide film has good adsorption properties, so the printing stability is very high. Therefore, in this embodiment, when the base layer 110 and / or the protective layer 120 are made of polyimide film, the circuit or antenna can be directly printed on the base layer 110 and / or the protective layer 120.

[0055] Preferably, if the base layer 110 is made of a material that is not easily adsorbed by the printed circuit, a layer of polyimide can be evenly coated on the base layer 110 and then the circuit can be printed. In this way, a better adsorption effect can be achieved and the printing of the printed circuit can be facilitated.

[0056] Preferably, in order to increase the stability and life of the electronic tag, in some embodiments, a composite resin mortar coating, a polyurethane coating, or an acrylic waterproof coating may be sprayed on the base layer 110 and / or the protective layer 120. These materials have waterproof and wear-resistant properties, thereby effectively improving the overall waterproof and wear-resistant properties of the electronic tag to increase the life of the electronic tag.

[0057] Preferably, a water-chromic material can be applied on the surface of the functional part 200, or / and the corresponding part of the base layer 110, the protective layer 120 and the functional part 200, or / and the inner side of the fixed connection between the base layer 110 and the protective layer 120. Once water enters between the base layer 110 and the protective layer 120, the water-chromic material will change color, thereby prompting the user to replace it to avoid affecting the use. The water-chromic material can be a water-chromic pigment, a water-chromic coating, a water-chromic ink, etc.

[0058] Preferably, in some embodiments, the base layer 110 and the protective layer 120 may be fixed by a connecting layer 610 , and the connecting layer 610 may be glue, that is, the base layer 110 and the protective layer 120 are bonded and fixed as a whole by glue.

[0059] Preferably, in some embodiments, the base layer 110 and the protective layer 120 may be integrally cast, so that the base layer 110 and the protective layer 120 are naturally fused and sealed as one.

[0060] See also Figure 3 In some embodiments, the outer end surface of the base layer 110 or the protective layer 120 may be covered with a layer of adhesive to form an adhesive layer 310, and the electronic tag 100 is attached and fixed to a position required by the user through the adhesive layer 310. The adhesive may be a self-adhesive adhesive.

[0061] Preferably, the end surface of the adhesive layer 310 away from the electronic tag 100 is covered with a release layer 640, which is used to protect the adhesive layer 310 and prevent the adhesive layer 310 from being unnecessarily bonded to other objects, which affects the use. The release layer 640 can be a thin film, and the specific reference can be made to the paper or film used to isolate the adhesive part on the outside of the double-sided tape in the prior art. When in use, tear off the release layer 640 and stick the adhesive layer 310 to the desired position.

[0062] Preferably, a printing layer 620 may be attached to the end surface of the electronic tag 100 away from the adhesive layer 310, and patterns and texts may be printed on the printing layer 620 for decoration or printing the ID information of the electronic tag, associated QR code, etc. Of course, some reflective patterns may be printed or the surface of the printing layer 620 may be printed with reflective materials to achieve the reflective function. Of course, the printing layer 620 may also be omitted and the corresponding patterns, texts, reflective materials, etc. may be directly printed or coated on the electronic tag 100.

[0063] See also Figure 4 In some embodiments, the electronic tag 100 can be fixed on the inner fixing layer 320, and then fixed on other objects such as clothes, shoes, hats, bags, etc. using the inner fixing layer 320.

[0064] Specifically, an adhesive layer may also be provided on the inner fixing layer 320 , so that the adhesive layer 310 can be used to adhere and fix it on other objects.

[0065] Or the inner fixing layer 320 is sewn on the fabric by needle and thread, such as being sewn and fixed on clothes, shoes, hats, bags, etc.

[0066] Or the inner fixing layer 320 and the fixing object 01 are provided with a hook surface and a fur surface of the tear buckle respectively, and then the hook surface and the fur surface are adhered and fixed to each other. The fixing object 01 can be clothes, shoes, hats, bags, user's skin, user's mobile phone and the like.

[0067] Or the inner fixing layer 320 is directly fixed on the fixed object 01 by glue. In short, the electronic tag 100 can be fixed on the fixed object 01 by the inner fixing layer 320 .

[0068] See also Figure 5 In some embodiments, an outer fixing layer 330 may be added, and the outer fixing layer 330 presses the electronic tag 100 onto the fixed object 01, and the outer fixing edge 331 of the outer fixing layer 330 is fixed to the fixed object 01. The outer fixing edge 331 may be fixed to the fixed object by gluing, adhesive bonding, tear-off buckle bonding, sewing, zipper connection, etc. In theory, as long as the outer fixing edge 331 is fixed or bonded to the fixed object 01, it will be sufficient.

[0069] Preferably, the electronic tag 100 can be adhered or fixed to the outer fixing layer 330 so that the electronic tag 100 and the outer fixing layer 330 become one.

[0070] Preferably, a printed layer 620 may be adhered, fixed or provided on the end surface of the outer fixed layer 330 away from the electronic tag 100 .

[0071] See also Figure 6 In order to prevent the electronic tag 100 from being torn off manually and ensure easy disassembly and assembly, a buckle seat 430 is fixed on the wearable object 410 in this embodiment, and a spring piece 431 and a buckle piece 451 are provided on the buckle seat 430. The spring piece 431 and the buckle piece 451 are elastic, and a magnetic block 440 is installed on the spring piece 431, which can be adsorbed by a magnet.

[0072] The buckle head 450 is mounted on the buckle piece 451, and the buckle head 450 is respectively provided with an outer bevel 452 and an inner bevel 453, and the inner bevel 453 is fitted on the switch bevel 441 of the magnetic block 440, so as to keep the buckle head 450 inserted into the buckle groove 131, and the buckle groove 131 is provided on the buckle sleeve 130, and the inner side of the end of the buckle sleeve 130 is provided with a chamfered bevel 132, and the inside of the buckle sleeve 130 is a hollow buckle cavity 133. The buckle cavity 133 is sleeved outside the buckle seat 430. The buckle sleeve 130 is fixed on the electronic tag 100.

[0073] The spring piece 431 exerts an elastic force on the magnetic attraction block 440 to push the inward inclined surface 453 , and the buckle piece 451 exerts an elastic force on the buckle head 450 to move away from the buckle groove 131 . Figure 6 The electronic tag is shown in its installed state. When the electronic tag needs to be removed, a magnet is placed on one side of the wearable object 410. The magnet attracts the magnetic block 440 to the wearable object 410 through magnetic force. The magnetic block 440 overcomes the elastic force of the spring 431 and moves away from the electronic tag 100, thereby separating the switch inclined surface 441 from the inner inclined surface 453. The buckle piece 451 carries the buckle head 450 and rotates away from the buckle slot 131 until it exits the buckle slot 131. At this time, the electronic tag 100 can be pulled out.

[0074] When the electronic tag needs to be installed, the buckle sleeve 130 is put on the buckle head 450 and then pressed against the wearable object 410, so that the chamfered inclined surface 132 and the outer inclined surface 452 are pressed together to apply a squeezing force to the spring piece on the magnetic block 440, and the magnetic block 440 moves downward to make the buckle sleeve 130 fit outside the buckle seat 430, and when the buckle slot 131 and the buckle head 450 are facing each other, the magnetic block 440 is reset to push the buckle head into the buckle slot 131 to complete the installation. This design can only remove the electronic tag by magnetic attraction, so it can effectively prevent it from being torn off by the user or others to ensure effective positioning.

[0075] See also Figure 7 , and Figure 6 The same is for anti-disassembly, but the structure has been simplified, as follows:

[0076] The electronic tag 100 is fixed with a needle seat 710, a plug pin 720 is fixed on the needle seat 710, and an arc-shaped annular groove 721 is provided on the plug pin 720;

[0077] The pin 720 passes through the fixture 01 and is inserted into the fixing box 730. The fixing box 730 has a hollow fixing box cavity 731. A locking member 740 is installed in the fixing box cavity 731. The locking member 740 is provided with a buckle strip 741, a diagonal brace strip 742, and a supporting spring piece 743. The buckle strip 741 is provided with a buckle protrusion 7411. One end of the diagonal brace strip 742 is connected to the buckle protrusion 7411 or the buckle strip 741. The other end of the diagonal brace strip 742 is fixed to the magnet block 750. The magnet block 750 is placed on the supporting spring piece 743. The supporting spring piece 743 and the buckle strip 741 are both elastic. The magnet block 750 is made of a material that can be adsorbed by a magnet, such as soft iron, steel, magnet, etc.

[0078] The buckle protrusion 7411 is engaged in the arc-shaped annular groove 721 to lock the pin 720 in the axial direction. At this time, the pin 720 cannot be pulled out of the fixing box 730. When unlocking is required, the outer side of the end of the locking member 740 ( Figure 7 A magnet is placed on the bottom of the plug or a magnetic field is applied. At this time, the magnetic field pulls the magnet block 750 downward through magnetic force, thereby driving the buckle strip 741 to bend, and the buckle strip 741 drives the buckle protrusion 7411 to exit the arc-shaped ring groove 721, and the pin 720 can be pulled out at this time. When the pin is inserted, the pin squeezes the buckle protrusion 7411 to push the buckle strip 741 to bend outward until the arc-shaped ring groove 721 and the buckle protrusion 7411 are aligned, and the buckle protrusion 7411 is inserted into the arc-shaped ring groove 721 to complete the locking.

[0079] Preferably, the buckle strip 741 is also provided with an assisting spring 760, which is elastic and is at the position 760-1 in the initial state. After the insertion of the plug 720, the assisting spring 760 will be pushed downward to bend elastically and store elastic force. Once the buckle protrusion 7411 exits the arc-shaped annular groove 721, the assisting spring 760 will exert an elastic force on the plug 720 to push the electronic tag, so that the arc-shaped annular groove 721 and the buckle protrusion 7411 are misaligned. At this time, even if there is no magnetic field, the arc-shaped annular groove 721 and the buckle protrusion 7411 will not be locked again, so that it is easy to remove the electronic tag.

[0080] Preferably, the entire locking member 740 can be integrally injection-molded using elastic plastic.

[0081] Figure 3-Figure 7 The methods are all detachable, which is mainly to be torn off when washing clothes or encountering water to avoid damage, and it is also easy to replace.

[0082] See also Figure 8-Figure 9In some embodiments, a temperature sensor is embedded in the electronic tag, which is attached to the user's skin to detect the user's body temperature when in use, and the corresponding body temperature value is subsequently read by a reader. This is a prior art, and the flexible electronic temperature measurement tag technology of Fudan Microelectronics Group can be referenced or adopted.

[0083] See also Fig.10 A layer of heat-insulating film 150 can be fixed on the outside of the electronic tag 100. The heat-insulating film 150 protrudes from the electronic tag 100 and forms an adhesive part 151. The adhesive part 151 is coated with adhesive and is attached to the fixed object 01 by the adhesive when in use. The heat-insulating film 150 is made of heat-insulating film, so that it has certain heat-insulating properties, thereby improving the accuracy of body temperature detection.

[0084] See also Fig.11 A layer of adhesive layer 160 may also be applied on the inner side of the electronic tag 100 by means of adhesive, and the electronic tag 100 may be attached to the fixed object 01 by means of the adhesive layer when in use.

[0085] This design is mainly for ease of use. After combining with the payment function, it can be pasted on the wrist to realize functions such as wireless payment, public transportation payment, body temperature monitoring, etc., and the body health status during use can be monitored through the body temperature monitoring function.

[0086] In some embodiments, see Fig.10 The functional part 200 is respectively mounted with a first temperature sensor 611 and a first humidity sensor 612, and the functional part 200 is mounted on the inner side of the thermal insulation film 150. The adhesive part 151 is attached to the skin, so that the first temperature sensor 611 and the first humidity sensor 612 are closely attached to the skin to detect body temperature and skin humidity.

[0087] See also Fig.11 The functional part 200 is respectively provided with a second temperature sensor 621 and a second humidity sensor 622. The end surface of the functional part 200 is provided with an adhesive layer 160, which is attached to the outside of the clothes so that the second temperature sensor 621 and the second humidity sensor 622 face the air. In this way, the ambient temperature and the ambient humidity can be measured by the second temperature sensor 621 and the second humidity sensor 622 respectively.

[0088] The first temperature sensor 611 and the second temperature sensor 621, the first humidity sensor 612 and the second humidity sensor 622 can be a thermistor and a humidifier, respectively; the thermistor and the humidifier are connected in series between the power supply (the end of the antenna coupling output current) and the pin of the RFID chip. The resistance of the thermistor and the humidifier will change when the temperature and humidity change, so you only need to know the power supply voltage and the current value obtained by the RFID chip to calculate the resistance, and then calculate the corresponding temperature and humidity. The voltage can be obtained by the spectrum analyzer method. The spectrum analyzer is mainly used to analyze the spectrum components of the electrical signal, that is, to convert the electrical signal in the time domain into a representation in the frequency domain, and display the power distribution of the signal at different frequencies. For the voltage signal generated by the coupling current, it can detect which frequency components the signal contains and the power, amplitude and other information corresponding to each frequency component. Its working principle is based on digital signal processing technology such as fast Fourier transform (FFT) of the input signal, and the signal is converted from the time domain to the frequency domain for analysis and display. In the radio frequency identification (RFID) system, there is electromagnetic coupling between the reader and the electronic tag. To detect the characteristics of the voltage signal generated by this coupling in the frequency domain, the input port of the spectrum analyzer can be connected to the location to be detected (such as the coupling area near the electronic tag antenna) through a suitable RF cable. After setting the center frequency, scanning bandwidth, resolution bandwidth and other parameters of the spectrum analyzer, the spectrum diagram of the coupled voltage signal can be seen on the display screen, and its frequency distribution and corresponding power can be understood, thereby having a deeper understanding of the characteristics of the coupled voltage.

[0089] Of course, how to realize passive measurement of temperature and passive measurement of humidity is already an existing technology. This embodiment can directly adopt the existing technology, such as the flexible electronic temperature measurement tag and passive humidity measurement tag of Fudan Microelectronics Group, which are relatively mature technologies. This embodiment can directly adopt the existing technology. In theory, it can measure temperature and humidity without being restricted to the structure and principle given in this embodiment. The above scheme is an optional scheme and is not a necessary technical feature of this embodiment.

[0090] See also Fig.12 , when used, usually Fig.10 , Fig.11 The electronic tags are used in pairs. Fig.10 The electronic tag is named as the internal electronic tag. Fig.11The electronic tag is named as the outer electronic tag. When in use, the inner electronic tag is attached to the inside of the clothes, preferably to the skin, and the outer electronic tag is attached to the outside of the clothes. At this time, the temperature and humidity of the skin, the temperature and humidity of the environment can be collected respectively, so as to infer the patient's body temperature health status, whether sweating, etc. If the body temperature is 40℃ when the ambient temperature is 25℃, then it is obvious that the patient has a fever. Or when the humidity in the clothes is significantly greater than the ambient humidity, it can be considered that the patient is sweating. Combined with the body temperature, if the body temperature does not increase significantly, it means that the humidity in the patient's clothes has increased significantly, and there may be urination, menstruation, etc. If the body temperature increases significantly, there may be sweating. Of course, these can be obtained through limited test methods. Detecting the ambient temperature and humidity is of great significance for subsequent positioning. This design can be integrated with the monitoring system to achieve passive positioning and health status monitoring of the monitored object.

[0091] In normal use, a QR code can be printed on the functional part 200, and the ID of the QR code can be scanned and identified to bind with the ID of the electronic tag, and the ID of the electronic tag is then bound with the user's ID, so that the ID of the electronic tag represents the user's ID.

[0092] In this embodiment, the fixing object is determined according to the actual situation. For some scenes that require puncture or scenes where parts are fixed on the fixing object, the skin is obviously not suitable. This is basic common sense.

[0093] See also Fig.13 In some embodiments, the user's location and walking path can be inferred through parameters such as walking speed, number of user passes, number of people passing the path and location, path relationship, purpose of passage, and path length, thereby providing an approximate positioning reference when positioning is impossible.

[0094] In actual use, the user's wireless payment information, public transportation payment information, etc. are bound to the ID information of the electronic tag, so that the corresponding payment and deduction can be continued after the user ID is identified. The reading device that identifies the user's electronic tag needs to be connected to the Internet, so that the user can be reversely located through the reading device. At the same time, multiple reading devices upload information to form a path, and can calculate parameters such as the user's walking speed and the number of times the user passes. The specific operation steps are as follows:

[0095] S100, obtaining data on the user's walking speed V, the number of times the user passes N1, the number of people passing the path and position N2, the path relationship R, the purpose of passing P, and the path length L;

[0096] S200, parameter weight setting:

[0097] S210, the weight (α) of walking speed (V), the value range is 0.1-0.3. In open areas with smooth traffic and low population density (such as urban parks, open squares, etc.), walking speed has a greater impact on position estimation, and α can be 0.25; in densely populated, traffic-congested or complex indoor environments (such as shopping malls, subway platforms, train station waiting halls, etc.), walking speed is easily disturbed, and α is 0.15.

[0098] S220, the weight (β1) of the number of user passes (N1) is in the range of 0.2-0.4. For users with fixed itineraries (such as office workers and students who commute for a long time), β1 can be set to 0.35; for users with more random activities (such as tourists and freelancers), β1 is set to 0.25.

[0099] S230, the weight (β2) of the number of people (N2) passing through the path and location, with a value range of 0.1-0.3. When analyzing traffic congestion and predicting the user's choice trend on different paths, the number of people has a larger weight. In crowded areas such as main roads and stations during peak traffic hours, B2 takes a value of 0.3; in sparsely populated areas (such as suburban roads late at night and remote parks), β2 takes a value of 0.1.

[0100] S240, the weight (γ) of the path relationship (R), the value range is 0.2-0.3. The path relationship is crucial to determining the possible walking routes of users. In complex transportation networks (such as the central areas of large cities and large transportation hubs), γ is 0.3; in relatively simple areas (such as small towns and villages), the path relationship (R) weight (γ): the value range is 0.2-0.3. The path relationship is crucial to determining the possible walking routes of users. In complex transportation networks (such as the central areas of large cities and large transportation hubs), γ is 0.3; in relatively simple areas (such as small towns and villages), γ is 0.2.

[0101] S250, the weight (δ) of the purpose of travel (P), the value range is 0.2-0.3. When the user's purpose of travel is clear (such as a medical treatment plan determined according to the calendar schedule, or commuting between home and work at a fixed time on weekdays), δ is 0.3, which plays an important guiding role in location estimation; when the purpose of travel is unclear (such as free time, rest days without a specific schedule), δ is 0.2.

[0102] S260, weight (ε) of path length (L), ranging from 0.1 to 0.2. Path length is an important factor when considering the user's physical exertion and total travel time. For long-distance travel scenarios (such as cross-city travel and long-distance commuting), ε is 0.2; for short-distance urban travel (such as activities around the community and short-distance shopping), ε is 0.1.

[0103] S300, dead reckoning

[0104] S310, Initialization

[0105] Suppose the user's current possible location set is S, and initially S is all the location points in the entire monitoring area. At the same time, a temporary location set T is initialized as an empty set.

[0106] S320, screening based on walking speed (V)

[0107] According to the current walking speed V and time interval △T, calculate the possible distance range of the user:

[0108] [d min ,d max ], where d min =(V-△V)△T,d max =(V+△V)△T, △V is the velocity error (determined based on environmental factors, sensor accuracy, and acceleration sensor data). i ∈S, calculate the distance from the last positioning position to p i The distance d i If d i <d min or i >d max , then p i Remove from S. At the same time, the velocity direction (determined by the velocity components on the coordinate axis) is combined to further narrow the range. For example, if the velocity direction points to the east, then the position points in the west direction can be excluded first.

[0109] S330: Optimization based on the number of user passes (N1) and the number of people on the path (N2)

[0110] S331, User Pass Times (N1) Optimization

[0111] Count the number of users passing through the path where the remaining location points are located N / 1. Calculate the difference rate with the current measured number of passes N1:

[0112] △N1=(|N / 1-N1|÷N1)*100%; if △N1 is greater than the set threshold (such as 30%, considering the large dynamic changes in the flow of people), the location point will be removed. For suspicious points with a difference rate within a certain range (such as 20%-30%), the subsequent changes in the number of passes will be continuously observed, and if they are still abnormal, they will be removed.

[0113] S332, Path Number of People (N2) Optimization

[0114] Get the current number of people N on the path and location of the remaining location points / 2. Calculate the difference rate with the number of people N2 passing through the path and location

[0115] △N2=(|N / 2-N2|÷N2)*100%; if △N2 is greater than the set threshold (such as 40%, considering the large dynamic changes in the flow of people), the location point is removed from S. The threshold setting here is relatively loose, because the flow of people itself changes frequently, but a more accurate judgment can be made by combining historical flow data of different time periods. For example, for roads around commercial areas during the morning rush hour on weekdays, the flow of people is usually large. If the current number of people is much lower than the historical level during the same period, the possibility of this location point is reduced.

[0116] S340, consider the constraints of path relationship (R)

[0117] Check whether the connection relationship between the paths where the remaining location points are located and the surrounding paths is consistent with the constructed path relationship, including the physical connection of the path (such as the connectivity of intersections and passages) and the logical connection (such as traffic control, one-way street restrictions, the direction of travel of stairs or elevators, etc.). If not, remove the location point from S. At the same time, consider the dynamic change information of the path (obtained through data interaction with traffic management departments, municipal construction departments, etc.), such as new roads, road construction, temporary control, etc., update the path relationship model in time and re-evaluate the rationality of the location points.

[0118] S350, Adjustment in combination with traffic purpose (P)

[0119] According to the inferred purpose of travel, analyze whether the environmental information around the remaining location points (such as the type of nearby facilities, operating status, etc.) matches the purpose of travel. Use machine learning models to analyze the user's behavior patterns near the location point (stay time, movement trajectory, interaction with surrounding devices, etc.). For example, if the purpose of travel is to seek medical treatment, then the location point where there is a hospital nearby and the user's stay time in the hospital area meets the medical treatment process is more likely to be the user's location; if it does not meet the purpose of travel, the location point is removed from S. Of course, the purpose of travel can also be input into the system for accurate inference.

[0120] S360, using the path length (L)

[0121] According to the current walking speed V and the length of the path where the remaining position point is located, the theoretical travel time T = L ÷ V required to pass through the path is calculated. Considering the dynamic changes in speed (adjusted according to the path type, environmental factors, user historical speed data, and the speed of transportation vehicles) and the actual travel conditions of the path (obtaining information such as congestion and obstacles through environmental sensors, combined with the real-time operation status of buses and subways), the time range [Tmin ,T max ]. Combined with the actual time interval from the last positioning to the current one, if the theoretical travel time of a path where a certain location point is located does not match the actual time interval (not in [T min ,T max ]), then remove the position point from S.

[0122] S370, temporary set update and finalization

[0123] The remaining location points after the above steps are assigned to the temporary set T. Probability evaluation is performed on the location points in T, and the probability score of each location point is calculated based on the relationship between each parameter and the weight. For example, location point p j ∈S score, where:

[0124] S core (p j )=α*V Score(pj) +β1*N 1Score(pj) +β2*N 2Score(pj) +γ*R Score(pj) +δ*P Score(pj) +ε*L Score(pj)

[0125] Where V Score(pj) 、N 1Score(pj) 、N 2Score(pj) , R Score(pj) , P Score(pj) , L Score(pj) They are the scoring functions of the location points based on walking speed, number of user passes, number of people on the path, path relationship, purpose of travel, and path length (calculated according to the matching degree between each parameter and the location point, with a value range of 0-1), S core (p j ) is the specific score. Select several locations with the highest scores (such as the first 10, determined according to the system accuracy requirements and computing resources) as the final user possible location set.

[0126] S400, prediction and correction

[0127] Based on the final determined set of possible user locations, path relationships, trends of various parameters, and the operating status of transportation tools, the location prediction model is used to predict the user's possible subsequent path and location. When the next electronic tag positioning information (through 5G-Α base stations, readers, bus card readers, subway card readers, etc.) is updated, the prediction results are compared with the actual data. If there is a deviation, the parameter weights, model parameters and algorithms are adjusted according to the deviation (such as learning deviation adjustment strategies through reinforcement learning algorithms) to achieve correction of the prediction. At the same time, the new positioning data is fed back to the big data storage system for further data analysis and model training to continuously improve the accuracy and adaptability of the system.

[0128] Preferably, S100 further includes S110, collecting and processing walking speed (V) parameters.

[0129] S111, Data Collection

[0130] The position data is obtained by timing the positioning of the electronic tag by the reading device or the position of the reading device that reads the electronic tag is directly used as the position data of the electronic tag. Assuming that the position coordinates of the electronic tag at time t1 are (x1, y1, z1) and the coordinates at time t2 are (x2, y2, z2), the position change is:

[0131] Time interval △t=t2-t1, walking speed Among them, x, y, and z represent the left and right directions, the front and back directions, and the up and down directions, respectively. The front of the coordinate is the direction of the user's walking, the back direction is the direction directly behind the user, the origin is the reading device, the left and right direction is the direction of the line connecting the left and right hands of the user, and the up and down direction is the height direction of the user. If the three parameters x, y, and z cannot be obtained, the length between the two positioning information is used as △d. When there are multiple paths corresponding to the origin or the path is uncertain, the left and right directions, the front and back directions, and the up and down directions are manually specified, and the coordinate relationship between each direction and each path is recorded. The direction of each reading device can be manually specified. This method has a small amount of calculation, unified coordinate values, and is easy to calculate, such as the east-west direction is the front and back direction, the north-south direction is the left and right direction, and the vertical direction is the up and down direction.

[0132] To improve the accuracy of speed data, the sliding window average method is used to smooth the speed value. Select an appropriate window size (such as the speed value calculated by the past 10 positioning calculations) and calculate the average speed in the window as the current effective speed. In addition, consider the impact of environmental factors on speed. For example, in crowded places, the speed is appropriately adjusted according to the number of people and the degree of environmental crowding (establish a crowding degree-speed adjustment coefficient model).

[0133] S112, establish speed error model:

[0134] The determination of the velocity error ∆V is a key factor. It is a function of many variables, including the ambient temperature T env 、Ambient humidity H env Etc. Through a large amount of experimental data and machine learning algorithms (such as multivariate linear regression), a velocity error model △V=f(T env , H env ). For example, in a high temperature and high humidity environment, the signal propagation speed may change, resulting in a decrease in positioning accuracy, which in turn affects speed calculation; when the signal interference intensity is high, the communication between the electronic tag and the reader / writer may experience packet loss or delay, which can also cause errors in speed calculation.

[0135] In different scenarios (such as indoors, outdoors, underground, etc.), data is collected to train and optimize the model. For example, for indoor scenarios, the focus is on the reflection and attenuation of the signal by the wall, as well as the relatively stable values ​​of indoor temperature and humidity; for outdoor scenarios, the impact of weather changes on the signal, as well as the differences in signal propagation in different geographical environments (such as mountains and plains) must be considered.

[0136] The user's travel direction and path are calculated based on the user's successive positioning information, combined with the last valid positioning information and the possible path of this positioning information. According to the speed direction, in the map data of the Geographic Information System (GIS), a sector area is delineated along the speed direction with the current position as the starting point. The angle of the sector is determined according to the stability and accuracy requirements of the speed direction. For example, if the speed direction does not change much in a short period of time, the sector angle can be smaller (such as 30%); if the speed direction changes frequently, the sector angle can be appropriately increased (such as 90°). Only the position points within this sector area are retained in the possible position set S.

[0137] Preferably, S100 further includes S120, collecting and processing the number of user passes (N1) parameters:

[0138] Data collection: Readers at key path nodes (such as entrances and exits of transportation hubs, passageways of large buildings, etc.) as well as bus and subway card readers are responsible for recording the passage of electronic tags. Each time an electronic tag is detected passing, the device records the passing time, the electronic tag identification, and its own location or transportation information. Through statistical analysis of these records, the number of times users pass through different path nodes and transportation vehicles can be obtained.

[0139] To avoid miscounting due to signal interference or equipment failure, multiple verification mechanisms are set up. For example, in addition to setting signal strength thresholds and time windows (such as signal strength above -60dBm and stable within 1-3 seconds), cross-verification is also performed in combination with information from adjacent readers or card readers. For bus and subway card readers, the number of rides is also checked for rationality based on the operating schedule and vehicle routes to remove abnormal data.

[0140] More preferably, S120 further includes

[0141] S121, User Pass Times (N1) Optimization:

[0142] S121. Passage frequency statistical model:

[0143] In addition to simple counting, a more complex statistical model of the number of passes is established. Considering the time factor, a day is divided into different time periods (such as peak hours, off-peak hours, and night time). The user's travel pattern may be different in each time period. For each time period, a Markov chain-based model is established, where the state is a different key path node or area, and the transition probability represents the possibility of the user transferring from one state to another. This model is trained with a large number of users' historical travel data to obtain the transition probability matrix P between different path nodes in each time period. ij (t), where i and j are different path nodes and t is the time period.

[0144] S122, abnormal passage times detection:

[0145] When calculating the difference rate △N1 between the current number of passes and the number of passes predicted by the model, the judgment of abnormal situations is not based only on a fixed threshold. A dynamic threshold method is used to adjust the threshold according to the time period, day of the week, and the stability of recent user behavior. For example, for peak hours on weekdays, the threshold can be set narrower because user behavior is relatively stable; while for weekends or nighttime hours, user behavior may be more random, and the threshold can be appropriately relaxed. At the same time, if the number of passes of users in the recent period (such as in the past week) fluctuates greatly, the threshold is also adjusted accordingly to reduce misjudgment.

[0146] S123, suspicious point processing:

[0147] For points whose difference rate is within the suspicious range, time series analysis methods are used for further observation. For example, the autoregressive moving average model (ΑRMΑ) is used to model and predict the number of passes of these points. If the subsequent actual number of passes continues to deviate from the model prediction value, and the deviation is getting larger and larger, the position points on the path where these suspicious points are located are removed from S.

[0148] More preferably, S121 further includes:

[0149] Markov decision process (MDP) model construction and training details: When building an MDP model, the definition of the state space needs to be more refined. In addition to the current location and possible path choices, time factors (such as different hours of the weekday, holidays, etc.), weather conditions (sunny, rainy, snowy, etc. affect the comfort and feasibility of the path) and the user's recent behavior patterns (such as whether a specific type of path is frequently selected recently, such as preferring paths with shading facilities) should also be considered. For each state, the action space is all the path options that can be connected to the current location.

[0150] The design of the reward function should take into account many factors. In terms of path length, the length of each path is accurately calculated based on the actual geographic information data and normalized. Shorter paths are given higher positive rewards, but the weight differences of different types of paths (such as pedestrian paths, bicycle paths, and motor vehicle paths) need to be considered. The flow of people factor is obtained through a real-time flow prediction model. Paths with less flow of people are rewarded. Different reward values ​​can be set according to the degree of crowding. For example, the flow of people can be divided into different levels (sparse, moderate, and crowded), corresponding to different reward value ranges.

[0151] The degree of matching between the purpose of travel and the attributes of the path is also key. If the purpose of travel is shopping, a higher reward will be given if there are many shopping malls around the path and it is easily accessible. For the purpose of medical treatment, the reward for the path connected to the hospital and with suitable entrances and parking facilities (for driving users) is higher. In addition, factors such as the safety of the path (such as the increase in the reward for the path corresponding to the area with good lighting and low crime rate) and the comfort of the environment (green and quiet paths are more popular) should also be considered.

[0152] When training the MDP model, a large amount of historical user path data is used and the data is preprocessed. Each historical path is labeled, including information such as the start time, weather conditions, and the user's purpose of travel (inferred through the various methods mentioned above). Monte Carlo method or temporal difference learning algorithm (such as Q-learning) is used for training. During the training process, the state-action value function Q(S, α) is continuously adjusted, and the model converges to a stable strategy through multiple iterations, that is, there is a relatively accurate value estimate for the best action (path selection) in each state.

[0153] Preferably, S100 further includes S130, parameter collection and processing of the number of people N2 passing through the path and location:

[0154] S131. The number of people passing through each reading device location and the number of people passing through each reading device are obtained through sensors, electronic tags, manual counting, base station connection data, manual input, etc., especially the paths and reading devices passed by the same person multiple times are marked. For points with a difference rate within a suspicious range, a time series analysis method is used for further observation, such as the number of repeated passes within 1 minute is not counted repeatedly.

[0155] S132, Optimization of the number of people on the path (N2):

[0156] S1321. Crowd flow prediction model:

[0157] A crowd flow prediction model is constructed using historical crowd flow data, time information, weather data, and surrounding activity information. A deep learning model (such as convolutional neural network (CNN) combined with long short-term memory network (LSTM)) is used, where CNN is used to extract temporal and spatial features (such as the change pattern of crowd flow at different times of the day, the correlation of crowd flow in different regions), and LSTM is used to process the long-term dependencies of time series data (such as the similarity of crowd flow in the same time period of each week, the impact of holidays on crowd flow). Weather data (such as temperature, humidity, whether it is raining, etc.) and surrounding activity information (such as shopping mall promotions, concerts, etc.) are used as additional input features to improve the prediction accuracy of the model.

[0158] S1322 Abnormal flow of people:

[0159] When calculating the difference rate △N2, the dynamic threshold method is also used. The threshold is adjusted according to the type of path (such as main roads, branch roads, commercial streets, etc.), time period, and weather conditions. For example, for commercial streets, during holidays and promotions, the normal fluctuation range of human traffic is large, and the threshold should be relaxed accordingly; while for branch roads near residential areas, the human traffic is relatively stable, and the threshold can be set narrower. At the same time, consider the seasonal changes in human traffic data. For example, during the peak tourist season, the human traffic on roads around certain attractions will increase significantly, and the model must be able to adapt to this change.

[0160] S1323. Optimization based on the trend of passenger flow:

[0161] In addition to the current difference in traffic flow, the traffic flow trend is analyzed. If the traffic flow on a certain path shows a rapid increase or decrease trend in a short period of time (such as the past 15 minutes), and the user's travel behavior does not match this trend (such as the traffic flow on the path increases rapidly, but the number of user passes does not increase accordingly), then the location point on the path is removed from S. This can be achieved by performing first-order and second-order difference analysis on the traffic flow data to determine the speed and acceleration of traffic flow changes.

[0162] Preferably, S100 further includes S140, constructing and updating path relationship (R) parameters:

[0163] S141, Initial Construction:

[0164] Collect a variety of geographic information data, including high-precision map data (such as city street maps, building floor plans), transportation network data (such as bus and subway line maps), three-dimensional terrain data, etc. These data are integrated through geographic information system (GIS) technology to build a path relationship model. In the model, the path is represented by a graph structure, the nodes represent important geographical locations (such as intersections, stations, building entrances and exits, etc.), and the edges represent the roads, channels or traffic routes connecting the nodes. Each edge is annotated with multiple attributes, such as length, width, slope, direction of travel, allowed types of transportation, road grade, etc.

[0165] S142, Dynamic Update:

[0166] Establish a data sharing mechanism with traffic management departments, municipal construction departments, bus and subway operating companies, etc., obtain real-time information on road construction, traffic control, new line opening, station changes, etc., and update the path relationship model in a timely manner. At the same time, optimize the path relationship based on the user's actual travel data (travel records of electronic tags). For example, when it is found that a large number of users choose a shortcut that is not marked on the map, the path relationship model is modified accordingly.

[0167] S143. Consider the constraint refinement of path relationship (R):

[0168] S1431, path relationship update mechanism:

[0169] Establish a real-time path relationship update mechanism and maintain connections with multiple data sources. In addition to data exchange with traffic management departments and municipal construction departments, it also connects to real-time traffic information platforms, road construction announcement websites, and traffic-related information on social media. Useful information is extracted from these data sources through natural language processing technology, such as the start and end times of road construction, specific measures for traffic control, and detailed information on newly opened roads. This information is promptly converted into updated content of the path relationship model, for example, modifying the traffic status of the road (from passable to construction closure), adjusting the connection relationship of the path (adding or deleting connections at intersections), and updating the attributes of the path (such as adjusting the width attribute when the road becomes narrower due to construction).

[0170] Optimize path relationships using user feedback and group wisdom. Encourage users to report road conditions (such as new obstacles, temporary traffic control, etc.) through mobile applications, and give higher credibility to the same information reported by a large number of users. At the same time, analyze the actual travel paths of users. If it is found that a large number of users have chosen a route that does not match the existing path relationship model, further investigate the cause. It may be that there are undiscovered new paths (such as newly opened channels between communities) or the attributes of existing paths are wrong (such as one-way street signs). Update the path relationship model in a timely manner.

[0171] S1432, path relationship checking algorithm:

[0172] When checking the matching of the location point and the path relationship, a graph search algorithm is used. The path relationship model is represented as a directed graph G = (V, E), where V is a set of nodes (representing intersections, stations, buildings, etc.) and E is a set of edges (representing roads, passages, traffic routes, etc.). Each edge has a series of attributes, such as travel direction, vehicle restrictions, construction status, etc. For each remaining location point p i ∈S, starting from the node where the position point is located, use the depth-first search (DFS) or breadth-first search (BFS) algorithm to traverse the graph and check whether all paths starting from the node meet the constraints of the path relationship. For example, if an edge is marked as a one-way street and the direction is opposite to the user's travel direction, then the subsequent nodes connected by the edge do not meet the requirements, and these nodes and the position points on the related paths are removed from S.

[0173] Considering the hierarchical structure of path relationships, for complex transportation hubs or large buildings, there are multiple layers of path relationships (such as the connections between different floors in subway stations and different areas in large shopping malls). When checking path relationships, not only the connections on the plane should be considered, but also the vertical connectivity. This hierarchical structure is represented by three-dimensional coordinates and floor information to ensure that the user's location and path in three-dimensional space conform to the path relationship model.

[0174] Preferably, S100 further includes S150, inferring and processing the traffic purpose (P) parameter:

[0175] S151, Data Collection:

[0176] Infer the purpose of travel by integrating multi-source data. Analyze the time and frequency of electronic tags in different types of places (such as shopping malls, hospitals, schools, office areas, etc.), combined with the boarding and alighting locations of buses and subways and payment information (such as consumption records at specific merchants, registration and payment records at hospitals, etc.). At the same time, obtain application usage data on users' smart devices (such as mobile phones), including navigation destination settings, query records of travel applications, calendar schedules, etc.

[0177] S152, inference model construction:

[0178] Use machine learning algorithms to build a travel purpose inference model. Algorithms such as decision trees, support vector machines, and neural networks can be used. The collected user behavior data is used as input features, and the travel purpose (such as going to work, shopping, medical treatment, entertainment, travel, etc.) is used as the output category to train the model. For example, by analyzing that users often stay in an office building area near a subway station for a long time on weekday mornings and have regular dining consumption records in the surrounding area, combined with the work schedule information in the mobile phone calendar, the training model recognizes this behavior pattern as the travel purpose of going to work.

[0179] S153, Model Update:

[0180] As user behavior changes, new data is constantly accumulated, and the purpose of travel inference model is regularly retrained and updated. For example, when a user changes his or her workplace or develops new travel habits (such as starting to exercise and going to the gym more often), the model can adapt to these changes in a timely manner and accurately infer the new purpose of travel.

[0181] S154, Adjustment and refinement in combination with the purpose of passage (P)

[0182] S1541. Inference of traffic purpose based on multimodal data fusion:

[0183] In addition to the aforementioned use of electronic tags to measure the length of stay in different places, bus and subway boarding and alighting locations, and payment information, more types of data can be further integrated. For example, the dynamics posted by users on social media (such as the location where the photos were taken, check-in information), and the interaction records with smart voice assistants (such as asking about nearby restaurants and hospitals, etc.) can be analyzed. Through multimodal data fusion technology (such as the multimodal fusion architecture in deep neural networks), data from these different sources can be integrated together to improve the accuracy of inference of travel purpose.

[0184] S1522. Inference of travel purpose based on situational awareness:

[0185] Consider the user's context information, including the current time, day of the week, weather conditions, and surrounding environment (such as nearby festivals, sports events, etc.). For example, if it is a sunny weekend day and the user appears near a park, his purpose of travel may be leisure and entertainment; if it is a weekday morning, near a school and there are a large number of students passing by, the user's purpose of travel may be to send the child to school. By establishing a context-travel purpose mapping model and using machine learning algorithms (such as decision trees, Bayesian networks, etc.) to learn the probability distribution of travel purposes in different contexts, the accuracy of inference can be improved.

[0186] S1543, location point and passage purpose matching assessment:

[0187] Environmental feature matching:

[0188] For each remaining position point p i ∈S, and analyze the environmental characteristics of its surroundings in detail. Use geographic information system (GIS) data to obtain information such as the type of surrounding facilities (such as hospitals, schools, shopping malls, restaurants, etc.), scale, business hours, etc. At the same time, use image recognition technology (if camera data is available) or user evaluation data (such as user comments on online maps) to further understand the details of the surrounding environment, such as the distribution of hospital departments and the characteristics of restaurant dishes. Match these environmental characteristics with the inferred purpose of travel. For example, if the purpose of travel is to seek medical treatment, there is a large hospital around the location point and the hospital has the department required by the user (based on the user's historical medical records or search information), then the matching degree of the location point is high.

[0189] Behavioral pattern matching:

[0190] The machine learning model is used to analyze the user's behavior patterns near the location point. In addition to the dwell time, movement trajectory, and interaction with surrounding devices, the user's activity intensity in the area is also considered (such as changes in walking speed, the number of stop locations, etc.). For example, if the purpose of the passage is shopping, the user's activity intensity in the mall area is high (frequent movement between different stores), and the dwell time is in line with the general shopping time range (based on the shopping behavior data statistics of a large number of users), then the location point is likely to meet the purpose of shopping. For each location point, a matching score with the passage purpose is calculated based on the results of matching environmental features and behavioral patterns, and location points with scores below a certain threshold are removed from S.

[0191] Preferably, S100 further includes S160, path length (L) parameter calculation and calibration:

[0192] S161, initial calculation:

[0193] For paths such as roads and passages, their lengths are calculated based on the scale and coordinate information in the geographic information data. For straight paths, the Euclidean distance formula is used for calculation; for curved paths, numerical integration methods (such as the trapezoidal method and the Simpson method) are used to divide the path into multiple small segments for approximate calculations. For transportation routes (such as bus and subway lines), the official mileage data provided by the operating company is used to determine the distance, while fine-tuning is performed based on factors such as station spacing and line turning radius.

[0194] S162, calibration mechanism:

[0195] The path length is calibrated using the user's actual travel time and walking speed data. A path length error model is established that takes into account the impact of environmental factors (such as detours caused by road construction, weather affecting walking speed), transportation operation conditions (such as temporary bus diversions, changes in itinerary caused by subway delays), etc. on path length. For example, if the distance calculated based on the user's walking speed and travel time on a certain path is significantly different from the initially calculated path length, analyze the possible influencing factors and adjust the path length accordingly.

[0196] S163. Accurate refinement using path length (L)

[0197] S1631, speed dynamic change model:

[0198] The dynamic change of speed is affected by many factors, so a comprehensive dynamic change model of speed is established. In addition to the path type, environmental factors and the user's historical speed data, the user's physical condition (age, disease condition, walking speed, fatigue level, etc. entered in advance) and the information of the items carried (if the user carries heavy objects, the walking speed may slow down). A fuzzy logic system or a neural network model is used to represent the relationship between the speed V and these factors:

[0199] V=g(PathType),T env , H env , V history , HealthStatus, HealthStatus), where

[0200] PathType is the path type (such as flat road, stairs, slope, etc.), HealthStatus is the user's physical condition, Load is the information of the items carried, V history is the speed that has been measured before, T env , H env are the ambient temperature and humidity measured previously respectively.

[0201] This model is trained and optimized through a large number of experiments and actual data collection in different scenarios (such as different weather conditions and different combinations of path types). For example, on a sloped road in rainy days, the user's walking speed can be accurately predicted based on their physical condition and the items they carry. For those who take public transportation, the speed change can be determined based on the type of transportation (such as bus, subway, taxi), the degree of crowding in the vehicle, and traffic conditions (such as road congestion and train delays).

[0202] S1632. Evaluation of actual traffic conditions on the route:

[0203] Congestion assessment based on multi-source data fusion:

[0204] Integrate multi-source data to evaluate the actual traffic conditions of the route, especially the congestion. In addition to the road congestion information obtained by environmental sensors, real-time operation data of buses and subways (such as vehicle intervals and delay information), taxi global positioning system (GPS) data (obtaining taxi speed and route information through cooperation with taxi dispatch platforms) and user feedback (such as traffic jams reported by users on mobile applications) are also used. Data fusion algorithms (such as weighted average method, Kalman filter, etc.) are used to integrate data from these different sources to obtain more accurate route congestion information.

[0205] Obstacle and construction information integration:

[0206] Integrate information such as road construction and obstacles into the assessment of actual path conditions. By sharing data with municipal construction departments and traffic management departments, obtain the specific location, scale and duration of road construction, as well as information on obstacles (such as traffic accident sites, temporary piles, etc.). In path length calculations and time estimates, consider the impact of these factors on travel speed. For example, if a section of a road is under construction, the speed calculated for passing through the section should be adjusted according to the type of construction (such as full closure construction, half-width construction) and traffic control measures (such as one-way traffic, detour instructions).

[0207] Time range calculation adjustments:

[0208] According to the speed dynamic change model and the evaluation results of the actual traffic conditions of the path, the time range of passing the path [T min ,T max ], the speed under the most ideal traffic conditions (such as clear roads, good physical condition of users, and punctual operation of transportation vehicles) is used for calculation; for T max , considering the speed under the worst traffic conditions (such as severe congestion, detours caused by road construction, user fatigue or carrying heavy objects). At the same time, considering the mutual influence between different paths (such as congestion on one road may cause changes in the flow of adjacent roads), a system dynamics model or a microscopic traffic simulation model is used for more accurate simulation when calculating the time range.

[0209] Preferably, in S300, S380 is also included, when the set is updated and the final determination is refined.

[0210] S381, location point score calculation refinement:

[0211] Score function V based on walking speed scpre (p j ): According to the speed error △V and position point p jIf the position point is on the center line of the sector area and the speed calculation error is small, then V scpre (p j ) is close to 1; if the position point deviates far from the center line or the speed error is large, V scpre (pj) is low. The specific calculation can use Gaussian function or other suitable similarity function, such as where d angle is the angular deviation between the position point and the center line of the velocity direction, σ angle and σ V is the standard deviation parameter of the angle and velocity errors.

[0212] S382. Dynamic adjustment considering real-time traffic information and environmental factors:

[0213] The acquisition and integration of real-time traffic information requires the support of multi-source data. The traffic flow data of major roads are obtained from the traffic flow monitoring system of the traffic management department. The real-time location, speed, delay and other information of public transportation vehicles are obtained through the interface with the bus company and subway operator. At the same time, user feedback (such as traffic jams reported by users of mobile applications) and crowdsourcing data of the vehicle navigation system (driving speed and route information of a large number of vehicles) are used. For each path, its current congestion level is evaluated based on these data. For example, for a road path, if the traffic flow exceeds a certain threshold and the average vehicle speed is lower than a certain proportion of the normal driving speed (such as 30%), the road is considered to be severely congested, and the selection probability of the path in the MDP model is reduced. For bus or subway paths, if the vehicle is delayed for more than a certain time or the degree of congestion in the vehicle is too high (judged by in-vehicle sensors or estimated number of passengers getting on and off the vehicle), its selection probability is adjusted accordingly.

[0214] The impact of environmental factors also needs to be considered dynamically. Weather information is obtained through the data interface with the meteorological department. If it rains, the probability of selecting open-air paths without shelter facilities is greatly reduced; if it snows or freezes, paths with large slopes or complex road conditions (such as roads that have not been cleared in time) are given a lower probability of selection. At the same time, temporary changes in the environment (such as dust and noise caused by road construction, or crowds caused by large-scale events nearby) will also affect the attractiveness of the path. The path selection probability is adjusted through real-time monitoring and data analysis (such as information about activities on social media and construction notices issued by municipal departments).

[0215] S383, location prediction distribution:

[0216] S3831. Position prediction model for walking conditions:

[0217] Based on the model based on Brownian motion or random walk, the user's walking behavior is further refined. Consider the changes in the user's stride length and cadence, and obtain more accurate stride length and cadence information through active input data fusion. Establish a stride length and cadence change model based on different terrains (such as flat ground, uphill, downhill) and environments (such as crowded crowds, open squares). For example, when going uphill, the stride length decreases and the cadence may increase; in a crowded crowd, both the stride length and cadence will be limited.

[0218] Combine the geometry and features of the path, such as the curvature of the path, the presence of obstacles, etc. For curved paths, predict the user's position change at the bend based on the curvature of the path and the user's turning habits (learned through historical data). Consider the user's pause behavior during walking, such as waiting for traffic lights at intersections, stopping at the door of a store, etc., and more accurately predict the user's position on the path by analyzing the distribution of the user's stay time in similar scenarios (based on historical data).

[0219] S3832. Location prediction when riding a vehicle:

[0220] For buses and subways, in addition to the vehicle's speed, stop information, and the user's boarding location and time, the flow of people in the vehicle must also be considered. The number of people getting on and off the vehicle and the distribution of passengers in the vehicle are monitored through on-board sensors (such as infrared sensors and pressure sensors installed near the doors) to infer the user's movement possibility in the vehicle. For example, if the front of the vehicle is crowded and the back is relatively loose, the user is more likely to move to the back of the vehicle, and this movement behavior will affect the user's position prediction on the driving route.

[0221] S3833. For taxis or online ride-hailing vehicles, in addition to the vehicle's GPS positioning information, the driver's driving habits (such as driving speed preference, whether he likes to take shortcuts, etc.) and the impact of traffic control on the driving route must also be considered. The vehicle's driving route and user's location changes are predicted by analyzing the driver's historical driving data and current traffic control information (such as one-way restrictions, temporary traffic control areas). At the same time, the interaction between the user and the driver (such as the user's request to change the route, adjust the destination, etc.) is considered, and this information is obtained through data interaction with the taxi platform and the location prediction is updated in a timely manner.

[0222] S384, Uncertain factors in processing path:

[0223] In the Bayesian network or hidden Markov model, the hidden state and observed state are defined in more detail. The hidden state includes the type of traffic control (such as temporary road closure, lane restriction), the progress of road construction (such as full closure, half-width construction stage), the failure of transportation tools (such as bus breakdown, subway failure), etc. The observed state is the actual positioning information of the user and the relevant data obtained by the electronic tag (such as transportation tool information obtained through the card reader).

[0224] The model is trained with a large amount of historical data to learn the probabilistic relationship between hidden states and observed states. For example, when it is observed that the card reader on a certain road has not detected any electronic tags for a period of time (possibly due to road closure), the probability of the hidden state (road closure) in the Bayesian network is adjusted in combination with the information released by the traffic management department. At the same time, the predictive ability of the model is used to infer the most likely hidden state based on the current observed state when new uncertainties arise, thereby more accurately predicting the probability distribution of the user's location on the path. For example, if a user is on a certain subway line and the positioning information is suddenly interrupted, the user's possible location and subsequent itinerary changes are predicted by combining the fault information released by the subway operator and the fault-location relationship learned in the model.

[0225] Preferably, in S400, a correction process is also included:

[0226] S410, Deviation Analysis:

[0227] S411, Multi-dimensional deviation calculation and quantification:

[0228] In addition to the Euclidean distance d error 、Direction angle deviation θ error and path selection bias Path diff In addition, the time deviation analysis is added. Calculate the difference between the predicted arrival time at a certain location and the actual arrival time t error , considering the impact of different travel modes and path lengths on time, the time deviation is normalized. For example, for short-distance walking paths, smaller time deviations have larger weights; for long-distance paths by public transportation, the tolerance for time deviations can be appropriately increased due to factors such as traffic congestion.

[0229] For the distance deviation d error , not just a simple straight-line distance calculation between two points. In complex environments (such as multi-story buildings and underground passages), consider the distance calculation in three-dimensional space, and adjust the distance calculation method according to the accessibility of the path (such as whether there are stairs, elevators, passage closures, etc.). For example, when calculating the distance deviation between different floors in a building, it is necessary to consider whether the user needs to use specific stairs or elevators, as well as the use of these facilities (such as elevator maintenance, crowded stairs).

[0230] Direction angle deviation θ error The calculation of the direction of the path and the geographic coordinate system must be combined. For complex intersections and curves, the deviation is accurately calculated based on the angle between the design angle of the path (obtained from the geographic information system) and the actual direction of the user's movement. At the same time, the user's inertia and possible pauses during the turning process (such as fine-tuning the direction when waiting for the traffic light at the intersection) are considered to conduct a more detailed analysis of the direction angle deviation.

[0231] Path selection deviation diff The evaluation should consider the similarity and substitution relationship of the paths. If the predicted path and the actual path are within a certain distance and have similar destinations (such as two different roads leading to the same shopping mall), the deviation weight is relatively low; if the two paths are completely different and the destinations are very different (such as predicting to go to the hospital but actually going to the school), the deviation weight is higher. In addition, it is necessary to analyze the possible reasons why users choose different paths, such as whether the deviation is caused by temporary road construction, traffic control or personal preference (such as choosing a quieter side road).

[0232] S412, Deviation weight setting and comprehensive deviation index calculation:

[0233] Deviation weights are set according to different scenarios and application requirements. In scenarios with high requirements for location accuracy (such as emergency rescue and high-precision navigation), the weights of distance deviation and direction angle deviation are higher; in scenarios that focus on travel efficiency (such as daily commuting), the weights of time deviation and path selection deviation are relatively increased. The specific values ​​of the weights are determined through a large number of simulation experiments and actual user data feedback. For example, in a high-precision navigation scenario, the distance deviation weight w1 can be set to 0.5, the direction angle deviation weight w2 to 0.3, the time deviation weight w3 to 0.1, and the path selection deviation weight w4 to 0.1.

[0234] The calculation of the comprehensive deviation index ErroeInder adopts the weighted summation method, but nonlinear adjustment can be added according to the severity of the deviation. For example, when a certain deviation value exceeds a certain threshold, its impact on the comprehensive deviation index increases exponentially to highlight the importance of serious deviation to system correction. This nonlinear adjustment can be achieved using piecewise functions, such as when d error <d threshold When participating in the calculation of w1*d error ; When d error ≥d threshold When w1*(d error ) 2 Participate in the calculation, where d threshold is the distance deviation threshold, which needs to be determined based on the system accuracy and actual conditions, that is, it can be obtained through a limited number of experiments.

[0235] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by technicians in the field to which this application belongs.

[0236] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A passive tracking and positioning method, including an electronic tag carried or worn by a user, with the user's ID information built into the electronic tag. Subsequently, the system associates the user with the ID information as the user's identity identification mark; A reading device reads the information of the electronic tag and uploads the read information to a server or stores it in a specified device through a network; A map or building structure diagram is stored in advance in a database. Combining the position of the reading device in the stored map or building, the electronic tag can be positioned. It is characterized by including the following steps: S100. Obtain the walking speed V of the user, the number of times N1 the user passes through, the number of people N2 passing through this path and position, the path relationship R between the possible passing paths, the purpose P of the user's passage, and the length L of the possible passing path; S200. Parameter weight setting: S210. The weight (α) of the walking speed (V); S220. The weight (β1) of the number of times N1 the user passes through; S230. The weight (β2) of the number of people N2 passing through this path and position; S240. The weight (γ) of the path relationship (R); S250. The weight (δ) of the purpose of passage (P); S260. The weight (ε) of the path length (L); S300, position estimation, the walking speed, the number of user passes, the number of people on the path, the path relationship, the purpose of the pass, and the path length are scored by the corresponding scoring functions. The scoring functions of the walking speed, the number of user passes, the number of people on the path, the path relationship, the purpose of the pass, and the path length are respectively denoted as V Score(pj) 、N 1Score(pj) 、N 2Score(pj) , R Score(pj) , P Score(pj) , L Score(pj) ; Calculate the probability score S of each location point based on the relationship between each parameter and weight core (p j ): S core (p j )=α*V Score(pj) +β1*N 1Score(pj) +β2*N 2Score(pj) +γ*R Score(pj) +δ*P Score(pj) +e*L Score(pj) Select several position points with the highest scores as the final set of possible user positions, and deduce the path the user passes through based on the predicted position, the previous positioning position, the walking speed, and the passable paths.

2. The passive tracking and positioning method according to claim 1, characterized in that: S300 also includes: S310. Initialization Set the current set of possible user positions as S. Initially, S is all the position points within the entire monitoring area; at the same time, initialize a temporary position set T as an empty set; S320. Screening based on the walking speed (V) According to the current walking speed V and the time interval △T, calculate the distance range that the user may move: [dmin, dmax], where dmin = (V - △V)△T, dmax = (V + △V)△T, and △V is the speed error; for each position point pi ∈ S, calculate the distance di from the previous positioning position to pi. If di < dmin or di > dmax, then remove pi from S; further narrow the range in combination with the speed direction; S330. Optimization based on the number of times N1 the user passes through and the number of people N2 on the path S331. Optimization of the number of times N1 the user passes through Count the number of users passing through the path where the remaining location points are located N / 1. Calculate the difference rate with the current measured number of passes N1: △N1=(|N / 1-N1|÷N1)*100%; if △N1 is greater than the set threshold, the location point is removed; for suspicious points with a difference rate within a certain range, the subsequent changes in the number of passes are continuously observed, and if they are still abnormal, they are removed; S332. Optimization of the number of people N2 on the path Get the current number of people N on the path and location of the remaining location points / 2. Calculate the difference rate between the number of people passing through the path and location N2: △N2 = (|N / 2-N2|÷N2)*100%; if △N2 is greater than the set threshold, the position point is removed from S; S340. Consider the constraint of the path relationship (R) Check whether the connection relationship between the path where the remaining position points are located and the surrounding paths conforms to the constructed path relationship, including physical connection and logical connection of the paths; if not, remove this position point from S; S350. Adjustment in combination with the purpose of passage (P) According to the inferred purpose of passage, analyze whether the environmental information around the remaining position points matches the purpose of passage, and use a machine learning model to analyze the user's behavior pattern near this position point; or input the purpose of passage through the system for accurate speculation; S360. Precision using the path length (L) According to the current walking speed V and the length L of the path where the remaining position point is located, the theoretical travel time T = L ÷ V required to pass the path is calculated; the time range [T min ,T max ]; Combined with the actual time interval from the last positioning to the current one, if the theoretical travel time of the path where a certain location point is located does not match the actual time interval, then the location point is removed from S; S370. Update and final determination of the temporary set The remaining location point set after the above steps is assigned to the temporary set T, and the probability of the location points in T is evaluated. The probability score S of each location point is calculated based on the relationship between each parameter and the weight. core (p j ).

3. The passive tracking and positioning method according to claim 1, characterized in that: It also includes S400. Prediction and correction: Based on the final determined set of possible user locations, path relationships, parameter trends, and vehicle operating status, the location prediction model is used to predict the user's possible subsequent path and location. When the next electronic tag positioning information is updated, the prediction result is compared with the actual data. If there is a deviation, the parameter weights, model parameters, and algorithms are adjusted according to the deviation to correct the prediction. At the same time, the new positioning data is fed back to the big data storage system for further data analysis and model training to continuously improve the accuracy and adaptability of the system.

4. The passive tracking and positioning method according to claim 3, characterized in that: In S400, the correction process is also included: S410, Deviation Analysis: S411, Multi-dimensional deviation calculation and quantification: In addition to the Euclidean distance d error 、Direction angle deviation θ error and path selection bias Path diff In addition, the time deviation analysis is added to calculate the difference between the predicted arrival time at a certain location and the actual arrival time t error ,Considering the influence of different travel modes and path length on time, the time deviation is normalized; For the distance deviation d error ,In complex environments, the distance calculation in three dimensions is considered, and the distance calculation method is adjusted according to the ,passability of the path; Direction angle deviation θ error The calculation should be combined with the path direction and geographic coordinate system. For complex intersections and curves, the deviation should be accurately calculated based on the angle between the path design angle and the user's actual moving direction. Path selection deviation diff The evaluation should consider the similarity and substitution relationship of the paths. If the predicted path and the actual path are within a certain distance and have similar destinations, the deviation weight is relatively low; if the two paths are completely different and the destinations are very different, the deviation weight is high; S412, Deviation weight setting and comprehensive deviation index calculation: Deviation weights are set according to different scenarios and application requirements. In scenarios with high requirements for location accuracy, the weights of distance deviation and direction angle deviation are higher. In scenarios that focus on travel efficiency, the weights of time deviation and path selection deviation are relatively increased. The specific values ​​of the weights are determined through a large number of simulation experiments and actual user data feedback. The calculation of the comprehensive deviation indicator ErroeInder adopts a weighted summation method.

5. The passive tracking and positioning method according to any one of claims 1 to 4, characterized in that: Also included in the S100: S110, walking speed (V) parameter collection and processing: The position data is obtained by timing the positioning of the electronic tag by the reading device or the position of the reading device that reads the electronic tag is directly used as the position data of the electronic tag; assuming that the position coordinates of the electronic tag at time t1 are (x1, y1, z1), and the coordinates at time t2 are (x2, y2, z2), then the position change is: Time interval △t=t2-t1, walking speed Among them, x, y, and z represent the left-right direction, front-back direction, and up-down direction respectively. The front of the coordinate is the direction of the user's walking, the back direction is the direction directly behind the user, the origin is the reading device, the left-right direction is the direction of the line connecting the left and right hands of the user, and the up-down direction is the direction of the user's height. If the three parameters x, y, and z cannot be obtained, the length between two positioning information is used as △d; S120, user pass times (N1) parameter collection and processing: The reading device at the key path node records the passing of the electronic tag. Each time the electronic tag is detected, the device records the passing time, the electronic tag identification, and its own location or vehicle information. Through statistical analysis of these records, the number of times the user passes through different path nodes and vehicles is obtained; S130, parameter collection and processing of the number of people passing through the path and location N2: Obtain the number of people passing through each reading device location and the number of people passing through each reading device path, especially marking the paths and reading devices that the same person passes through multiple times; for points with a difference rate within the suspicious range, use time series analysis methods for further observation; S140, path relationship (R) parameter construction and update: S141, Initial Construction: Collect a variety of geographic information data, including high-precision map data, traffic network data, and three-dimensional terrain data; integrate these data through geographic information system technology to build a path relationship model; in the model, the path is represented by a graph structure, with nodes representing important geographical locations and edges representing roads, channels or traffic routes connecting nodes; S142, Dynamic Update: Establish a data sharing mechanism with traffic management departments, municipal construction departments, and bus and subway operating companies to obtain real-time information on road construction, traffic control, new line openings, and station changes, and update the path relationship model in a timely manner; at the same time, optimize the path relationship based on the user's actual travel data; S150, obtaining the user's travel purpose (P); S160: Obtain the length (L) of the possible traversable path.

6. The passive tracking and positioning method according to claim 5, characterized in that: S110 also includes: calculating the user's travel direction and path based on the user's successive positioning information, combined with the last valid positioning information and the possible path of this positioning information; based on the speed direction, in the map data of the geographic information system, taking the current position as the starting point, a sector-shaped area is delineated along the speed direction, and only the position points within this sector-shaped area are retained in the possible position set S.

7. The passive tracking and positioning method according to claim 5, characterized in that: S120 also includes: User pass times (N1) optimization: S121, Passage Count Statistical Model In addition to simple counting, a more complex statistical model of the number of passes is established; considering the time factor, a day is divided into different time periods, and the user's travel mode may be different in each time period; for each time period, a Markov chain-based model is established, in which the state is a different key path node or area, and the transition probability represents the possibility of the user transferring from one state to another; this model is trained through a large number of users' historical travel data to obtain the transition probability matrix P between different path nodes in each time period ij (t), where i and j are different path nodes and t is the time period; S122, abnormal passage times detection: When calculating the difference rate △N1 between the current number of passes and the number of passes predicted by the model, the judgment of abnormal situations is not based only on a fixed threshold; a dynamic threshold method is used to adjust the threshold according to the time period, day of the week, and the stability of recent user behavior; S123, suspicious point processing: For points whose difference rate is within the suspicious range, the time series analysis method is used for further observation; if the subsequent actual number of passes continues to deviate from the model prediction value, and the deviation becomes larger and larger, the location points on the path where these suspicious points are located are removed from S.

8. The passive tracking and positioning method according to claim 5, characterized in that: S130 also includes the optimization of the number of people on the path (N2): Crowd flow prediction model: Use historical crowd flow data, time information, weather data, and surrounding activity information to build a crowd flow prediction model; use a deep learning model to take weather data and surrounding activity information as additional input features to improve the model's prediction accuracy; Abnormal flow judgment: When calculating the difference rate △N2, the dynamic threshold method is also used; the threshold is adjusted according to the type of path, time period and weather conditions; Optimization based on the trend of passenger flow: In addition to the current passenger flow difference, the trend of passenger flow is analyzed. If the user's traffic behavior does not match this trend, the location point on the path is removed from S.

9. The passive tracking and positioning method according to claim 5, characterized in that: S150 also includes the inference and processing of the purpose (P) parameter: S151, Data Collection: Infer the purpose of travel by integrating multi-source data, analyzing the residence time and frequency of electronic tags in different types of places, combining the boarding and alighting locations of buses and subways and payment information; and the purpose of travel input into the system; S152, inference model construction: Use machine learning algorithms to build a travel purpose inference model. Use decision trees, support vector machines, and neural network algorithms to train the model using the collected user behavior data as input features and travel purpose as output categories. S153, Model Update: As user behavior changes, new data is constantly accumulated, and the traffic purpose inference model is retrained and updated regularly.

10. The passive tracking and positioning method according to claim 5, characterized in that: S160 also includes path length (L) parameter calculation and calibration: S161, initial calculation: For paths such as roads and passages, their lengths are calculated based on the scale and coordinate information in the geographic information data; For straight paths, the Euclidean distance formula is used for calculation; for curved paths, the numerical integration method is used to divide the path into multiple small segments for approximate calculation; for traffic routes, the official mileage data provided by the operating company is used for determination, while fine-tuning is performed by considering factors such as station spacing and line turning radius; S162, calibration mechanism: The path length is calibrated through the user's actual travel time and walking speed data, and a path length error model is established. This model takes into account the impact of environmental factors and transportation operation conditions on path length. S163. Accurate refinement using path length (L) S1631, speed dynamic change model: The dynamic change of speed is affected by many factors. A comprehensive dynamic change model of speed is established. In addition to the path type, environmental factors and user historical speed data, the user's physical condition and the information of the items carried are also considered. A fuzzy logic system or a neural network model is used to represent the relationship between the speed V and these factors: V = g((PathType), T env , H env , V history , HealthStatus, HealthStatus), where PathType is the path type, HealthStatus is the user's physical condition, Load is the information of the items carried, V history is the speed that has been measured before, T env , H env are respectively the ambient temperature and humidity measured previously; S1632. Evaluation of actual traffic conditions on the route: Congestion assessment based on multi-source data fusion: Integrate multi-source data to evaluate the actual traffic conditions of the route, especially the congestion situation. In addition to the road congestion information obtained by environmental sensors, it also uses real-time operation data of buses and subways, GPS data of taxis, and user feedback. The data fusion algorithm is used to integrate these different sources of data to obtain more accurate route congestion information. S164. Time range calculation adjustment: According to the speed dynamic change model and the evaluation results of the actual traffic conditions of the path, the time range of passing the path [T min ,T max ]; for T min , considering the speed under the most ideal traffic conditions; for T max , calculated considering the speed under the worst traffic conditions.

Images