An indoor and outdoor adaptive passive internet of things reflection positioning method and system

Abstract

The application provides an indoor and outdoor adaptive passive Internet of Things reflection positioning method, comprising passive tag deployment and data collection, network architecture design, environment correction model, space-time dynamic scene evaluation function and positioning mode dynamic switching. The application also discloses an indoor and outdoor adaptive passive Internet of Things reflection positioning system. The application realizes adaptive discrimination and positioning mode switching of indoor and outdoor environments through deployment of passive Internet of Things temperature and humidity sensing tags and fusion of Beidou / inertial navigation positioning systems (GNSS / SINS). Passive Internet of Things tags acquire energy through environment radio frequency energy collection technology, do not need external power supply, can continuously perceive environment temperature and humidity data and transmit to a positioning terminal. A comprehensive scene discrimination function is constructed in two dimensions of perception information and signal vector, the most suitable positioning method is adopted for different scenes, and finally, seamless indoor and outdoor positioning switching with ultra-low power consumption and high robustness is realized.

Description

Technical Field

[0001] This invention relates to the field of navigation and positioning technology, and in particular to an indoor and outdoor adaptive passive Internet of Things reflective positioning method and system. Background Technology

[0002] In the current era of deep integration between digital industrialization and industrial digitalization, the Internet of Things (IoT) has made significant progress, with its connectivity and coverage continuously expanding. Simultaneously, the demand for location-based services is experiencing explosive growth. Whether in intelligent transportation, logistics management, smart homes, or industrial manufacturing, accurate location information is crucial, becoming a key element driving efficient operation and innovative development across various industries.

[0003] In outdoor environments, the Global Positioning System (GPS) is the primary means of obtaining location information due to its mature technology and wide coverage. However, once indoors, obstacles such as building walls can severely obstruct GPS signals, causing signal attenuation and interference, making it difficult for GPS to accurately identify the location of the target. In indoor settings such as homes, supermarkets, and warehouses, accurate positioning is crucial for applications such as goods management, personnel guidance, and intelligent security. The limitations of GPS have prompted people to actively explore other effective positioning technologies.

[0004] Positioning solutions based on wireless technologies such as ultrasound, infrared, ultra-wideband (UWB), Wi-Fi, and Bluetooth have emerged. However, each of these technologies has its own limitations. For example, while ultrasonic positioning offers high accuracy, it suffers from significant indoor transmission attenuation and requires a large amount of expensive dedicated hardware. Infrared positioning offers high accuracy but is only suitable for line-of-sight and short-range transmission. UWB technology boasts strong anti-interference capabilities and centimeter-level positioning accuracy, but its deployment costs are exorbitant. Wi-Fi positioning is limited by power consumption issues, and Bluetooth positioning suffers from limited coverage.

[0005] In contrast, passive IoT technologies, represented by Radio Frequency Identification (RFID), exhibit unique advantages. Utilizing backscattering for communication, they feature extremely low power consumption, low cost, and ease of deployment. Passive IoT technologies hold immense promise for applications in scenarios requiring location information services for massive numbers of targets, such as asset management, logistics tracking, and industrial manufacturing. By deploying passive tags on objects and using a reader to transmit and receive the tag's backscattered signals, the location and tracking of objects can be achieved, providing strong support for enterprises to improve management efficiency and reduce costs.

[0006] In recent years, the rise of cellular passive IoT has injected new vitality into passive IoT positioning technology. It combines the wide coverage and powerful connectivity of cellular networks with the advantages of passive IoT, further expanding the application scenarios of passive IoT positioning technology. Whether for asset monitoring in remote areas or device tracking in complex environments, cellular passive IoT positioning technology has great development potential, meeting more diverse positioning needs.

[0007] Against this backdrop, the research and development of passive IoT positioning technology has attracted increasing attention. A thorough analysis of the principles, characteristics, and application scenarios of different positioning technologies, and an exploration of their future development trends, is of great significance for promoting innovation in passive IoT positioning technology, meeting the ever-growing demand for location services, and facilitating the digital transformation of various related industries. Summary of the Invention

[0008] To address the aforementioned issues, this invention aims to propose an indoor-outdoor adaptive passive IoT reflective positioning method and system. By fusing passive IoT tags with GNSS / Inertial Navigation (SINS) and combining environmental perception with dynamic evaluation functions, seamless positioning switching can be achieved indoors, outdoors, and in transitional areas.

[0009] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0010] An indoor / outdoor adaptive passive IoT reflective positioning method includes the following steps:

[0011] S1. Passive Tag Deployment and Data Acquisition: Deploy passive IoT tags in indoor and outdoor areas. The tags are powered by radio frequency energy harvesting technology and periodically backscatter modulated signals containing unique identifiers and environmental sensing data. The positioning terminal captures the backscattered signals of the passive tags through a radio frequency receiver, and analyzes the tag ID, temperature and humidity, air pressure, received signal strength RSSI, and tag response time—unique characteristics of passive IoT—to form a positioning information database.

[0012] S2. Network Architecture Design: GNSS augmentation base stations are deployed synchronously in outdoor areas to improve satellite positioning accuracy; satellite signals are acquired through GNSS modules to calculate three-dimensional position coordinates and position accuracy factor PDOP; accelerometer and gyroscope data are collected through SINS to calculate the terminal's motion trajectory.

[0013] S3. Environmental Correction Model: Considering the correction model of environmental parameters on the refractive index of electromagnetic waves, establish formulas for the influence of temperature, humidity and air pressure on the signal strength of passive IoT tag signal reception.

[0014] S4. Spatiotemporal dynamic scene evaluation function: Construct a comprehensive discrimination function, integrate signal strength, environmental parameters and motion state, and output scene evaluation function F∈[0,1] to distinguish indoor areas, outdoor areas and transition areas;

[0015] S5. Dynamic switching of positioning mode: The positioning mode is dynamically switched according to the evaluation function F.

[0016] In indoor environments, the F-value approaches 0: localization is achieved using the WKNN algorithm based on modified RSSI fingerprints and tag response times;

[0017] Transitional region, F is the intermediate value: adaptive fusion positioning of BeiDou / SINS and tag data;

[0018] In outdoor environments, the F-value approaches 1: BeiDou / SINS fusion anti-interference dynamic positioning.

[0019] Furthermore, the positioning terminal integrates a GNSS module, a SINS inertial navigation system, a radio frequency radiation source, and a multi-band radio frequency receiver. The overall positioning network structure in the network architecture design includes:

[0020] Passive IoT tag access layer: Passive IoT tags are densely deployed in indoor areas and indoor-outdoor transition areas. Each tag uploads a modulated signal containing a unique identifier, preset anchor point coordinates, and environmental sensing data such as temperature, humidity, and air pressure through backscatter communication technology.

[0021] Terminal layer: The positioning terminal integrates a GNSS receiver module, an inertial navigation module (SINS), and a multi-band passive IoT reader module to achieve signal reception, data fusion, and positioning calculation; all positioning modes: outdoor GNSS / SINS, transition area fusion positioning, and indoor passive tag positioning calculations are all completed independently by the terminal;

[0022] Auxiliary enhancement layer: Deploy GNSS enhancement base stations in outdoor areas to improve satellite positioning accuracy.

[0023] Furthermore, in step S1, the passive IoT tag is powered by environmental radio frequency energy harvesting technology, and its energy harvesting efficiency satisfies:

[0024]

[0025] Among them, P harvest η is the energy harvesting power; η is the tag energy conversion efficiency; P RF This refers to the reader's transmit power; G tag λ is the tag antenna gain, which is affected by the antenna directivity; λ is the radio frequency wavelength; d is the distance between the tag and the positioning terminal.

[0026] The lifespan of a passive IoT tag without external power is:

[0027]

[0028] T life For tag lifespan; E storage For energy storage capacity; P tagsleep E represents the power consumption of the tag in sleep mode. harvest P represents the energy collected by the tag per unit time. tagactive This represents the power consumption of the tag when it is active.

[0029] Furthermore, in step S1, the absorption model formula for passive IoT signals by water vapor is as follows:

[0030]

[0031] Where, γ water ρ is the water vapor absorption coefficient; f is the water vapor density; and f is the radio frequency. T represents temperature;

[0032] The effects of temperature and air pressure on electromagnetic waves are similar in principle; both cause changes in air density, which in turn affect the refractive index of the medium, leading to bending of the electromagnetic wave path. The formula for this effect is:

[0033]

[0034] Where N is the corrected refractive index; T is the temperature; P is the atmospheric pressure; and e is the water vapor pressure.

[0035] Constructing the environmental impact factor function:

[0036] n=n0+α·γ water +(1-α)·(N-N0)

[0037] Where n0 is the basic attenuation factor; α is the humidity weighting coefficient, which is experimentally calibrated; and N0 is the standard atmospheric corrected refractive index.

[0038] The final mapping function between the received signal strength RSSI(d) and the distance d between the tag and the terminal device is:

[0039] RSSI(d) = -10n·log 10 (d)+C+X σ

[0040] Where n is the environmental degradation factor; C is the environmental reference constant; X σ It is random noise that follows a normal distribution.

[0041] Furthermore, in step S1, the tag response time of the passive tag includes the tag charging time, circuit startup time, data processing time, and backscatter modulation time. The circuit startup time is the time for the internal circuit to start and initialize after the capacitor voltage reaches the target, typically in the microsecond range. Since the passive IoT tag uses a low-power communication protocol, its data processing time is also typically in the microsecond range. The backscatter modulation time is typically in the nanosecond range. Therefore, the response time T of the passive tag... resp Tag charging time should be given priority, i.e.:

[0042] T resp ≈t harvest

[0043] Passive IoT tag charging time t harvest The formula can be expressed as:

[0044]

[0045] Among them, C tag For the tag capacitor, V threshold P is the tag activation threshold voltage. rf For the received power, under the premise of using passive tags of the same standard, the effects of tag capacitance, threshold voltage and conversion efficiency can be regarded as the same. However, the received power of the receiving antenna is affected by the channel state and gain direction factors and is different. Therefore, the difference in tag response time reflects the tag spatial characteristics to a certain extent, and is included in the fingerprint database as the radio frequency fingerprint feature.

[0046] Based on the effects of channel conditions and gain directivity, the antenna received power is adjusted as follows:

[0047]

[0048] Among them, P RF This refers to the reader's transmit power; G tx With G tag denoted by , respectively, are the gains of the transmitting antenna and the tag receiving antenna, which are affected by antenna directivity; d is the distance between the tag and the positioning terminal; and n is an environmental impact factor function.

[0049] Furthermore, in step S4, it is necessary to construct a spatiotemporal dynamic scene evaluation function that integrates perceived information and signal vectors. The evaluation function involves two types of data: one is the dynamic change of perceived information in the spatiotemporal dimension; the other is the dynamic change of signal vectors in the spatiotemporal dimension. Therefore, relevant influencing factor functions are constructed for information at a single time and a single location.

[0050] Based on sensor data such as temperature, humidity, and air pressure, an environmental influencing factor function is constructed:

[0051] n=n0+α·γwater +(1-α)·(N-N0)

[0052] Where n0 is the basic attenuation factor; α is the humidity weighting coefficient, which is determined experimentally; and N0 is the standard atmospheric corrected refractive index.

[0053] Based on signal strength RSSI and tag density, construct a signal vector factor function:

[0054]

[0055] Where, N tags N represents the number of tags received. max Maximum number of tags; RSSI avg The average signal strength; RSSI max The maximum RSSI value in the sample set; RSSI min The minimum RSSI value in the sample set; after normalization, S∈[0,1];

[0056] Spatiotemporal dynamic factor function based on continuous dynamic changes in tag time and space:

[0057]

[0058] Where v is the terminal's moving speed; v max The maximum terminal mobility rate in the sample set; Δn is the rate of change of environmental factors over time; Δt is the amount of change of environmental factors; δ is the weighting coefficient.

[0059] By integrating the above influencing factor functions, the final comprehensive evaluation function for positioning is determined as follows:

[0060]

[0061] Wherein, β1, β2, and β3 are weighting coefficients, which need to be calibrated experimentally; k is the slope parameter; n is the environmental influence factor function; S is the signal vector factor function; and D is the spatiotemporal dynamic factor function.

[0062] Based on the different values ​​of F, the changes in the positioning scene can be determined: F value close to 0: indoor environment; F value in the middle: transition area; F value close to 1: outdoor environment.

[0063] Furthermore, in step S5, when performing localization using the WKNN algorithm based on the corrected RSSI fingerprint and tag response time, indoor fingerprint localization is divided into the following two stages:

[0064] The first stage is the offline stage, which involves building a fingerprint database based on the corrected RSSI data, environmental perception parameter data, GNSS system data, and inertial navigation data.

[0065] The second stage is the online matching stage. First, it is necessary to collect current RSSI, label response time, and environmental parameter data. Then, it is necessary to fuse short-term motion data from SINS and optimize the WKNN weights.

[0066]

[0067] RSSI corrected This represents the corrected real-time signal strength value currently acquired; RSSI i Δt is the signal strength value of the i-th reference point in the fingerprint database; v is the terminal velocity vector, provided by SINS; Δt is the time interval; Δp i Candidate reference point (x) i ,y i ,z i The displacement difference between the positioning coordinates and the previous time step; φ is the attenuation coefficient, which controls the influence of motion consistency on the weight, and is determined experimentally; T resp The response time of the currently collected passive tag; Let be the response time of the passive label at the i-th reference point;

[0068] Subsequently, Top-K reference points are selected based on the dynamic K-value, which is driven by a combination of label density and environmental awareness factors. The formula is as follows:

[0069]

[0070] Among them, K base The base K value; σ RSSI σ is the current sample standard deviation of RSSI; max The preset maximum fluctuation threshold; N tags N represents the number of tags received. max ω is the maximum number of tag densities. n Δn is the environmental weighting coefficient; n is the environmental degradation factor; Δn is the change in environmental factors; n max ω represents the maximum permissible environmental factor deviation. n The environmental weighting coefficient, ω, is determined experimentally. n The larger the value of ω, the more sensitive the K value is to environmental fluctuations, and the more reference points will be selected to improve accuracy, making it more suitable for indoor environments with frequent fluctuations in temperature, humidity, and air pressure; conversely, the smaller the value of ω, the more sensitive the K value is to environmental fluctuations. n The smaller the value, the less sensitive the K value is to environmental fluctuations, and the more suitable it is for stable environments;

[0071] Finally, the weights w are combined. i The weighted average yields the relative positioning coordinates (x). tag ,y tag ,z tag ):

[0072]

[0073] Where w i Let x be the fusion weight of the i-th reference point; i ,y i ,z i (i) represents the positioning coordinates of the i-th reference point in the local coordinate system based on the reference point.

[0074] This coordinate system is based on a local coordinate system with a reference point; therefore, anchor point labels are required. The absolute coordinates are used to convert the relative position to global absolute coordinates:

[0075]

[0076] in, It is the relative position between the i-th reference point and the anchor label, determined by the anchor label's positioning coordinates (x, y) in the reference point coordinate system. m ,y m ,z m To decide:

[0077]

[0078] Therefore, the absolute coordinates of the label The formula is modified as follows:

[0079]

[0080] in, The reference point's absolute coordinates are corrected; anchor tags are deployed in critical locations, and passive IoT tags with known absolute coordinates are selected as anchor points.

[0081] Furthermore, in step S5, the adaptive fusion localization in the transition region is divided into the following two stages:

[0082] The first phase dynamically determines scene bias based on the comprehensive evaluation function F, assigns weights to multi-source data, and confirms the GNSS / SINS fusion anti-interference dynamic positioning weights:

[0083]

[0084] Where F is the comprehensive evaluation function; Q GNSS Score the quality of the BeiDou signal; Q max σ represents the theoretical maximum value of BeiDou signal quality. GNSS σ represents the standard deviation of GNSS position. max The preset GNSS error threshold;

[0085] By combining signal strength RSSI and tag density, tag fingerprint localization weights are defined:

[0086]

[0087] Where, N tags N represents the number of tags received. max Maximum number of tags; RSSI avg The average signal strength; RSSI max The maximum RSSI value in the sample set; RSSI min The minimum RSSI value in the sample set;

[0088] The second stage involves normalizing the weights:

[0089]

[0090] Among them, w GNSS For GNSS / SINS fusion anti-interference dynamic positioning weights; w tag Weights are assigned to the tag fingerprint for location;

[0091] Output the final positioning coordinates:

[0092]

[0093] Among them, w′ G / S For normalized GNSS / SINS fusion anti-interference dynamic positioning weights; w′ tag To normalize the label fingerprint location weights; (x G / S ,y G / S ,z G / S (This refers to the GNSS / SINS fusion anti-interference dynamic positioning coordinate results.) The coordinates of the tag fingerprint are obtained.

[0094] Furthermore, in step S5, the BeiDou / SINS fusion anti-interference dynamic positioning in the outdoor environment is divided into the following three stages:

[0095] The first phase involves receiving BeiDou signals and dynamically evaluating the quality of those signals.

[0096]

[0097] Where, N vis N represents the number of visible satellites. max This is the maximum number of visible satellites, with a default value of 12. Mean signal-to-noise ratio; SNR max The preset maximum signal-to-noise ratio; D PDOP β is the position accuracy factor. posThe suppression coefficient of the position accuracy factor is determined experimentally; This represents the average elevation angle of the satellite. It is the three-dimensional position variance of the receiver in the ECEF coordinate system;

[0098] The second stage involves establishing an adaptive weighted fusion model. First, error estimation is performed on the inertial navigation system. The error estimation formula is as follows:

[0099]

[0100] Where, ν SINS For SINS calculation speed; ν GNSS To improve the BeiDou Doppler resolution speed; The zero bias standard deviation of the accelerometer; α is the standard deviation of gyroscope drift. SINS These are dynamic weighting coefficients, calibrated according to the specific instrument's accuracy.

[0101] Subsequently, an adaptive weighted fusion model for BeiDou / SINS was established:

[0102]

[0103] Q GNSS Score the quality of the BeiDou signal; E SINS For inertial navigation error estimation; γ S This is the error suppression coefficient;

[0104] The third stage outputs the final 3D positioning coordinates:

[0105] (x G / S ,y G / S ,z G / S ) = w GNSS ·(x GNSS ,y GNSS ,z GNSS )+w SINS ·(x SINS ,y SINS ,z SINS )

[0106] Among them, w GNSS w represents the weighting coefficient for the BeiDou signal. SINS For the inertial navigation system weighting coefficients; (x GNSS ,y GNSS ,z GNSS (x) represents the three-dimensional coordinates of the BeiDou signal calculated by the terminal; SINS ,y SINS ,z SINS ) represents the three-dimensional coordinates of the inertial navigation system calculated by the terminal.

[0107] To achieve the above objectives, the present invention also provides an indoor / outdoor adaptive passive IoT reflective positioning system, comprising the following modules:

[0108] Passive tag deployment and data acquisition module: used to deploy passive IoT tags in indoor and outdoor areas. The tags are powered by radio frequency energy harvesting technology and periodically backscatter modulated signals containing unique identifiers and environmental sensing data. The positioning terminal captures the backscattered signals of the passive tags through radio frequency receivers, and analyzes the tag ID, temperature and humidity, air pressure, received signal strength RSSI, and tag response time—unique characteristics of passive IoT—to form a positioning information database.

[0109] Network architecture module: Improves satellite positioning accuracy by synchronously deploying GNSS augmentation base stations in outdoor areas; acquires satellite signals through the GNSS module to calculate three-dimensional position coordinates and position accuracy factor (PDOP); and collects accelerometer and gyroscope data through SINS to calculate the terminal's motion trajectory.

[0110] Environmental correction module: Considering the correction model of environmental parameters on the refractive index of electromagnetic waves, establish formulas for the influence of temperature, humidity and air pressure on the signal strength of passive IoT tag signal reception;

[0111] Spatiotemporal dynamic scene evaluation module: Construct a comprehensive discrimination function, integrate signal strength, environmental parameters and motion state, and output a scene evaluation function F∈[0,1] to distinguish indoor areas, outdoor areas and transition areas;

[0112] Dynamic positioning mode switching module: Dynamically switches positioning modes based on the evaluation function F.

[0113] In indoor environments, the F-value approaches 0: localization is achieved using the WKNN algorithm based on modified RSSI fingerprints and tag response times;

[0114] Transitional region, F is the intermediate value: adaptive fusion positioning of BeiDou / SINS and tag data;

[0115] In outdoor environments, the F-value approaches 1: BeiDou / SINS fusion anti-interference dynamic positioning.

[0116] Beneficial Effects: This invention fully considers the characteristics of passive IoT—simple deployment and strong adaptability—eliminating the need for complex wiring or external power supplies, thus reducing deployment costs and difficulties, greatly improving deployment efficiency, and enabling rapid application in various scenarios. This invention also considers the characteristics of temperature and humidity sensor data, taking into account the influence of temperature, humidity, and air pressure on electromagnetic radiation in the backscattering compensation model of passive IoT tags. Based on multiple dimensions of the spatiotemporal dynamic changes of sensing information and signal vectors, a unified discrimination function for continuous indoor, outdoor, and indoor-outdoor transition areas is constructed. Furthermore, corresponding anti-interference positioning algorithms are designed for different scenarios, ensuring high accuracy while effectively reducing environmental interference, making it suitable for complex and ever-changing environments. Attached Figure Description

[0117] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0118] Figure 1 This is a flowchart of the main body of the indoor and outdoor adaptive passive IoT reflective positioning method according to an embodiment of the present invention;

[0119] Figure 2 This is a schematic diagram of the backscatter communication principle in the indoor and outdoor adaptive passive IoT reflective positioning method described in this embodiment of the invention;

[0120] Figure 3 This is a flowchart of the fingerprint positioning process in the indoor and outdoor adaptive passive IoT reflective positioning method described in this embodiment of the invention.

[0121] Figure 4 This is a diagram showing the positioning results of the backend platform in the indoor / outdoor adaptive passive IoT reflective positioning method described in this embodiment of the invention.

[0122] Figure 5 This is a schematic diagram of the indoor and outdoor adaptive passive Internet of Things reflective positioning system according to an embodiment of the present invention. Detailed Implementation

[0123] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0124] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0125] Example 1

[0126] See Figure 1-4 An indoor / outdoor adaptive passive IoT reflective positioning method includes the following steps:

[0127] S1. Passive Tag Deployment and Data Acquisition: Deploy passive IoT tags in indoor and outdoor areas. The tags are powered by radio frequency energy harvesting technology and periodically backscatter modulated signals containing unique identifiers and environmental sensing data. The positioning terminal captures the backscattered signals of the passive tags through a radio frequency receiver, and analyzes the tag ID, temperature and humidity, air pressure, received signal strength RSSI, and tag response time—unique characteristics of passive IoT—to form a positioning information database.

[0128] It should be noted that, as Figure 2 As shown, when powered by an external carrier transmitter, the tag can periodically backscatter a modulated signal containing a unique identifier and environmental sensing data. The signal is demodulated and decoded at the reflection receiver to read the tag's temperature and humidity sensing information.

[0129] S2. Network Architecture Design: GNSS augmentation base stations are deployed synchronously in outdoor areas to improve satellite positioning accuracy; satellite signals are acquired through GNSS modules to calculate three-dimensional position coordinates and position accuracy factor PDOP; accelerometer and gyroscope data are collected through SINS to calculate the terminal's motion trajectory.

[0130] S3. Environmental Correction Model: Considering the correction model of environmental parameters on the refractive index of electromagnetic waves, establish formulas for the influence of temperature, humidity and air pressure on the signal strength of passive IoT tag signal reception.

[0131] S4. Spatiotemporal dynamic scene evaluation function: Construct a comprehensive discrimination function, integrate signal strength, environmental parameters and motion state, and output scene evaluation function F∈[0,1] to distinguish indoor areas, outdoor areas and transition areas;

[0132] S5. Dynamic switching of positioning mode: The positioning mode is dynamically switched according to the evaluation function F.

[0133] In indoor environments, the F-value approaches 0: localization is achieved using the WKNN algorithm based on modified RSSI fingerprints and tag response times;

[0134] Transitional region, F is the intermediate value: adaptive fusion positioning of BeiDou / SINS and tag data;

[0135] In outdoor environments, the F-value approaches 1: BeiDou / SINS fusion anti-interference dynamic positioning.

[0136] This embodiment has the following advantages:

[0137] Environmental adaptive positioning: By using an environmental parameter correction model, the influence of temperature, humidity and air pressure on electromagnetic wave propagation is effectively compensated, thereby improving the positioning robustness in complex environments.

[0138] Seamless multi-mode switching: Intelligent discrimination of indoor / transitional / outdoor scenes is achieved based on dynamic evaluation function, and WKNN, GNSS / SINS and tag data fusion is combined to ensure positioning continuity.

[0139] The low power consumption advantage of passive tags: By utilizing backscatter communication technology, the tag power supply requirement is eliminated, significantly reducing deployment and maintenance costs.

[0140] In a specific example, the positioning terminal integrates a GNSS module, a SINS inertial navigation system, a radio frequency radiation source, and a multi-band radio frequency receiver. The overall positioning network structure in the network architecture design includes:

[0141] Passive IoT tag access layer: Passive IoT tags are densely deployed in indoor areas and indoor-outdoor transition areas. Each tag uploads a modulated signal containing a unique identifier, preset anchor point coordinates, and environmental sensing data such as temperature, humidity, and air pressure through backscatter communication technology.

[0142] Terminal layer: The positioning terminal integrates a GNSS receiver module, an inertial navigation module (SINS), and a multi-band passive IoT reader module to achieve signal reception, data fusion, and positioning calculation; all positioning modes: outdoor GNSS / SINS, transition area fusion positioning, and indoor passive tag positioning calculations are all completed independently by the terminal;

[0143] Auxiliary enhancement layer: Deploy GNSS enhancement base stations in outdoor areas to improve satellite positioning accuracy.

[0144] Terminal-level autonomous computing: Integrating GNSS / SINS and multi-band readers and writers to achieve terminal-level data fusion and positioning calculation, avoiding cloud dependence and improving real-time performance and privacy.

[0145] Layered architecture scalability: Through a three-tiered architecture of access layer (tag), terminal layer (computing), and auxiliary layer (base station), it supports large-scale tag deployment and high-concurrency positioning requests.

[0146] In a specific example, in step S1, because passive IoT tags need to have advantages in energy harvesting and tag lifespan, be able to be deployed in harsh, demanding environments where battery replacement is difficult, and maintain low costs, passive IoT tags are powered by environmental radio frequency energy harvesting technology, and their energy harvesting efficiency meets the following requirements:

[0147]

[0148] Among them, P harvest η is the energy harvesting power; η is the tag energy conversion efficiency; P RF This refers to the reader's transmit power; G tagλ is the tag antenna gain, which is affected by the antenna directivity; λ is the radio frequency wavelength; d is the distance between the tag and the positioning terminal.

[0149] The lifespan of a passive IoT tag without external power is:

[0150]

[0151] T life For tag lifespan; E storage For energy storage capacity; P tagsleep E represents the power consumption of the tag in sleep mode. harvest P represents the energy collected by the tag per unit time. tagactive This represents the power consumption of the tag when it is active.

[0152] Energy model precision: Formulas describe the energy harvesting efficiency and lifespan of tags, providing a quantitative basis for tag deployment density and wake-up cycle configuration, and optimizing system endurance.

[0153] Directional gain compensation: By correcting the received power model through antenna gain parameters, the accuracy of energy harvesting prediction is improved, and the performance fluctuation caused by tag orientation differences is reduced.

[0154] In a specific example, in step S1, since the signal strength received by the passive tag cannot be directly obtained, it is necessary to use the distance between the tag and the terminal device for mapping and to consider the influence of many environmental factors on the backscattered signal:

[0155] The water molecule content in the air, i.e., air humidity, has a significant absorption effect on electromagnetic waves, leading to attenuation of reflected signals from passive IoT tags. The absorption model formula for passive IoT signals by water vapor is as follows:

[0156]

[0157] Where, γ water ρ is the water vapor absorption coefficient; f is the water vapor density; and f is the radio frequency. T represents temperature;

[0158] The effects of temperature and air pressure on electromagnetic waves are similar in principle; both cause changes in air density, which in turn affect the refractive index of the medium, leading to bending of the electromagnetic wave path. The formula for this effect is:

[0159]

[0160] Where N is the corrected refractive index; T is the temperature; P is the atmospheric pressure; and e is the water vapor pressure.

[0161] Constructing the environmental impact factor function:

[0162] n=n0+α·γwater +(1-α)·(N-N0)

[0163] Where n0 is the basic attenuation factor; α is the humidity weighting coefficient, which is experimentally calibrated; and N0 is the standard atmospheric corrected refractive index.

[0164] The final mapping function between the received signal strength RSSI(d) and the distance d between the tag and the terminal device is:

[0165] RSSI(d) = -10n·log 10 (d)+C+X σ

[0166] Where n is the environmental attenuation factor; C is the environmental reference constant, which is hardware-related such as antenna gain; X σ It is random noise that follows a normal distribution.

[0167] Dynamic refractive index correction: The refractive index correction formula is introduced by combining temperature, humidity, air pressure and water vapor pressure to accurately quantify the influence of the environment on the curvature of the electromagnetic wave path and improve the accuracy of RSSI ranging.

[0168] Noise suppression design: Add a normally distributed noise term to the RSSI-distance mapping function to enhance the algorithm's robustness to random disturbances.

[0169] In a specific example, in step S1, the tag response time of the passive tag includes the tag charging time, circuit startup time, data processing time, and backscatter modulation time. The circuit startup time is the time for the internal circuit to start and initialize after the capacitor voltage reaches the target, typically in the microsecond range. Passive IoT tags use low-power communication protocols, and their data volume and protocol complexity are not high; therefore, the data processing time is also typically in the microsecond range, and the backscatter modulation time is typically in the nanosecond range. Therefore, the response time T of the passive tag... resp Tag charging time should be given priority, i.e.:

[0170] T resp ≈t harvest

[0171] Passive IoT tag charging time t harvest The formula can be expressed as:

[0172]

[0173] Among them, C tag For the tag capacitor, V threshold P is the tag activation threshold voltage. rfFor the received power, under the premise of using passive tags of the same standard, the effects of tag capacitance, threshold voltage and conversion efficiency can be regarded as the same. However, the received power of the receiving antenna is affected by the channel state and gain direction factors and is different. Therefore, the difference in tag response time reflects the tag spatial characteristics to a certain extent, and is included in the fingerprint database as the radio frequency fingerprint feature.

[0174] Based on the effects of channel conditions and gain directivity, the antenna received power is adjusted as follows:

[0175]

[0176] Among them, P RF This refers to the reader's transmit power; G tx With G tag denoted by , respectively, are the gains of the transmitting antenna and the tag receiving antenna, which are affected by antenna directivity; d is the distance between the tag and the positioning terminal; and n is an environmental impact factor function.

[0177] Response time characterization: The difference in charging time is transformed into spatial fingerprint features, increasing the dimension of the fingerprint database and solving the problem of fuzzy label recognition under dense deployment.

[0178] Innovative Directional Modeling: Quantifying the impact of antenna directivity on received power, overcoming the shortcomings of traditional path loss models that only consider distance.

[0179] In a specific example, in step S4, it is necessary to construct a spatiotemporal dynamic scene evaluation function that integrates sensing information and signal vectors. The evaluation function involves two types of data: one is the dynamic change of sensing information in the spatiotemporal dimension; the other is the dynamic change of signal vectors in the spatiotemporal dimension. Therefore, relevant influence factor functions are constructed for information at a single time and a single location.

[0180] Based on sensor data such as temperature, humidity, and air pressure, an environmental influencing factor function is constructed:

[0181] n=n0+α·γ water +(1-α)·(N-N0)

[0182] Where n0 is the basic attenuation factor; α is the humidity weighting coefficient, which is determined experimentally; and N0 is the standard atmospheric corrected refractive index.

[0183] Based on signal strength RSSI and tag density, construct a signal vector factor function:

[0184]

[0185] Where, N tags N represents the number of tags received. max Maximum number of tags; RSSI avgThe average signal strength; RSSI max The maximum RSSI value in the sample set; RSSI min The minimum RSSI value in the sample set; after normalization, S∈[0,1];

[0186] Spatiotemporal dynamic factor function based on continuous dynamic changes in tag time and space:

[0187]

[0188] Where v is the terminal's moving speed; v max The maximum terminal mobility rate in the sample set; Δn is the rate of change of environmental factors over time; Δt is the amount of change of environmental factors; δ is the weighting coefficient.

[0189] By integrating the above influencing factor functions, the final comprehensive evaluation function for positioning is determined as follows:

[0190]

[0191] Wherein, β1, β2, and β3 are weighting coefficients, which need to be calibrated experimentally; k is the slope parameter; n is the environmental influence factor function; S is the signal vector factor function; and D is the spatiotemporal dynamic factor function.

[0192] Based on the different values ​​of F, the changes in the positioning scene can be determined: F value close to 0: indoor environment; F value in the middle: transition area; F value close to 1: outdoor environment.

[0193] Multi-factor dynamic fusion: The evaluation function is constructed by integrating environmental factors, signal strength, and spatiotemporal change rate to achieve multi-dimensional perception of scene discrimination and reduce the misjudgment rate.

[0194] Parameter configurability: By calibrating the weight coefficients through experiments, the algorithm can be optimized and adapted for different building structures or climate conditions.

[0195] In a specific example, during step S5, when the WKNN algorithm based on the modified RSSI fingerprint and tag response time is used for localization, the indoor fingerprint localization is divided into the following two stages:

[0196] The first stage is the offline stage, which involves building a fingerprint database based on the corrected RSSI data, environmental perception parameter data, GNSS system data, and inertial navigation data.

[0197] The second stage is the online matching stage. First, it is necessary to collect current RSSI, label response time, and environmental parameter data. Then, it is necessary to fuse short-term motion data from SINS and optimize the WKNN weights.

[0198]

[0199] RSSI corrected This represents the corrected real-time signal strength value currently acquired; RSSI i Δt is the signal strength value of the i-th reference point in the fingerprint database; v is the terminal velocity vector, provided by SINS; Δt is the time interval; Δp i Candidate reference point (x) i ,y i ,z i The displacement difference between the positioning coordinates and the previous time step; φ is the attenuation coefficient, which controls the influence of motion consistency on the weight, and is determined experimentally; T resp The response time of the currently collected passive tag; Let be the response time of the passive label at the i-th reference point;

[0200] Subsequently, Top-K reference points are selected based on the dynamic K-value, which is driven by a combination of label density and environmental awareness factors. The formula is as follows:

[0201]

[0202] Among them, K base The base K value; σ RSSI σ is the current sample standard deviation of RSSI; max The preset maximum fluctuation threshold; N tags N represents the number of tags received. max ω is the maximum number of tag densities. n Δn is the environmental weighting coefficient; n is the environmental degradation factor; Δn is the change in environmental factors; n max ω represents the maximum permissible environmental factor deviation. n The environmental weighting coefficient, ω, is determined experimentally. n The larger the value of ω, the more sensitive the K value is to environmental fluctuations, and the more reference points will be selected to improve accuracy, making it more suitable for indoor environments with frequent fluctuations in temperature, humidity, and air pressure; conversely, the smaller the value of ω, the more sensitive the K value is to environmental fluctuations. n The smaller the value, the less sensitive the K value is to environmental fluctuations, and the more suitable it is for stable environments, such as indoor environments like warehouses with constant temperature and humidity.

[0203] Compared to selecting label deployment points as center values, dynamically selecting the K value not only reduces the number of centers and computational complexity, making the algorithm more lightweight, but also allows for dynamic adjustment of the K value based on environmental parameters, label density, and other parameters, thereby improving the algorithm's generalization ability.

[0204] Finally, the weights w are combined. i The weighted average yields the relative positioning coordinates (x). tag ,y tag ,z tag ):

[0205]

[0206] Where w i Let x be the fusion weight of the i-th reference point; i ,y i ,z i (i) represents the positioning coordinates of the i-th reference point in the local coordinate system based on the reference point.

[0207] This coordinate system is based on a local coordinate system with a reference point; therefore, anchor point labels are required. The absolute coordinates are used to convert the relative position to global absolute coordinates:

[0208]

[0209] in, It is the relative position between the i-th reference point and the anchor label, determined by the anchor label's positioning coordinates (x, y) in the reference point coordinate system. m ,y m ,z m To decide:

[0210]

[0211] Therefore, the absolute coordinates of the label The formula is modified as follows:

[0212]

[0213] in, The absolute coordinates of the corrected reference point are obtained; finally, the following is obtained: Figure 4 The final positioning coordinate result diagram shown shows that during the deployment process, anchor tags were deployed in relatively critical positions, and passive IoT tags with known absolute coordinates (rectangular coordinates) were selected as anchor points.

[0214] Motion consistency optimization: By introducing SINS velocity vector and displacement difference constraints, path ambiguity in WKNN fingerprint matching is suppressed, which is particularly suitable for pedestrian continuous movement scenarios.

[0215] Dynamic K-value strategy: Adaptively adjust the number of reference points based on tag density and environmental fluctuations to balance positioning accuracy and computational efficiency.

[0216] Local-Global Coordinate Mapping: By transforming the absolute coordinates of anchor point labels, the challenge of connecting the indoor local coordinate system with the global positioning system is solved.

[0217] In a specific example, in step S5, the adaptive fusion localization in the transition region is divided into the following two stages:

[0218] The first phase dynamically determines scene bias based on the comprehensive evaluation function F, assigns weights to multi-source data, and confirms the GNSS / SINS fusion anti-interference dynamic positioning weights:

[0219]

[0220] Where F is the comprehensive evaluation function; Q GNSS Score the quality of the BeiDou signal; Q max σ represents the theoretical maximum value of BeiDou signal quality. GNSS σ represents the standard deviation of GNSS position. max The preset GNSS error threshold;

[0221] By combining signal strength RSSI and tag density, tag fingerprint localization weights are defined:

[0222]

[0223] Where, N tags N represents the number of tags received. max Maximum number of tags; RSSI avg The average signal strength; RSSI max The maximum RSSI value in the sample set; RSSI min The minimum RSSI value in the sample set;

[0224] The second stage involves normalizing the weights:

[0225]

[0226] Among them, w GNSS For GNSS / SINS fusion anti-interference dynamic positioning weights; w tag Weights are assigned to the tag fingerprint for location;

[0227] Output the final positioning coordinates:

[0228]

[0229] Among them, w′ G / S For normalized GNSS / SINS fusion anti-interference dynamic positioning weights; w′ tag To normalize the label fingerprint location weights; (x G / S ,y G / S ,z G / S (This refers to the GNSS / SINS fusion anti-interference dynamic positioning coordinate results.) The coordinates of the tag fingerprint are obtained.

[0230] Dynamic weight allocation: The fusion weights are adjusted in real time based on the quality of the BeiDou signal and the density of the tags to ensure a smooth transition in positioning in the transition area.

[0231] Anti-satellite interference design: Introducing GNSS position standard deviation and error threshold constraints to reduce the impact of multipath effects on the fusion results.

[0232] In a specific example, in step S5, the BeiDou / SINS fusion anti-interference dynamic positioning in the outdoor environment is divided into the following three stages:

[0233] The first stage involves receiving BeiDou signals, such as demodulating pseudorange, carrier phase, Doppler shift, and navigation messages, and dynamically evaluating the quality of the BeiDou signals.

[0234]

[0235] Where, N vis N represents the number of visible satellites. max This is the maximum number of visible satellites, with a default value of 12. Mean signal-to-noise ratio; SNR max The preset maximum signal-to-noise ratio; D PDOP β is the position accuracy factor. pos The suppression coefficient of the position accuracy factor is determined experimentally; This represents the average elevation angle of the satellite. It is the three-dimensional position variance of the receiver in the ECEF coordinate system;

[0236] The second stage involves establishing an adaptive weighted fusion model. First, error estimation is performed on the inertial navigation system. The error estimation formula is as follows:

[0237]

[0238] Where, ν SINS For SINS calculation speed; ν GNSS To improve the BeiDou Doppler resolution speed; The zero bias standard deviation of the accelerometer; α is the standard deviation of gyroscope drift. SINS These are dynamic weighting coefficients, calibrated according to the specific instrument's accuracy.

[0239] Subsequently, an adaptive weighted fusion model for BeiDou / SINS was established:

[0240]

[0241] Q GNSS Score the quality of the BeiDou signal; E SINS For inertial navigation error estimation; γ S This is the error suppression coefficient;

[0242] The third stage outputs the final 3D positioning coordinates:

[0243] (x G / S ,y G / S ,z G / S ) = w GNSS ·(x GNSS ,y GNSS ,z GNSS )+w SINS ·(x SINS ,y SINS ,z SINS )

[0244] Among them, w GNSS w represents the weighting coefficient for the BeiDou signal. SINS For the inertial navigation system weighting coefficients; (x GNSS ,y GNSS ,z GNSS (x) represents the three-dimensional coordinates of the BeiDou signal calculated by the terminal; SINS ,y SINS ,z SINS ) represents the three-dimensional coordinates of the inertial navigation system calculated by the terminal.

[0245] Signal quality quantitative assessment: Construct a BeiDou scoring model with multiple dimensions including satellite number, signal-to-noise ratio, and PDOP to accurately quantify positioning reliability.

[0246] Inertial navigation error suppression: By dynamically estimating the SINS error using the zero bias standard deviation of accelerometers / gyroscopes, the scientific nature of the fusion weight allocation is improved.

[0247] Anti-dynamic interference capability: Utilizing Doppler velocity verification and ECEF variance analysis, it effectively resists positioning jumps caused by rapid vehicle acceleration or obstruction by tall buildings.

[0248] In summary, this embodiment overcomes the bottleneck of traditional positioning systems experiencing a sharp drop in accuracy in transitional indoor and outdoor areas by employing three core technologies: passive tag backscattering, environmental parameter compensation, and multi-source data fusion. Its core innovation lies in:

[0249] Heterogeneous integration of passive IoT and active positioning: For the first time, the low-cost characteristics of passive tags are deeply combined with the high-precision advantages of GNSS / SINS.

[0250] Environmental perception closed-loop correction: Establish a quantifiable environmental-electromagnetic wave propagation model to achieve adaptability in dynamic environments.

[0251] Lightweight terminal computing: All algorithms are executed locally on the terminal, meeting the stringent requirements of real-time performance and privacy in scenarios such as emergency rescue and industrial inspection.

[0252] Example 2

[0253] To achieve the above objectives, see Figure 5This embodiment also provides an indoor / outdoor adaptive passive IoT reflective positioning system, including the following modules:

[0254] Passive tag deployment and data acquisition module: used to deploy passive IoT tags in indoor and outdoor areas. The tags are powered by radio frequency energy harvesting technology and periodically backscatter modulated signals containing unique identifiers and environmental sensing data. The positioning terminal captures the backscattered signals of the passive tags through radio frequency receivers, and analyzes the tag ID, temperature and humidity, air pressure, received signal strength RSSI, and tag response time—unique characteristics of passive IoT—to form a positioning information database.

[0255] Network architecture module: Improves satellite positioning accuracy by synchronously deploying GNSS augmentation base stations in outdoor areas; acquires satellite signals through the GNSS module to calculate three-dimensional position coordinates and position accuracy factor (PDOP); and collects accelerometer and gyroscope data through SINS to calculate the terminal's motion trajectory.

[0256] Environmental correction module: Considering the correction model of environmental parameters on the refractive index of electromagnetic waves, establish formulas for the influence of temperature, humidity and air pressure on the signal strength of passive IoT tag signal reception;

[0257] Spatiotemporal dynamic scene evaluation module: Construct a comprehensive discrimination function, integrate signal strength, environmental parameters and motion state, and output a scene evaluation function F∈[0,1] to distinguish indoor areas, outdoor areas and transition areas;

[0258] Dynamic positioning mode switching module: Dynamically switches positioning modes based on the evaluation function F.

[0259] In indoor environments, the F-value approaches 0: localization is achieved using the WKNN algorithm based on modified RSSI fingerprints and tag response times;

[0260] Transitional region, F is the intermediate value: adaptive fusion positioning of BeiDou / SINS and tag data;

[0261] In outdoor environments, the F-value approaches 1: BeiDou / SINS fusion anti-interference dynamic positioning.

[0262] The indoor-outdoor adaptive passive IoT reflective positioning system of this embodiment has the same advantages over the prior art as the above-mentioned indoor-outdoor adaptive passive IoT reflective positioning method, and will not be repeated here.

[0263] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An indoor-outdoor adaptive passive Internet of Things (IoT) reflection positioning method, characterized in that, Comprise the following steps: S1, passive tag deployment and data collection: deploy passive Internet of Things tags in indoor and outdoor areas, the tag is powered by radio frequency energy collection technology, and periodically backscatters modulated signals containing unique identifiers and environmental sensing data; the positioning terminal captures the backscattered signals of the passive tag through the radio frequency receiver, analyzes the tag ID, temperature and humidity, air pressure, received signal strength RSSI and passive Internet of Things unique feature information, and forms a positioning information database; in step S1, the absorption model formula of water vapor to passive Internet of Things signal is: wherein, is the water vapor absorption coefficient; is the water vapor density, is the radio frequency, , is the temperature; The influence of temperature and air pressure on electromagnetic waves is the same, both of which cause changes in air density and thus affect the refractive index of the medium, resulting in bending of the electromagnetic wave path, and the influence formula is: wherein is the refractive index; is the temperature; is the atmospheric pressure, is the water vapor pressure; Construct the environmental impact factor function: wherein is a base attenuation factor; is a humidity weight coefficient, calibrated by experiment; is a standard atmospheric correction refractive index; The final label is the received signal strength between the terminal device and the terminal device The mapping function is: The mapping function is: wherein, is an environmental decay factor; is an environmental reference constant; is a random noise subject to a normal distribution; S2, network architecture design: deploy GNSS enhanced base stations in outdoor areas to improve satellite positioning accuracy; obtain satellite signals through the GNSS module to calculate three-dimensional position coordinates and position dilution of precision PDOP; collect accelerometer and gyroscope data through SINS to calculate terminal motion trajectory; S3, environmental correction model: consider the correction model of environmental parameters on the refractive index of electromagnetic waves, and establish the influence formula of temperature and humidity, air pressure on the received signal strength of passive Internet of Things tag signals; S4, constructing a comprehensive evaluation function, fusing signal strength, environmental parameters and motion state, outputting a scene evaluation function to distinguish indoor area, outdoor area and transition area; in step S4, a spatio-temporal dynamic scene evaluation function of comprehensive perception information and signal vector needs to be constructed, and the evaluation function involves two types of data: one is the dynamic change of perception information along the spatio-temporal dimension; the other is the dynamic change of signal vector along the spatio-temporal dimension; therefore, a related influence factor function is constructed for information at a single time and a single location; Based on temperature and humidity, air pressure and other sensing information, construct the environmental impact factor function: wherein is a base attenuation factor; is a humidity weight coefficient, calibrated by experiment, is a standard atmospheric correction refractive index; Based on signal strength RSSI and label density, construct the signal vector factor function: wherein, is the number of received tags; is the maximum number of tags; is the average signal strength; is the maximum RSSI value in the sample set; is the minimum RSSI value in the sample set; normalized to ; Based on the time and space dynamic changes of the label, construct the space-time dynamic factor function: wherein, is a terminal moving speed; is a maximum terminal moving speed in the sample set; is an environmental factor time change rate, is an environmental factor change amount; is a transform time; is a weight coefficient; Fusion of the above influence factor functions, the final positioning comprehensive evaluation function is: wherein, , , is a weight coefficient, which needs to be calibrated by experiment; is a slope parameter; is an environmental impact factor function; is a signal vector factor function; is a space-time dynamic factor function; According to the different values of the positioning scene, the change of the positioning scene can be determined, the value tends to 0: indoor environment; the value is an intermediate value: transition area; the value tends to 1: outdoor environment; S5, dynamic switching of positioning mode: according to evaluation function Dynamic switching of positioning mode: Indoor environment, Value approaching 0: positioning based on WKNN algorithm with corrected RSSI fingerprints and tag response time Transition region, For the intermediate value: Beidou / SINS and adaptive fusion positioning of label data; Outdoor environment, The value tends to 1: Beidou / SINS fusion anti-jamming dynamic positioning.

2. The indoor-outdoor adaptive, passive Internet of Things reflective positioning method of claim 1, wherein, The positioning terminal integrates a GNSS module, an inertial navigation system SINS, a radio frequency radiation source, and a multi-band radio frequency receiver. The overall positioning network structure in the network architecture design includes: Passive Internet of Things tag access layer: densely deploy passive Internet of Things tags in indoor areas and indoor-outdoor transition areas. Each tag uploads modulated signals containing unique identifiers, pre-set anchor point coordinates, and environmental sensing data temperature and humidity, air pressure through backscatter communication technology; Terminal layer: the positioning terminal integrates a GNSS receiving module, an inertial navigation module SINS, and a multi-band passive Internet of Things reader module, realizing signal reception, data fusion, and positioning calculation; all positioning modes: outdoor GNSS / SINS, transition area fusion positioning, indoor passive tag positioning calculation are completed independently by the terminal; Auxiliary enhancement layer: deploy GNSS enhanced base stations in outdoor areas to improve satellite positioning accuracy.

3. The indoor-outdoor adaptive, passive Internet of Things reflective positioning method of claim 1, wherein, In step S1, passive Internet of Things tags are powered by environmental radio frequency energy collection technology, and the energy collection efficiency meets: wherein, is the energy harvesting power; is the tag energy conversion efficiency; is the reader transmit power; is the tag antenna gain, affected by antenna directivity; is the radio frequency wavelength; is the tag to positioning terminal distance; The survival time of passive Internet of Things tags without external power supply is: the tag's lifetime; the tag's energy storage capacity; the tag's power consumption in the dormant state; the energy harvested by the tag per unit time; the tag's power consumption in the active state.

4. The indoor-outdoor adaptive, passive Internet of Things reflective positioning method of claim 2, wherein, In step S1, the tag response time of the passive tag includes tag charging time, circuit starting time, data processing time, and backscattering modulation time, wherein the circuit starting time is the starting and initialization time of the internal circuit after the capacitor voltage reaches the standard, usually in the order of microseconds, the passive Internet of Things tag adopts a low-power communication protocol, and the data processing time is usually also in the order of microseconds, and the backscattering modulation time is usually in the order of nanoseconds; therefore, the response time of the passive tag The tag charging time should be given priority, that is: Passive internet of things tag charging time The formula can be expressed as: wherein, Ctag is a tag capacitance, Vth is a tag activation threshold voltage, Prcv is a received power, under the premise of a passive tag adopting a uniform standard, the influences of the tag capacitance, the threshold voltage, and the conversion efficiency can be considered as the same, but the received power of the receiving antenna is different due to the influences of the channel state and the gain direction factor, therefore, the differences in the response time of the tag to a certain extent reflect the spatial characteristics of the tag, and are included in the fingerprint library as the radio frequency fingerprint characteristics. Based on the influence of channel and gain directionality, the antenna receiving power is corrected as follows: in, This refers to the reader's transmission power. and These are the gains of the transmitting antenna and the tag receiving antenna, respectively, which are affected by the antenna directivity; The distance between the tag and the positioning terminal; This is a function of environmental impact factors.

5. The indoor-outdoor adaptive, passive Internet of Things reflective positioning method of claim 2, wherein, In step S5, based on the WKNN algorithm for positioning based on the corrected RSSI fingerprint and label response time, indoor fingerprint positioning is divided into the following two stages: The first stage is an offline stage, and a fingerprint library is constructed based on corrected RSSI data, environment perception parameter data, GNSS system data and inertial navigation data; The second stage is an online matching stage. First, current RSSI, tag response time and environment parameter data are collected, and then SINS short-time motion data is fused to optimize the WKNN weight: wherein is the current collected corrected real-time signal strength value; is the signal strength value of the th reference point in the fingerprint library; is the terminal velocity vector, provided by SINS; is the time interval; is the displacement difference between the candidate reference point and the positioning coordinates at the last time; is the attenuation coefficient, controlling the influence strength of motion consistency on the weight, calibrated by experiment; is the current collected passive tag response time; and is the response time of the th reference point passive tag; Then, Top-K reference points are selected according to a dynamic K value, which is driven by tag density and environment perception factor, and the formula is: wherein, is the base value; is the current RSSI sample standard deviation; is a preset maximum fluctuation threshold; is the number of received tags; is the maximum number of tag densities; is the environmental weight coefficient; is the environmental attenuation factor; is the environmental factor change amount; is the maximum allowed environmental factor deviation; is the environmental weight coefficient, calibrated by experiment, The greater the value, the more sensitive the K value to environmental fluctuations, which will tend to select more reference points to improve accuracy, and is more suitable for indoor environments with frequent temperature, humidity, and air pressure fluctuations. Conversely, The smaller the value, the less sensitive the K value to environmental fluctuations, and it is more suitable for stable environments; final fusion weight : relative positioning coordinates are obtained by weighted averaging : wherein is a fusion weight of the th reference point; is a positioning coordinate of the th reference point in a local coordinate system based on the reference points; The coordinates are local coordinates based on a reference point, so the relative positions need to be converted to global absolute coordinates using the anchor tags of absolute coordinates: in, , , It is the first The relative position between each reference point and the anchor point label is determined by the anchor point label's positioning coordinates in the reference point coordinate system. To decide: Thus, the absolute coordinates of the tag The formula is modified to: wherein, is the corrected reference point absolute coordinate; the anchor point is deployed in a more critical position, and a passive Internet of Things tag with a known absolute coordinate is selected as the anchor point.

6. The indoor-outdoor adaptive, passive Internet of Things reflective positioning method of claim 1, wherein, In step S5, adaptive fusion positioning in the transition area is divided into the following two stages: The first stage dynamically determines the scene bias based on a comprehensive evaluation function F, and assigns weights to multiple sources of data to confirm the GNSS / SINS fusion anti-interference dynamic positioning weight: wherein, is a comprehensive evaluation function; is a Beidou signal quality score; is a Beidou signal quality theoretical maximum value; is a GNSS position standard deviation; is a preset GNSS error threshold value; In combination with signal strength RSSI and tag density, a tag fingerprint positioning weight is defined: wherein, is the number of received tags; is the maximum number of tags; is the average signal strength; is the maximum RSSI value in the sample set; is the minimum RSSI value in the sample set; The second stage normalizes the weight: wherein, GNSS / SINS fusion anti-jamming dynamic positioning weight; tag fingerprint positioning weight; The final positioning coordinates are output: wherein, is a normalized GNSS / SINS fusion anti-jamming dynamic positioning weight; is a normalized tag fingerprint positioning weight; is a GNSS / SINS fusion anti-jamming dynamic positioning coordinate result; is a tag fingerprint positioning coordinate result.

7. The indoor-outdoor adaptive, passive Internet of Things reflective positioning method of claim 1, wherein, In step S5, Beidou / SINS fusion anti-interference dynamic positioning in an outdoor environment is divided into the following three stages: The first stage receives Beidou signals and dynamically evaluates the quality of the Beidou signals: wherein, is the number of visible satellites; is the maximum number of visible satellites, default is 12; is the mean signal-to-noise ratio; is the preset maximum signal-to-noise ratio; is the position dilution of precision; is the damping factor of the position dilution of precision, calibrated by experiment; is the mean satellite elevation angle; , , is the three-dimensional position variance of the receiver in the ECEF coordinate system; The second stage is to establish an adaptive weighted fusion model. First, the error of inertial navigation is estimated, and the error estimation formula is: wherein, is the SINS estimated velocity; is the Beidou Doppler velocity; is the accelerometer bias standard deviation; is the gyroscope drift standard deviation; is the dynamic weight coefficient, calibrated according to the accuracy of the specific instrument; Then, a Beidou / SINS adaptive weighted fusion model is established: wherein is a Beidou signal quality score; is an inertial navigation error estimate; is an error suppression coefficient; The third stage outputs the final three-dimensional positioning coordinates: wherein, is a weight coefficient of the Beidou signal; is a weight coefficient of the inertial navigation system; is a three-dimensional coordinate of the Beidou signal calculated by the terminal; is a three-dimensional coordinate of the inertial navigation system calculated by the terminal.

8. An indoor-outdoor adaptive passive Internet of Things reflective positioning system, characterized in that, The following modules are included: Passive tag deployment and data collection module: used for deploying passive Internet of Things tags in indoor and outdoor areas. The tags are powered by radio frequency energy harvesting technology, and periodically backscatter modulated signals containing unique identifiers and environmental sensing data. The positioning terminal captures the backscattered signals of the passive tags through the radio frequency receiver, analyzes the tag ID, temperature and humidity, air pressure, received signal strength RSSI and tag response time, and forms a positioning information database. In the passive tag deployment and data collection module, the formula for the absorption model of water vapor on passive Internet of Things signals is: wherein, is the water vapor absorption coefficient; is the water vapor density, is the radio frequency, , is the temperature; The influence of temperature and air pressure on electromagnetic waves is the same, which is to cause changes in air density and thus affect the refractive index of the medium, causing the path of electromagnetic waves to bend. The influence formula is: wherein is the refractive index; is the temperature; is the atmospheric pressure, is the water vapor pressure; An environmental impact factor function is constructed: wherein is a base attenuation factor; is a humidity weight coefficient, calibrated by experiment; is a standard atmospheric correction refractive index; The final label is the received signal strength between the terminal device and the terminal device The mapping function is: The mapping function is: wherein, is an environmental decay factor; is an environmental reference constant; is a random noise subject to a normal distribution; Network architecture module: GNSS enhanced base stations are deployed synchronously in outdoor areas to improve satellite positioning accuracy. Satellite signals are obtained through the GNSS module to calculate three-dimensional position coordinates and position dilution of precision PDOP. Accelerometer and gyroscope data are collected through SINS to calculate the terminal motion trajectory. Environment correction module: considering the correction model of environmental parameters on the refractive index of electromagnetic waves, the influence formula of temperature and humidity, air pressure on the received signal strength of passive Internet of Things tag signals is established. A spatiotemporal dynamic scene evaluation module: a comprehensive discriminant function is constructed to fuse signal intensity, environmental parameters and motion state, and output a scene evaluation function to distinguish indoor areas, outdoor areas and transition areas; in the spatiotemporal dynamic scene evaluation module, a spatiotemporal dynamic scene evaluation function of comprehensive perception information and signal vectors needs to be constructed, and the evaluation function involves two types of data: one is the dynamic change of perception information along the spatiotemporal dimension; the other is the dynamic change of signal vectors along the spatiotemporal dimension; therefore, a related influence factor function is constructed for information at a single time and a single location; Based on temperature and humidity, air pressure and other sensing perception information, an environmental impact factor function is constructed: wherein is a base attenuation factor; is a humidity weight coefficient, calibrated by experiment, is a standard atmospheric correction refractive index; Based on signal strength RSSI and tag density, a signal vector factor function is constructed: wherein, is the number of received tags; is the maximum number of tags; is the average signal strength; is the maximum RSSI value in the sample set; is the minimum RSSI value in the sample set; normalized to ; Based on the time and space dynamic changes of the tag, a time-space dynamic factor function is constructed: wherein, is a terminal moving speed; is a maximum terminal moving speed in the sample set; is an environmental factor time change rate, is an environmental factor change amount; is a transform time; is a weight coefficient; After fusing the above impact factor functions, the final positioning comprehensive evaluation function is determined as wherein, , , are weight coefficients, which need to be calibrated by experiments; is a slope parameter; is an environmental impact factor function; is a signal vector factor function; is a spatiotemporal dynamic factor function; According to different values, the change of positioning scene can be determined, tends to 0: indoor environment; is the intermediate value: transition area; tends to 1: outdoor environment; Positioning mode dynamic switching module: according to evaluation function Dynamic switching positioning mode: Indoor environment, Value approaching 0: positioning based on WKNN algorithm with corrected RSSI fingerprints and tag response time Transition region, For the intermediate value: Beidou / SINS and adaptive fusion positioning of label data; Outdoor environment, Values approach 1: Beidou / SINS fusion anti-jamming dynamic positioning.

Images