A data error correction method and system for RFID reading terminals

Abstract

This invention discloses a data error correction method and system for RFID ranging terminals, relating to the field of data processing technology. The method includes: performing radio frequency scanning of the working environment to establish error reference data; calibrating the original phase data of the target tag to obtain a target tag data matrix; performing sparse reconstruction optimization calculations on the target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag; constructing a set of phase observation equations, solving the set of phase observation equations to generate multiple candidate distance estimates; and performing consistency verification and fusion of the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI to output the final corrected distance estimate. This solves the technical problem of insufficient accuracy of RFID ranging results in complex multipath environments in existing technologies, achieving the technical effect of improving the accuracy of ranging results.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to a data error correction method and system for RFID reading terminals. Background Technology

[0002] RFID (Radio Frequency Identification) technology is widely used in warehousing and logistics, smart manufacturing, asset inventory, and indoor positioning due to its advantages such as non-contact operation, batch identification, and low deployment cost. Measuring tag distance using RFID readers is a crucial foundation for item location, trajectory tracking, and spatial awareness. However, in real-world applications, RFID systems typically operate in environments with dense metallic reflectors and complex spatial structures. Radio frequency signals are prone to reflection, diffraction, and scattering during propagation, resulting in significant multipath effects. In complex multipath environments, the signal received by the RFID reader is often a superposition of direct path signals and signals from multiple reflected paths. Its phase and amplitude information are affected by environmental changes, frequency switching, and terminal hardware deviations, leading to large ranging errors and unstable results in traditional ranging methods based on phase, delay, or Received Signal Strength Indication (RSSI). Especially in multi-frequency or frequency-hopping modes, phase shifts and system errors between different frequencies are difficult to correct uniformly, making the ranging results highly sensitive to environmental disturbances. In existing technologies, some solutions attempt to reduce ranging errors through simple filtering, averaging, or correction of single signal features. However, such methods usually fail to effectively distinguish between direct paths and multipath components, making it difficult to achieve stable and reliable distance estimation in complex environments. Therefore, they still cannot meet the accuracy and robustness requirements of high-precision ranging applications. Summary of the Invention

[0003] This application provides a data error correction method and system for RFID reading terminals, which solves the technical problem in the prior art that the phase measurement of RFID reading terminals in complex multipath environments is easily affected by frequency drift, environmental reflection and the superposition of equipment hardware errors, resulting in insufficient accuracy of reading results.

[0004] A first aspect of this application provides a data error correction method for an RFID reading terminal, the method comprising: The working environment is scanned by radio frequency to assess multipath complexity and identify available frequency bands. Error baseline data is established using reference tag signals at known locations. A set of optimal frequency points is selected from the available frequency bands to generate a frequency hopping sequence. Based on the frequency hopping sequence, the original phase and amplitude data of the target tag and at least one neighboring reference tag are synchronously acquired. The original phase data of the target tag is calibrated using the error baseline data of the reference tag to obtain a calibrated target tag data matrix. Using the known ideal channel response of the neighboring reference tag as a reference, sparse reconstruction optimization calculation is performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag. Using the separated direct path component, a set of phase observation equations is constructed at different frequency hopping points. Combined with the real-time phase offset information of the reference tag, the set of phase observation equations is solved to generate multiple candidate distance estimates. The candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused for consistency verification to output the final corrected distance estimate.

[0005] A second aspect of this application provides a data error correction system for an RFID reading terminal, the system comprising: The system comprises the following modules: a baseline establishment module, a calibration module, and a verification module. The baseline module performs a radio frequency scan of the operating environment, assesses multipath complexity, identifies available frequency bands, and establishes error baseline data using reference tag signals from known locations. The calibration module selects an optimal set of frequencies from the available frequency bands to generate a frequency hopping sequence. Based on this sequence, it synchronously acquires the original phase and amplitude data of the target tag and at least one neighboring reference tag. The error baseline data of the reference tag is used to calibrate the original phase data of the target tag, resulting in a calibrated target tag data matrix. The optimization module uses the known ideal channel response of a neighboring reference tag as a reference baseline to perform sparse reconstruction optimization calculations on the calibrated target tag data matrix, separating the direct path component from the mixed multipath signal of the target tag. The calculation module uses the separated direct path component to construct a set of phase observation equations at different frequency hopping points. Combined with the real-time phase offset information of the reference tag, the phase observation equations are solved to generate multiple candidate distance estimates. The verification module performs consistency verification and fusion of the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI, outputting the final corrected distance estimate.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, an RF scan of the working environment is performed to assess multipath complexity and identify available frequency bands. Error baseline data is established using reference tag signals from known locations. Next, a set of optimal frequency points is selected from the available frequency bands to generate a frequency hopping sequence. Based on this sequence, the raw phase and amplitude data of the target tag and at least one neighboring reference tag are simultaneously acquired. The raw phase data of the target tag is calibrated using the error baseline data of the reference tag, resulting in a calibrated target tag data matrix. Further, using the known ideal channel response of a neighboring reference tag as a reference benchmark, sparse reconstruction optimization calculations are performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag. Then, using the separated direct path component, a set of phase observation equations is constructed at different frequency hopping points. These equations are solved using real-time phase offset information from the reference tag, generating multiple candidate distance estimates. Finally, the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused for consistency verification, outputting the final corrected distance estimate. This invention solves the technical problem in existing technologies where RFID reading terminals are susceptible to the effects of frequency drift, environmental reflection, and the superposition of equipment hardware errors in complex multipath environments, resulting in insufficient accuracy of reading results. It achieves the technical effect of adaptive calibration of RFID reading phase error and effective separation of direct path components in complex multipath environments, and improves reading accuracy through consistent fusion of multi-source distance estimation. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This application provides a schematic flowchart of a data error correction method for an RFID reading terminal. Figure 2 This is a schematic diagram of the data error correction system structure of an RFID reading terminal provided in an embodiment of this application.

[0009] Explanation of reference numerals in the attached diagram: 11. Benchmark establishment module; 12. Calibration module; 13. Optimization module; 14. Calculation module; 15. Verification module. Detailed Implementation

[0010] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0011] Example 1, as Figure 1 As shown, this application provides a data error correction method for an RFID reading terminal, wherein the method includes: The working environment is scanned by radio frequency to assess multipath complexity and identify available frequency bands. Error baseline data is established using reference tag signals at known locations.

[0012] In this embodiment, the RFID reading terminal is controlled to sequentially switch carrier frequencies at preset frequency step intervals within its permitted operating frequency band, performing channel measurements at each carrier frequency point no less than a preset number of times. At each frequency point, a reference tag deployed at least at a known location in the working environment is read, and the corresponding received signal strength indication value, phase measurement value, and signal round-trip time stamp information are collected. Statistical analysis is performed on the multiple measurement results obtained at the same frequency point, and the mean and variance of the received signal strength indication value, the standard deviation of the phase measurement value, and the channel delay spread parameter estimated based on the phase change rate or time domain correlation function are calculated to characterize the multipath interference intensity and phase stability at that frequency point. The multipath interference intensity, phase stability, and time delay spread parameters corresponding to each frequency point are used as multipath complexity evaluation indicators. A comprehensive score is calculated for each scanning frequency point within the working frequency band, and the set of frequency points with multipath complexity below a preset threshold is identified as usable frequency bands based on the scoring results. Within the usable frequency bands, the theoretical phase value of the reference tag at each frequency point is calculated based on the known spatial location of the reference tag and the geometric relationship between the reading terminal and the reference tag. The theoretical phase value is then compared with the corresponding actual phase measurement value to obtain the instantaneous phase error of the reference tag at different frequency points. Statistical modeling of the instantaneous phase error is performed in the time, frequency, and spatial location dimensions to generate error benchmark data describing the inherent phase deviation of the system hardware, environmental multipath disturbances, and the influence of individual tag differences. This error benchmark data is associated and stored with the corresponding frequency point, spatial region, and timestamp for subsequent real-time calibration of the target tag phase measurement data.

[0013] Furthermore, the operating environment is scanned using radio frequency (RF) technology to assess multipath complexity and identify available frequency bands, including: Within the operating frequency band of the RFID reading terminal, a step frequency scan is performed at a preset step interval. At each scan frequency point, channel response data is collected a predetermined number of times. Based on the channel response data, the mean and variance of the received signal strength indication value, the standard deviation of the phase measurement value, and the delay spread value estimated based on the channel impulse response are calculated to obtain the characteristic data of all scan frequency points. Based on the characteristic data of all scan frequency points, multipath complexity assessment and available frequency band identification are performed to generate channel quality parameters for each frequency point.

[0014] Within the operating frequency band of the RFID reading terminal, the RFID reading terminal is controlled to switch the carrier frequency sequentially according to a preset frequency step interval, and perform step frequency scanning for each scanning frequency point; at each scanning frequency point, continuous reading of at least one reference tag at a known location is triggered, and channel response data is collected no less than a predetermined number of times, wherein the channel response data includes at least the received signal strength indication value, the phase measurement value, and relevant sampling information for time domain analysis.

[0015] For each channel response data acquired at a scanning frequency, the received signal strength indication value is statistically processed based on an AI large model to calculate its mean and variance, thereby characterizing the stability of the signal amplitude at that frequency. The phase measurement value is also statistically processed to calculate its standard deviation, thereby characterizing the fluctuation of the phase observation at that frequency. Simultaneously, a corresponding channel impulse response is constructed based on the channel response data, and the time delay spread value corresponding to that frequency is estimated according to the time distribution of the multipath components in the channel impulse response, thereby characterizing the multipath propagation characteristics of the signal at that frequency.

[0016] The mean and variance of the received signal strength indication value, the standard deviation of the phase measurement value, and the delay spread value corresponding to each scanning frequency point are used as frequency point feature data to construct a feature data set for all scanning frequency points. Based on the feature data set, each frequency point is comprehensively evaluated according to the preset multipath complexity evaluation rules, and frequency points with multipath complexity below the preset threshold and phase stability meeting the ranging requirements are identified as usable frequency bands. Finally, an evaluation result containing the channel quality parameters of each scanning frequency point is generated for subsequent frequency hopping sequence construction and ranging error correction processes.

[0017] Based on the feature dataset, a comprehensive evaluation of each frequency point is performed according to a preset multipath complexity evaluation rule. Specifically, for each scanning frequency point, the mean, variance, standard deviation of the phase measurement, and delay spread of the corresponding received signal strength indication (RSSI) value are read using an AI large-scale model. These feature parameters are then dimensionless to eliminate the influence of different physical dimensions on the comprehensive evaluation results. The dimensionless processing includes normalizing each feature parameter based on historical statistical intervals or preset physical upper limits. According to the mechanism by which multipath propagation affects ranging accuracy, corresponding weight coefficients are assigned to each feature parameter to construct a multipath complexity evaluation function. The weight coefficients for the standard deviation of the phase measurement and the delay spread are higher than those for the mean and variance of the RSI value to highlight the dominant influence of phase stability and multipath delay spread on the ranging error. The dimensionless feature parameters of each frequency point are then weighted and summed with their corresponding weight coefficients to obtain the multipath complexity score for that frequency point. The multipath complexity score is compared with a preset complexity threshold. When the multipath complexity score is less than the complexity threshold, the corresponding frequency point is determined to be a preferred frequency point with low multipath interference. When the multipath complexity score is greater than or equal to the complexity threshold, the corresponding frequency point is determined to be a non-preferred frequency point with high multipath interference. Based on the above determination results, all scanning frequency points are screened and sorted to generate a set of available frequency bands for subsequent frequency hopping sequence construction.

[0018] Furthermore, using reference tag signals from known locations, error baseline data is established, including: The RFID reading terminal is controlled to sequentially read signals from multiple reference tags deployed at known locations at multiple discrete frequency points within the available frequency band. For each reference tag at each frequency point, the phase measurement value, received signal strength indication value, and reading delay are recorded. The phase measurement value is compared with the theoretical phase value calculated based on the precise known location of the reference tag to obtain the instantaneous phase error of the reference tag at the frequency point. The instantaneous phase errors of all reference tags at all frequency points are statistically analyzed to establish an error field model describing the phase error as a function of spatial location and frequency, including real-time error reference data for system hardware deviation, environmental phase disturbance, and inherent tag offset.

[0019] After completing the identification of available frequency bands, the RFID reading terminal is controlled to switch the working frequency sequentially on multiple discrete frequency points within the available frequency band, and read reference tags deployed at multiple known locations in the working environment at each frequency point; for each reference tag at each frequency point, the corresponding phase measurement value, received signal strength indication value, and reading delay information from transmission to reception are collected and recorded to form an original measurement dataset indexed by "reference tag-frequency point-time stamp".

[0020] Based on the precise known spatial location of each reference tag and the installation parameters of the reading terminal antenna, the theoretical propagation distance at the corresponding frequency point is calculated according to the electromagnetic wave propagation model. The theoretical phase value of the reference tag at that frequency point is then converted from the theoretical propagation distance. The electromagnetic wave propagation model describes the propagation relationship of the radio frequency signal between the reading terminal antenna and the reference tag. Based on the known spatial location of the reference tag, the installation parameters of the reading terminal antenna, and the carrier frequency of the corresponding frequency point, the electromagnetic wave propagation model calculates the theoretical propagation distance and propagation delay of the radio frequency signal. The theoretical phase value at the corresponding frequency point is then converted from the propagation delay and used as a reference benchmark for phase error calculation. AI-assisted calculation is used to compare the theoretical phase value with the corresponding actual phase measurement value to calculate the instantaneous phase error of the reference tag at that frequency point. The instantaneous phase error is used to characterize the combined effects of the fixed phase deviation introduced by the system hardware, the phase disturbance caused by environmental multipath reflection, and the inherent offset of the reference tag itself.

[0021] Based on AI large model assistance, the instantaneous phase errors of all reference tags at all frequency points are statistically analyzed. The phase error distribution characteristics are modeled in both spatial and frequency dimensions to establish an error field model describing the relationship between phase error and spatial location and operating frequency. The error field model generates real-time error benchmark data to characterize system hardware deviations, environmental phase disturbances, and inherent tag offsets by estimating the mean, performing variance analysis, or regression fitting on the instantaneous phase errors. The error benchmark data is then associated and stored with the corresponding frequency points, spatial regions, and time windows for subsequent calibration of the target tag phase measurement data.

[0022] Select a set of optimal frequency points from the available frequency bands to generate a frequency hopping sequence. Based on the frequency hopping sequence, synchronously collect the original phase and amplitude data of the target tag and at least one nearby reference tag. Use the error reference data of the reference tag to calibrate the original phase data of the target tag to obtain the calibrated target tag data matrix.

[0023] Within the identified available frequency bands, the frequency points are sorted and filtered according to the channel quality parameters corresponding to each frequency point. Frequency points with low multipath complexity, high phase stability, and frequency spacing that meet the preset minimum spacing constraint are selected first to form a candidate frequency point set. From the candidate frequency point set, a frequency hopping sequence is generated according to a preset frequency hopping strategy. The number of frequency points and the switching rate included in the frequency hopping sequence are configured according to the motion state of the target tag to ensure that multi-frequency observation is completed within the ranging time window.

[0024] The RFID reading terminal is controlled to switch operating frequencies sequentially according to the frequency hopping sequence. At each frequency hopping point, signal data of the target tag and at least one reference tag whose spatial distance from the target tag is less than a preset proximity threshold are collected simultaneously. For each frequency hopping point, the original phase measurement value and received signal strength indication value corresponding to the target tag and the reference tag are recorded respectively, forming an original observation dataset indexed by "frequency point-tag identifier-time stamp".

[0025] Based on the aforementioned established error benchmark data, reference tag reference phase error values ​​corresponding to each frequency point in the current frequency hopping sequence are extracted. The actual phase measurement values ​​obtained by the reference tag during the current frequency hopping measurement are compared with the benchmark phase error values ​​to calculate the real-time phase offset reflecting the current environmental changes and system state. Using the real-time phase offset, the original phase measurement values ​​of the target tag at the corresponding frequency point are corrected frequency-by-frequency, i.e., the real-time phase offset of the corresponding frequency point is subtracted from the original phase measurement values ​​of the target tag to obtain the calibrated target tag phase data. The calibrated target tag phase data at each frequency hopping point is combined with the corresponding received signal strength indication values ​​according to the frequency hopping order to construct a calibrated target tag data matrix. This data matrix is ​​used for subsequent direct path component separation, multi-frequency phase observation equation construction, and range estimation calculation.

[0026] Furthermore, selecting a set of optimal frequency points from the available frequency bands to generate a frequency hopping sequence includes: Based on the multipath interference complexity score of each frequency point, frequency points with interference below the first threshold are selected as base frequencies. From the base frequency point set, frequency points with frequency intervals that meet the preset minimum interval requirement are selected to form an initial frequency hopping sequence. According to the motion state of the target tag, the length of the frequency hopping sequence and the switching frequency are adjusted. When the target tag is stationary or moving at low speed, a long sequence containing 4-8 frequency points is used; when the target tag is moving at high speed, a short sequence containing 2-4 frequency points is used to generate the frequency hopping sequence.

[0027] Based on the channel quality parameters of each frequency point obtained by the aforementioned radio frequency scanning, a corresponding multipath interference complexity score is calculated for each frequency point. The multipath interference complexity score is obtained by weighting the variance of the received signal strength indication value, the standard deviation of the phase measurement value, and the time delay spread value according to a preset weight. Frequency points with multipath interference complexity scores lower than a first threshold are identified as low interference frequency points, and the low interference frequency points are selected as the base frequency points to ensure that the selected frequency points have high phase stability and low multipath impact.

[0028] In the basic frequency point set, each frequency point is sorted according to its frequency value, and frequency points with a frequency interval greater than or equal to the preset minimum frequency interval are selected in sequence to avoid the frequency correlation between adjacent frequency points from interfering with the phase observation results, thereby forming a preliminary frequency hopping sequence; wherein, the preset minimum frequency interval is configured according to the system bandwidth, phase resolution requirements and hardware switching capabilities.

[0029] The initial frequency hopping sequence is dynamically adjusted based on the target tag's motion state. The target tag's motion state is determined by the rate of change or Doppler frequency shift characteristics of multiple consecutive ranging results. When the target tag is determined to be stationary or moving at low speed, the single ranging time window is extended, and a long frequency hopping sequence containing 4 to 8 frequency points is used to improve the redundancy and noise resistance of multi-frequency phase observations. When the target tag is determined to be moving at high speed, the single ranging time window is shortened, and a short frequency hopping sequence containing 2 to 4 frequency points is used to reduce the risk of phase mismatch caused by motion. Finally, a frequency hopping sequence adapted to the current motion state of the target tag is generated.

[0030] Furthermore, the calibrated target label data matrix is ​​obtained, including: Based on the error reference data, the reference phase error value of the currently used reference tag at each frequency point of the current frequency hopping sequence is extracted; the phase matrix obtained by the reference tag in the current actual measurement is compared with the reference phase error value to calculate the real-time phase offset of the current environment; the real-time phase offset is applied to the original phase matrix of the target tag, and the real-time phase offset of the corresponding frequency point is subtracted from each phase measurement value of the target tag to obtain the calibrated target tag data matrix.

[0031] Based on the aforementioned established and stored error reference data, and according to the currently selected frequency hopping sequence, reference phase error values ​​corresponding to the currently used reference tag and each frequency point in the frequency hopping sequence are extracted from the error reference data to form a reference phase error vector for the reference tag under the current frequency hopping conditions. Within the current ranging cycle, the RFID reading terminal is controlled to synchronously measure the reference tag according to the frequency hopping sequence, obtaining the actual phase measurement values ​​of the reference tag at each frequency hopping point, and constructing the actual phase matrix of the reference tag according to the frequency hopping order. The actual phase matrix of the reference tag is compared frequency-by-frequency with the reference phase error vector to calculate the phase offset of the reference tag at each frequency point. The phase offset is used to characterize the real-time phase offset introduced by current environmental changes, instantaneous multipath disturbances, and system state fluctuations.

[0032] The real-time phase offset is mapped to the frequency dimension corresponding to the target tag and applied to the original phase matrix of the target tag. Specifically, the real-time phase offset of the corresponding frequency point is subtracted from the original phase measurement value of the target tag at each frequency hopping point to eliminate common errors caused by system hardware drift and environmental phase disturbances. The phase measurement values ​​of the calibrated target tag at each frequency point are arranged in the frequency hopping order and together with the corresponding received signal strength indication value to form the calibrated target tag data matrix. The calibrated target tag data matrix serves as the input data for subsequent direct path component separation and distance estimation calculations.

[0033] Furthermore, when real-time data from nearby reference tags is unavailable, calibration is performed using the regional historical average phase offset stored in the error baseline data, or interpolation estimation is performed using the phase change trends of other calibrated target tags in the same environment.

[0034] When the real-time phase measurement data of the reference tag cannot be obtained in the current ranging cycle due to temporary failure, obstruction, or communication abnormality, the historical average phase offset of the region corresponding to the current spatial region of the target tag and the frequency hopping sequence used is retrieved from the error reference data. The historical average phase offset of the region is obtained by statistically analyzing the phase offset data of the reference tag collected in the region within a preset historical time window. The original phase measurement value of the target tag at each frequency hopping point is calibrated frequency by frequency using the historical average phase offset of the region.

[0035] When the historical average phase offset of the region cannot meet the current ranging accuracy requirements, at least one other target tag that has been calibrated in the same working environment is selected, and its phase change trend at the same or adjacent frequency points is extracted. Based on the phase change trend, combined with the spatial distance or ranging result similarity between the target tag and other target tags, the phase offset of the target tag at each frequency point is interpolated and estimated to obtain the estimated phase offset of the target tag, and the phase data of the target tag is calibrated accordingly.

[0036] Through the aforementioned downgrade calibration mechanism, even in the absence of real-time data from the reference tag, the ability to continuously calibrate the phase measurement data of the target tag can still be maintained, ensuring the stability and reliability of the subsequent direct path component separation and distance estimation process.

[0037] Using the known ideal channel response of a neighboring reference tag as a reference benchmark, sparse reconstruction optimization calculations are performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag.

[0038] Furthermore, using the known ideal channel response of a neighboring reference tag as a reference benchmark, sparse reconstruction optimization calculations are performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag, including: Based on the principles of electromagnetic propagation and system operating parameters, a guidance vector dictionary containing different combinations of time delay and angle of arrival is established. Based on at least one neighboring reference tag at a known location, the theoretical time delay and angle of arrival of the direct path are calculated, and the corresponding reference atom is matched from the guidance vector dictionary. Guided by the reference atom, the calibrated observation signal of the target tag is decomposed by solving an optimization problem containing sparsity constraints and reference consistency constraints to obtain a sparse coefficient vector. Based on the sparse coefficient vector, the coefficients of the direct path are identified, and the direct path signal components of the target tag are reconstructed using the guidance vector dictionary. Based on the energy proportion of the direct path components and the consistency of the parameters with the reference, the confidence evaluation value of the direct path separation result is calculated.

[0039] Based on the propagation mechanism of electromagnetic waves in free space and complex environments, and combined with the operating frequency, antenna layout parameters, and system sampling parameters of the RFID reading terminal, a steering vector dictionary containing multiple combinations of different time delay parameters and angle of arrival parameters is constructed. The steering vector dictionary is used to describe the channel response characteristics formed at the receiving end by the direct path and various possible reflection paths.

[0040] Based on at least one reference tag whose spatial distance from the target tag is less than a preset proximity threshold, the theoretical direct path propagation distance of the reference tag in the current operating frequency band is calculated using its precisely known spatial location and the installation parameters of the reading terminal. The corresponding theoretical delay and angle of arrival are then converted from the propagation distance. Using the theoretical delay and angle of arrival as index conditions, the corresponding steering vector is matched from the steering vector dictionary and used as a reference atom to characterize the ideal channel response features of the direct path. Using the reference atom as a guiding constraint, and the calibrated target tag data matrix as the observation input, an optimization problem containing sparsity constraints and reference consistency constraints is constructed. The sparsity constraint is used to limit the channel decomposition result to contain only a finite number of main propagation path components, and the reference consistency constraint is used to guide the solution to converge in a direction consistent with the direct path parameters of the reference tag. The optimization problem is solved iteratively to obtain the corresponding sparse coefficient vector, where each non-zero coefficient in the sparse coefficient vector corresponds to a potential propagation path in the steering vector dictionary.

[0041] Based on the sparse coefficient vector, the path component that best matches the reference atom in terms of time delay and angle of arrival parameters and has the largest coefficient amplitude is identified, and the path component is determined to be the direct path of the target tag. Using the corresponding steering vector and its sparse coefficient in the steering vector dictionary, the signal of the path component is reconstructed to obtain the direct path signal component of the target tag.

[0042] Based on the energy proportion of the direct path signal component in the overall reconstructed signal, and the degree of consistency between the time delay parameters and angle of arrival parameters of the direct path component and the reference atom, the confidence evaluation value of the direct path separation result is calculated. The confidence evaluation value is used to measure the reliability of the direct path separation result and serves as the weight basis for the subsequent phase observation equation construction and distance estimation fusion process.

[0043] Using the separated direct path components, a set of phase observation equations is constructed at different frequency hopping points. By combining the real-time phase offset information of the reference tag, the set of phase observation equations is solved to generate multiple candidate distance estimates.

[0044] Furthermore, multiple candidate distance estimates are generated, including: From the direct path signal component of the target tag, extract phase measurement values ​​at at least three different frequency hopping points, and simultaneously acquire phase measurement values ​​of at least one neighboring reference tag at the same frequency point. Use the known distance of the reference tag to perform real-time system error correction on the phase measurement values ​​of each frequency point to obtain calibrated multi-frequency phase values ​​of the target tag. Based on the calibrated multi-frequency phase values ​​and their corresponding frequencies, construct a set of phase observation equations with the true distance of the target tag and integer ambiguity as unknowns, and solve the set of phase observation equations to calculate multiple candidate distance estimates. Based on each candidate distance estimate obtained, verify its physical rationality and geometric consistency, eliminate candidate values ​​that do not meet the verification conditions, and generate the multiple candidate distance estimates.

[0045] Based on the signal components of the target tag's direct path, at least three different frequency hopping points are selected in the frequency hopping sequence, and the phase measurement values ​​of the target tag at each of the frequency points are extracted. Simultaneously, the phase measurement values ​​of at least one neighboring reference tag are acquired at the same frequency hopping point. Combined with the known real distance between the reference tag and the reading terminal, a systematic error analysis is performed on the phase measurement values ​​of the reference tag to obtain the real-time phase correction amount at each frequency point. The real-time phase correction amount is applied to the phase measurement values ​​of the target tag at the corresponding frequency point to obtain the calibrated multi-frequency phase values ​​of the target tag.

[0046] Based on the calibrated multi-frequency phase values ​​of the target tag and the corresponding carrier frequency, a phase observation equation is constructed at each frequency point. The phase observation equation uses the actual propagation distance between the target tag and the reading terminal and the integer ambiguity corresponding to the phase integer as unknowns. The phase observation equations corresponding to each frequency point are combined to form a multi-frequency phase observation equation set. By jointly solving or traversing the phase observation equation set, multiple distance solutions that satisfy the equation constraints are obtained, thereby calculating multiple candidate distance estimates.

[0047] For each candidate distance estimate, physical rationality and geometric consistency verification are performed sequentially. Physical rationality verification is used to determine whether the candidate distance estimate falls within the measurable distance range of the system and whether it matches the historical ranging trend of the target label. Geometric consistency verification is used to determine whether the spatial geometric relationship between the candidate distance estimate and the reference label meets the preset spatial constraints. Candidate distance estimates that fail either verification condition are eliminated, and the candidate distance estimates that pass the verification are retained as the input results for subsequent distance consistency fusion.

[0048] The candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused together for consistency verification, and the final corrected distance estimate is output.

[0049] Furthermore, the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused together after consistency verification to output the final corrected distance estimate, including: Consistency verification is performed based on the deviation characteristics of the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI from historical samples. After passing the consistency verification, a fusion weight is configured according to the confidence level of each distance evaluation value. The candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are weighted and averaged using the fusion weight to obtain the final corrected distance estimate.

[0050] Based on historical samples, the deviation distribution characteristics of candidate distance estimates, distances calculated based on direct path delay, and distances estimated based on received signal strength indicators are statistically analyzed under stable ranging conditions. The deviation distribution characteristics include at least mean deviation, fluctuation amplitude, or variance index. Within the current ranging period, the various distance estimation results are compared with their corresponding historical deviation distribution characteristics to determine whether their deviations fall within a preset consistency threshold range, thereby completing the consistency verification.

[0051] When at least two of the following distance estimation results pass the consistency verification: the candidate distance estimate, the distance calculated based on the direct path delay, and the distance estimated based on the received signal strength indication (RSI), a fusion weight is configured according to the confidence evaluation value corresponding to each distance estimation result. The confidence evaluation value is determined based on the historical stability of the corresponding distance estimation result, the current deviation level, and the reliability of the ranging method; the higher the confidence evaluation value, the larger the corresponding fusion weight. Using the fusion weight, a weighted average is calculated for the candidate distance estimates that have passed the consistency verification, the distance calculated based on the direct path delay, and the distance estimated based on the RSI, to obtain the final corrected distance estimate. When a distance estimation result fails the consistency verification, its corresponding fusion weight is reduced or it is excluded from the weighted fusion process. The remaining fusion weights are then normalized before weighted averaging to ensure the stability and reliability of the final corrected distance estimate. Through the above consistency verification and fusion process, adaptive filtering and fusion of multi-source distance estimation results are achieved, effectively suppressing the impact of single ranging method error on the final ranging result, and improving the ranging accuracy and robustness of RFID ranging terminals in complex environments.

[0052] In summary, the embodiments of this application have at least the following technical effects: First, an RF scan of the working environment is performed to assess multipath complexity and identify available frequency bands. Error baseline data is established using reference tag signals from known locations. Next, a set of optimal frequency points is selected from the available frequency bands to generate a frequency hopping sequence. Based on this sequence, the raw phase and amplitude data of the target tag and at least one neighboring reference tag are simultaneously acquired. The raw phase data of the target tag is calibrated using the error baseline data of the reference tag, resulting in a calibrated target tag data matrix. Further, using the known ideal channel response of a neighboring reference tag as a reference benchmark, sparse reconstruction optimization calculations are performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag. Then, using the separated direct path component, a set of phase observation equations is constructed at different frequency hopping points. These equations are solved using real-time phase offset information from the reference tag, generating multiple candidate distance estimates. Finally, the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused for consistency verification, outputting the final corrected distance estimate. This invention solves the technical problem in existing technologies where RFID reading terminals are susceptible to the effects of frequency drift, environmental reflection, and the superposition of equipment hardware errors in complex multipath environments, resulting in insufficient accuracy of reading results. It achieves the technical effect of adaptive calibration of RFID reading phase error and effective separation of direct path components in complex multipath environments, and improves reading accuracy through consistent fusion of multi-source distance estimation.

[0053] Example 2, based on the same inventive concept as the data error correction method for an RFID reading terminal in the foregoing examples, such as... Figure 2 As shown, this application provides a data error correction system for an RFID reading terminal, wherein the system includes: The benchmark establishment module 11 performs RF scanning of the working environment, evaluates multipath complexity, identifies available frequency bands, and establishes error benchmark data using reference tag signals at known locations. The calibration module 12 selects a set of optimal frequency points from the available frequency bands to generate a frequency hopping sequence. Based on the frequency hopping sequence, it synchronously collects the original phase and amplitude data of the target tag and at least one neighboring reference tag. It uses the error benchmark data of the reference tag to calibrate the original phase data of the target tag, obtaining a calibrated target tag data matrix. The optimization module 13 uses the known ideal channel response of the neighboring reference tag as a reference benchmark to perform sparse reconstruction optimization calculations on the calibrated target tag data matrix, separating the direct path component from the mixed multipath signal of the target tag. The calculation module 14 uses the separated direct path component to construct a set of phase observation equations at different frequency hopping points. It solves the set of phase observation equations in conjunction with the real-time phase offset information of the reference tag, generating multiple candidate distance estimates. The verification module 15 performs consistency verification and fusion of the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI, and outputs the final corrected distance estimate.

[0054] Furthermore, the benchmark establishment module 11 is used to perform the following method: Within the operating frequency band of the RFID reading terminal, a step frequency scan is performed at a preset step interval. At each scan frequency point, channel response data is collected a predetermined number of times. Based on the channel response data, the mean and variance of the received signal strength indication value, the standard deviation of the phase measurement value, and the delay spread value estimated based on the channel impulse response are calculated to obtain the characteristic data of all scan frequency points. Based on the characteristic data of all scan frequency points, multipath complexity assessment and available frequency band identification are performed to generate channel quality parameters for each frequency point.

[0055] Furthermore, the benchmark establishment module 11 is used to perform the following method: The RFID reading terminal is controlled to sequentially read signals from multiple reference tags deployed at known locations at multiple discrete frequency points within the available frequency band. For each reference tag at each frequency point, the phase measurement value, received signal strength indication value, and reading delay are recorded. The phase measurement value is compared with the theoretical phase value calculated based on the precise known location of the reference tag to obtain the instantaneous phase error of the reference tag at the frequency point. The instantaneous phase errors of all reference tags at all frequency points are statistically analyzed to establish an error field model describing the phase error as a function of spatial location and frequency, including real-time error reference data for system hardware deviation, environmental phase disturbance, and inherent tag offset.

[0056] Furthermore, the calibration module 12 is used to perform the following method: Based on the multipath interference complexity score of each frequency point, frequency points with interference below the first threshold are selected as base frequencies. From the base frequency point set, frequency points with frequency intervals that meet the preset minimum interval requirement are selected to form an initial frequency hopping sequence. According to the motion state of the target tag, the length of the frequency hopping sequence and the switching frequency are adjusted. When the target tag is stationary or moving at low speed, a long sequence containing 4-8 frequency points is used; when the target tag is moving at high speed, a short sequence containing 2-4 frequency points is used to generate the frequency hopping sequence.

[0057] Furthermore, the calibration module 12 is used to perform the following method: Based on the error reference data, the reference phase error value of the currently used reference tag at each frequency point of the current frequency hopping sequence is extracted; the phase matrix obtained by the reference tag in the current actual measurement is compared with the reference phase error value to calculate the real-time phase offset of the current environment; the real-time phase offset is applied to the original phase matrix of the target tag, and the real-time phase offset of the corresponding frequency point is subtracted from each phase measurement value of the target tag to obtain the calibrated target tag data matrix.

[0058] Furthermore, the calibration module 12 is used to perform the following method: When real-time data from nearby reference tags is unavailable, calibration is performed using the regional historical average phase offset stored in the error baseline data, or interpolation estimation is performed using the phase change trends of other calibrated target tags in the same environment.

[0059] Furthermore, the optimization module 13 is used to perform the following method: Based on the principles of electromagnetic propagation and system operating parameters, a guidance vector dictionary containing different combinations of time delay and angle of arrival is established. Based on at least one neighboring reference tag at a known location, the theoretical time delay and angle of arrival of the direct path are calculated, and the corresponding reference atom is matched from the guidance vector dictionary. Guided by the reference atom, the calibrated observation signal of the target tag is decomposed by solving an optimization problem containing sparsity constraints and reference consistency constraints to obtain a sparse coefficient vector. Based on the sparse coefficient vector, the coefficients of the direct path are identified, and the direct path signal components of the target tag are reconstructed using the guidance vector dictionary. Based on the energy proportion of the direct path components and the consistency of the parameters with the reference, the confidence evaluation value of the direct path separation result is calculated.

[0060] Furthermore, the calculation module 14 is used to perform the following method: From the direct path signal component of the target tag, extract phase measurement values ​​at at least three different frequency hopping points, and simultaneously acquire phase measurement values ​​of at least one neighboring reference tag at the same frequency point. Use the known distance of the reference tag to perform real-time system error correction on the phase measurement values ​​of each frequency point to obtain calibrated multi-frequency phase values ​​of the target tag. Based on the calibrated multi-frequency phase values ​​and their corresponding frequencies, construct a set of phase observation equations with the true distance of the target tag and integer ambiguity as unknowns, and solve the set of phase observation equations to calculate multiple candidate distance estimates. Based on each candidate distance estimate obtained, verify its physical rationality and geometric consistency, eliminate candidate values ​​that do not meet the verification conditions, and generate the multiple candidate distance estimates.

[0061] Furthermore, the verification module 15 is used to perform the following methods: Consistency verification is performed based on the deviation characteristics of the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI from historical samples. After passing the consistency verification, a fusion weight is configured according to the confidence level of each distance evaluation value. The candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are weighted and averaged using the fusion weight to obtain the final corrected distance estimate.

[0062] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A data error correction method for an RFID reading terminal, characterized in that, The method includes: The working environment is scanned by radio frequency to assess multipath complexity and identify available frequency bands. Error baseline data is established using reference tag signals at known locations. Select a set of optimal frequency points from the available frequency bands to generate a frequency hopping sequence. Based on the frequency hopping sequence, synchronously collect the original phase and amplitude data of the target tag and at least one nearby reference tag. Use the error reference data of the reference tag to calibrate the original phase data of the target tag to obtain the calibrated target tag data matrix. Using the known ideal channel response of a neighboring reference tag as a reference benchmark, sparse reconstruction optimization calculations are performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag. Using the separated direct path components, a set of phase observation equations is constructed at different frequency hopping points. The phase observation equations are then solved by combining the real-time phase offset information of the reference tag to generate multiple candidate distance estimates. The candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused together for consistency verification, and the final corrected distance estimate is output. Using the known ideal channel response of a neighboring reference tag as a reference benchmark, sparse reconstruction optimization calculations are performed on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag, including: Based on the electromagnetic propagation principle and system operating parameters, a guidance vector dictionary containing different combinations of time delay and angle of arrival is established; Based on at least one neighboring reference tag at a known location, calculate the theoretical time delay and angle of arrival of the direct path, and match the corresponding reference atom from the guide vector dictionary; Guided by the aforementioned reference atom, the calibrated observation signal of the target label is decomposed by solving an optimization problem that includes sparsity constraints and reference consistency constraints to obtain a sparse coefficient vector; Based on the sparse coefficient vector, the coefficients of the direct path are identified, and the direct path signal components of the target label are reconstructed using the guide vector dictionary.

2. The data error correction method for an RFID reading terminal according to claim 1, characterized in that, Perform RF scanning of the operating environment to assess multipath complexity and identify available frequency bands, including: Within the operating frequency band of the RFID reading terminal, a step frequency scan is performed at a preset step interval, and channel response data is collected a predetermined number of times at each scan frequency point; Based on the channel response data, the mean and variance of the received signal strength indication value of the frequency point, the standard deviation of the phase measurement value, and the delay spread value estimated based on the channel impulse response are calculated to obtain the characteristic data of all scanning frequency points. Based on the feature data of all scanned frequency points, multipath complexity is evaluated and available frequency bands are identified, generating channel quality parameters for each frequency point.

3. The data error correction method for an RFID reading terminal according to claim 2, characterized in that, Using reference tag signals from known locations, establish error baseline data, including: The RFID reading terminal is controlled to sequentially read signals from multiple reference tags deployed at known locations at multiple discrete frequency points within the available frequency band. For each reference tag at each frequency point, the phase measurement value, received signal strength indication value, and reading delay are recorded. The instantaneous phase error of the reference tag at the frequency point is obtained by comparing the phase measurement value with the theoretical phase value calculated based on the precise known position of the reference tag. Statistical analysis was performed on the instantaneous phase errors of all reference tags at all frequencies to establish an error field model describing the phase error as a function of spatial location and frequency. This model includes real-time error benchmark data for system hardware deviations, environmental phase disturbances, and inherent tag offsets.

4. The data error correction method for an RFID reading terminal according to claim 3, characterized in that, Generate a frequency hopping sequence by selecting a set of optimal frequency points from the available frequency bands, including: Based on the multipath interference complexity score of each frequency point, the frequency points with interference below the first threshold are selected as the base frequency points. From the base frequency set, select frequency points whose frequency spacing meets the preset minimum spacing requirement to form an initial frequency hopping sequence; The length and switching frequency of the frequency hopping sequence are adjusted according to the motion state of the target tag. When the target tag is stationary or moving at low speed, a long sequence containing 4-8 frequency points is used; when the target tag is moving at high speed, a short sequence containing 2-4 frequency points is used to generate the frequency hopping sequence.

5. The data error correction method for an RFID reading terminal according to claim 4, characterized in that, The calibrated target label data matrix is ​​obtained, including: Based on the error reference data, the reference phase error value of the currently used reference label at each frequency point of the current frequency hopping sequence is extracted from it. The phase matrix obtained from the current actual measurement using the reference label is compared with the reference phase error value to calculate the real-time phase offset of the current environment; The real-time phase offset is applied to the original phase matrix of the target tag, and the real-time phase offset of the corresponding frequency point is subtracted from each phase measurement value of the target tag to obtain the calibrated target tag data matrix.

6. The data error correction method for an RFID reading terminal according to claim 5, characterized in that, Also includes: When real-time data from nearby reference tags is unavailable, calibration is performed using the regional historical average phase offset stored in the error baseline data, or interpolation estimation is performed using the phase change trends of other calibrated target tags in the same environment.

7. The data error correction method for an RFID reading terminal according to claim 1, characterized in that, Generate multiple candidate distance estimates, including: From the direct path signal component of the target tag, extract the phase measurement values ​​at at least three different frequency hopping points, and simultaneously acquire the phase measurement value of at least one neighboring reference tag at the same frequency point. Use the known distance of the reference tag to perform real-time system error correction on the phase measurement values ​​of each frequency point to obtain the calibrated multi-frequency phase value of the target tag. Based on the calibrated multi-frequency phase values ​​and their corresponding frequencies, a set of phase observation equations is constructed with the true distance of the target label and the integer ambiguity as unknowns. The set of phase observation equations is then solved to obtain multiple candidate distance estimates. Based on each candidate distance estimate obtained from the solution, physical rationality and geometric consistency verification are performed, candidate values ​​that do not meet the verification conditions are eliminated, and the multiple candidate distance estimates are generated.

8. The data error correction method for an RFID reading terminal according to claim 1, characterized in that, The candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI are fused together after consistency verification to output the final corrected distance estimate, including: Consistency verification is performed based on the deviation characteristics of the candidate distance estimates in historical samples, the distance calculated based on the direct path delay, and the distance estimated based on RSSI. Once the consistency verification is passed, the fusion weights are configured based on the confidence level of each distance evaluation value. The final corrected distance estimate is obtained by weighting the candidate distance estimate, the distance calculated based on the direct path delay, and the distance estimated based on RSSI using the fusion weights.

9. A data error correction system for an RFID reading terminal, characterized in that, The system is used to implement the data error correction method for an RFID reading terminal according to any one of claims 1-8, the system comprising: Benchmark establishment module: Performs RF scanning of the working environment, assesses multipath complexity and identifies available frequency bands, and establishes error benchmark data using reference tag signals at known locations; Calibration module: Selects a set of optimal frequency points from the available frequency bands to generate a frequency hopping sequence. Based on the frequency hopping sequence, it synchronously acquires the original phase and amplitude data of the target tag and at least one nearby reference tag. It uses the error reference data of the reference tag to calibrate the original phase data of the target tag and obtains the calibrated target tag data matrix. Optimization module: Using the known ideal channel response of the neighboring reference tag as a reference benchmark, it performs sparse reconstruction optimization calculation on the calibrated target tag data matrix to separate the direct path component from the mixed multipath signal of the target tag. Calculation module: Using the separated direct path components, a set of phase observation equations is constructed at different frequency hopping points. Combined with the real-time phase offset information of the reference tag, the set of phase observation equations is solved to generate multiple candidate distance estimates. Verification module: Performs consistency verification and fusion on the candidate distance estimates, the distance calculated based on the direct path delay, and the distance estimated based on RSSI, and outputs the final corrected distance estimate.

Images