Positioning method and system based on LED array

Abstract

This invention relates to the field of mobile carrier positioning technology, and specifically to a positioning method and system based on an LED array. The method includes: acquiring a composite visible light signal emitted by an LED array received by the mobile carrier at the location to be located; filtering and down-converting the initial modulation frequency to obtain a channel signal with a unified frequency; constructing a first analytical signal based on the unified frequency channel signal corresponding to a reference emitting unit, and constructing a second analytical signal based on the unified frequency channel signals corresponding to the remaining emitting units in the LED emitting units; determining a phase difference estimate based on the first and second analytical signals; performing time unwrapping processing on the multi-frame phase difference estimates to obtain an unwrapped phase difference; constructing a phase difference fingerprint for the location to be located based on the unwrapped phase difference; and determining a set of candidate reference points based on the similarity between the phase difference fingerprint for the location to be located and a database fingerprint, in order to construct a query center graph structure and estimate the positioning result of the location to be located.

Description

Technical Field

[0001] This invention relates to the field of mobile carrier positioning technology, and more specifically, to a positioning method and system based on an LED array. Background Technology

[0002] Underground or tunnel engineering spaces are characterized by their narrow and elongated shape, repetitive structures, severe obstruction, unavailability of global navigation satellite system signals, complex multipath radio signals, and continuous distribution of lighting facilities. When personnel, vehicles, inspection robots, robotic dogs, and construction equipment operate within these spaces, continuous positioning systems with low deployment costs, easy maintenance, resistance to electromagnetic interference, and high precision are required.

[0003] However, current visible light positioning methods require precise location mapping, numbering, management, and synchronous control of multiple lighting fixtures. In long-distance engineering spaces such as tunnels, engineering corridors, and underground caverns, the large number of lighting fixtures and their long deployment range make reliance on multi-base station collaborative positioning costly in the early stages and difficult in engineering implementation. Visible light positioning methods based on received signal strength are susceptible to changes in ambient light, receiver attitude, light source power fluctuations, surface reflection, and obstruction. Especially in situations such as tunnel slopes, vehicle vibrations, robot turning, and tilting of personnel's handheld terminals, changes in received light intensity are not entirely caused by positional changes, easily leading to fingerprint drift and increased positioning errors, resulting in insufficient positioning accuracy. Therefore, current positioning methods suffer from high costs and insufficient accuracy.

[0004] It should be noted that the information in the background section above is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a positioning method and system based on LED arrays, thereby overcoming the problems existing in related technologies, reducing positioning costs and complexity, and improving positioning accuracy.

[0006] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0007] According to one aspect of the present invention, a positioning method based on an LED array is provided, comprising: acquiring a composite visible light signal emitted by an LED array received by a mobile carrier at a location to be positioned, wherein the LED array is obtained by integrating multiple LED light-emitting units from the same illumination unit, and the initial modulation frequencies corresponding to each LED light-emitting unit are non-overlapping; filtering and down-converting each initial modulation frequency obtained by frequency decomposition of the composite visible light signal to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency; determining a reference light-emitting unit among the multiple LED light-emitting units, and constructing a first analytical signal based on the channel signal with the unified frequency corresponding to the reference light-emitting unit; and determining the remaining light-emitting units among the multiple LED light-emitting units. The channel signals corresponding to the unified frequency of the element are used to construct a second analytical signal. Based on the first and second analytical signals, the phase difference estimate between multiple LED light-emitting units inside the LED array is determined. Multi-frame phase difference estimates of the position to be located are obtained. The multi-frame phase difference estimates are subjected to time unwrapping processing to obtain multiple unwrapped phase differences. The phase difference fingerprint to be located is constructed based on the multiple unwrapped phase differences. Based on the similarity between the phase difference fingerprint to be located and the pre-constructed database fingerprint, a set of candidate reference points is determined from the database fingerprint. A query center graph structure is constructed based on the set of candidate reference points, the position to be located, and the phase difference fingerprint to be located. The positioning result of the position to be located is estimated based on the query center graph structure.

[0008] According to one aspect of the present invention, a positioning system based on an LED array is provided, comprising: a photoelectric receiving module for acquiring a composite visible light signal emitted by an LED array received by a mobile carrier at a location to be positioned, wherein the LED array is obtained by integrating multiple LED light-emitting units from the same illumination unit, and the initial modulation frequencies corresponding to each LED light-emitting unit are non-overlapping; a signal processing module for filtering and down-converting each initial modulation frequency obtained by frequency decomposition of the composite visible light signal to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency; and a phase difference estimation module for determining a reference light-emitting unit among the multiple LED light-emitting units, constructing a first analytical signal based on the channel signal with the unified frequency corresponding to the reference light-emitting unit, and estimating the phase difference based on the phase difference of the multiple LED light-emitting units. The channel signals of the remaining light-emitting units corresponding to the unified frequency are used to construct second analytical signals, so as to determine the phase difference estimate between multiple LED light-emitting units inside the LED array based on the first analytical signal and the second analytical signal; the phase difference fingerprint construction module is used to obtain multi-frame phase difference estimates of the position to be located, perform time unwrapping processing on the multi-frame phase difference estimates to obtain multiple unwrapped phase differences, and construct the phase difference fingerprint to be located based on the multiple unwrapped phase differences; the positioning module is used to determine the candidate reference point set from the database fingerprint based on the similarity between the phase difference fingerprint to be located and the pre-constructed database fingerprint, and construct a query center graph structure based on the candidate reference point set, the position to be located and the phase difference fingerprint to be located, so as to estimate the positioning result of the position to be located based on the query center graph structure.

[0009] According to one aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements any of the methods described above.

[0010] According to one aspect of the present invention, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform any of the above methods by executing the executable instructions.

[0011] The LED array-based positioning method in the exemplary embodiments of the present invention, on the one hand, utilizes the LED array within a single lighting unit to construct the phase difference fingerprint to be located. This eliminates the need for strict synchronization and complex geometric coordination between multiple lighting units, reducing the workload of multi-base station mapping, calibration, synchronization, and maintenance, lowering deployment costs and engineering maintenance complexity, and making it more suitable for large-scale deployment in long-distance scenarios such as tunnels, engineering corridors, and integrated utility tunnels. Furthermore, by setting distinguishable modulation frequencies for different LED light-emitting units within the same lighting unit and extracting the phase difference estimate within the LED array, a single lighting unit is no longer just a regular lighting point, but forms an array positioning base station with spatial coding capabilities, thereby enhancing the observability of single-lamp positioning. On the other hand, the phase difference fingerprint to be located belongs to the relative phase characteristics between different light-emitting units within the LED array, which can, to a certain extent, weaken the impact of overall light source brightness changes, receiver gain changes, slowly varying ambient light interference, and some receiver intensity fluctuations on the positioning results, thus improving positioning stability in complex lighting environments such as underground areas and tunnels. Furthermore, by unwrapping the phase and constructing the phase difference fingerprint to be located, the impact of the original phase period jump on subsequent fingerprint matching is reduced, enabling the phase difference fingerprint to be stored and matched in a more stable and continuous form, thereby improving the accuracy of positioning.

[0012] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0013] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation.

[0014] Figure 1 The diagram illustrates an application environment related to an LED array-based positioning method, an exemplary embodiment of the present invention.

[0015] Figure 2 A flowchart of a positioning method based on an LED array according to an exemplary embodiment of the present invention is shown.

[0016] Figure 3 A schematic diagram of a single-lamp LED array and its frequency encoding according to an exemplary embodiment of the present invention is shown.

[0017] Figure 4 A flowchart illustrating the acquisition of a channel signal with a unified frequency according to an exemplary embodiment of the present invention is shown.

[0018] Figure 5A flowchart illustrating an embodiment of the present invention for obtaining an estimate of the phase difference between multiple LED light-emitting units is shown.

[0019] Figure 6 A flowchart illustrating the construction of a phase difference fingerprint to be located is shown according to an exemplary embodiment of the present invention.

[0020] Figure 7 A flowchart illustrating the determination of a set of candidate reference points according to an exemplary embodiment of the present invention is shown.

[0021] Figure 8 A flowchart illustrating the construction of a query center graph structure according to an exemplary embodiment of the present invention is shown.

[0022] Figure 9 A flowchart illustrating the localization result based on a query center graph structure for estimating the location to be located, according to an exemplary embodiment of the present invention, is shown.

[0023] Figure 10 A schematic diagram illustrating a location based on a query center graph structure according to an exemplary embodiment of the present invention is shown.

[0024] Figure 11 A schematic diagram of the composition of a positioning system based on an LED array according to an exemplary embodiment of the present invention is shown.

[0025] Figure 12 A schematic diagram of another LED array-based positioning system according to an exemplary embodiment of the present invention is shown. Detailed Implementation

[0026] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the invention will be more comprehensive and complete, and will fully convey the concept of exemplary embodiments to those skilled in the art. The same reference numerals in the drawings denote the same or similar structures, and therefore their detailed description will be omitted.

[0027] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a full understanding of embodiments of the invention. However, those skilled in the art will recognize that the technical solutions of the invention can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known structures, methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of the invention.

[0028] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, or in one or more software-hardened modules, or in different network and / or processor devices and / or microcontroller devices.

[0029] Current visible light positioning methods require precise location mapping, numbering, management, and synchronous control of multiple lighting fixtures. In long-distance engineering spaces such as tunnels, engineering corridors, and underground caverns, the large number of lighting fixtures and their long deployment range make reliance on multi-base station collaborative positioning costly in the early stages and difficult in engineering implementation. Visible light positioning methods based on received signal strength are susceptible to changes in ambient light, receiver attitude, light source power fluctuations, surface reflection, and obstruction. Especially in situations such as tunnel slopes, vehicle vibrations, robot turning, and tilting of personnel's handheld terminals, changes in received light intensity are not entirely caused by positional changes, easily leading to fingerprint drift and increased positioning errors, resulting in insufficient positioning accuracy. Furthermore, existing single-light fixture or single-base station positioning methods have limited observable information, making it difficult to provide sufficient spatial discrimination capabilities without adding extra sensors and complex hardware.

[0030] To address one or more of the aforementioned problems, exemplary embodiments of the present invention provide a positioning method based on an LED array. This method constructs an LED array within a single lighting unit and utilizes the phase difference between different LED light-emitting units within the array as a positioning fingerprint, replacing methods that rely solely on received signal strength or multi-lamp geometric positioning. This improves the position discrimination capability and positioning robustness under single-lamp conditions. Furthermore, by constructing a query center graph structure using the similarity between the phase difference fingerprint to be located and a pre-built database fingerprint to determine a set of candidate reference points, along with the location to be located, the positioning result is estimated through the query center graph structure. This avoids the problem of insufficient local structural representation, thereby achieving low-cost, high-precision, and highly robust positioning estimation of mobile vehicles in underground and tunnel engineering spaces.

[0031] The LED array-based positioning method provided in the exemplary embodiments of the present invention can be applied to, for example... Figure 1 The application environment shown is as follows. The deployment environment can be underground or enclosed spaces where global navigation satellite system signals are unavailable or limited, such as tunnels, underground chambers, underground factories, mine roadways, engineering corridors, integrated utility tunnels, underground transportation spaces, long-distance construction passages, and other enclosed or semi-enclosed environments for the positioning of personnel terminals, inspection robots, robot dogs, vehicles, construction equipment, and emergency equipment.

[0032] The single illumination unit 110 integrates multiple LED light-emitting units 120 to form a visible light positioning base station with a fixed array geometry. A receiving end is provided on the mobile carrier 140 located in the carrier working space 130 to receive composite visible light signals from the LED light-emitting units. The mobile carrier 140 can be a portable mobile terminal, robot, robot dog, vehicle, or other equipment, etc., and the present invention does not impose any special limitations on it.

[0033] like Figure 2 The diagram shown is a flowchart of an LED array-based positioning method according to an exemplary embodiment of the present invention. (Refer to...) Figure 2 As shown, the method includes steps S210 to S250: Step S210: Obtain the composite visible light signal emitted by the LED array received by the mobile carrier at the location to be positioned. The LED array is obtained by integrating multiple LED light-emitting units from the same lighting unit, and the initial modulation frequencies corresponding to each LED light-emitting unit do not overlap.

[0034] Step S220: Filter and down-convert each initial modulation frequency obtained by frequency decomposition of the composite visible light signal to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency.

[0035] Step S230: Determine a reference light-emitting unit among multiple LED light-emitting units, and construct a first analytical signal based on the channel signal of the reference light-emitting unit after unifying the frequency. Construct a second analytical signal based on the channel signal of the remaining light-emitting units among the multiple LED light-emitting units after unifying the frequency, so as to determine the phase difference estimate between multiple LED light-emitting units inside the LED array based on the first analytical signal and the second analytical signal.

[0036] Step S240: Obtain multi-frame phase difference estimates of the location to be located, perform time unwrapping processing on the multi-frame phase difference estimates to obtain multiple unwrapped phase differences, and construct the phase difference fingerprint of the location to be located based on the multiple unwrapped phase differences.

[0037] Step S250: Based on the similarity between the phase difference fingerprint to be located and the pre-built database fingerprint, determine the candidate reference point set from the database fingerprint, and construct a query center graph structure based on the candidate reference point set, the location to be located, and the phase difference fingerprint to be located, so as to estimate the location result of the location to be located based on the query center graph structure.

[0038] The LED array-based positioning method in the exemplary embodiments of the present invention, on the one hand, utilizes the LED array within a single lighting unit to construct the phase difference fingerprint to be located. This eliminates the need for strict synchronization and complex geometric coordination among multiple luminaires, reducing the workload of multi-base station mapping, calibration, synchronization, and maintenance, lowering deployment costs and engineering maintenance complexity, and making it more suitable for large-scale deployment in long-distance scenarios such as tunnels, engineering corridors, and integrated utility tunnels. Furthermore, by setting distinguishable modulation frequencies for different LED light-emitting units within the same lighting unit and extracting the phase difference estimate within the LED array, a single lighting unit is no longer just a regular lighting point, but forms an array positioning base station with spatial coding capabilities, thereby enhancing the observability of single-lamp positioning. On the other hand, the phase difference fingerprint to be located belongs to the relative phase characteristics between different light-emitting units within the LED array, which can, to a certain extent, weaken the impact of overall light source brightness changes, receiver gain changes, slowly varying ambient light interference, and some receiver intensity fluctuations on the positioning results, thus improving positioning stability in complex lighting environments such as underground areas and tunnels. Furthermore, by unwrapping the phase and constructing the phase difference fingerprint to be located, the impact of the original phase period jump on subsequent fingerprint matching is reduced, enabling the phase difference fingerprint to be stored and matched in a more stable and continuous form, thereby improving the accuracy of positioning.

[0039] Steps S210 to S250 will be described in more detail below.

[0040] In step S210, the composite visible light signal emitted by the LED array received by the mobile carrier at the location to be positioned is acquired. The LED array is obtained by integrating multiple LED light-emitting units from the same lighting unit, and the initial modulation frequencies corresponding to each LED light-emitting unit do not overlap.

[0041] In an exemplary embodiment of the present invention, the LED array, i.e., a single-lamp LED array, includes multiple LED light-emitting units installed within the same lighting unit, such as multiple LED light-emitting units installed within the same lamp housing. These multiple LED light-emitting units are controlled by a unified clock from the same modulation drive module, ensuring a stable relative phase relationship between the LED light-emitting units within the array. Each LED light-emitting unit corresponds to a frequency channel at the receiving end of the mobile carrier.

[0042] For example, the LED array may include the following set of LED light-emitting units: Formula 1 In the formula, It is a collection of LED light-emitting units within an LED array; Indicates the first One LED light-emitting unit; Indicates the LED light-emitting unit number; This represents the total number of LED light-emitting units within a single lighting unit. The position of the i-th LED light-emitting unit in the lighting unit coordinate system is represented as: Formula 2 In the formula, Indicates the first Each LED light-emitting unit is in the coordinate system of the lighting unit. The three-dimensional coordinates below; , and They represent the first Each LED light-emitting unit is in the coordinate system of the lighting unit. lower edge axis, shaft and Coordinate components of the axis; superscript Indicates the coordinate system of the lighting fixture; This represents the transpose of a vector.

[0043] In an exemplary embodiment, the method may further include: for each LED light-emitting unit, determining the initial modulation frequency of the LED light-emitting unit based on a preset basic modulation frequency and the odd frequency sequence number corresponding to the LED light-emitting unit; the odd frequency sequence number is predetermined based on the number of the LED light-emitting unit in a plurality of LED light-emitting units or the number of the driving channel, and the odd frequency sequence number does not change with the position of the moving carrier.

[0044] To enable the receiver to distinguish different LED light-emitting units from the composite visible light signal, this invention assigns non-overlapping modulation frequencies to each LED light-emitting unit. Specifically, an odd harmonic frequency allocation method can be used: Formula 3 In the formula, For the first The initial modulation frequency of each LED light-emitting unit; Indicates the preset basic modulation frequency; Indicates the LED light-emitting unit number; Indicates the first Each LED light-emitting unit corresponds to an odd-order frequency sequence number. The fundamental modulation frequency is a preset parameter, the value of which can be determined based on the bandwidth of the LED driver circuit, the sampling rate of the photoelectric receiving module at the receiving end, the frequency resolution, and the environmental interference frequency band, or can be flexibly set based on historical experience. The odd-order frequency sequence number can be predetermined based on the LED light-emitting unit's number in the array or the driver channel number, and does not change with the position of the moving carrier. Based on this sequence number, non-overlapping modulation frequencies can be assigned to different LED light-emitting units, enabling frequency separation at the receiving end.

[0045] Correspondingly, the emitted light power of the i-th LED light-emitting unit is: Formula 4 In the formula, Indicates the first Each LED light-emitting unit at time The instantaneous emitted optical power; Indicates the first The DC light power component of each LED light-emitting unit, which can be determined by the LED drive current setting value, the luminaire factory calibration parameters or the on-site light power measurement results, is used to describe the reference luminous intensity of the LED light-emitting unit; Indicates the first The amplitude of the AC modulated optical power of each LED light-emitting unit is determined by both the modulation depth and the DC optical power. It can be determined by multiplying the modulation depth by the DC optical power. This value can also be preset by the modulation parameters of the LED driver circuit, or obtained by frequency domain amplitude estimation after the signal is acquired at the receiving end. In subsequent phase difference fingerprinting, this amplitude is used to determine signal quality, such as whether a particular LED channel is valid. Indicates the first The initial modulation frequency of each LED light-emitting unit; Indicates time; Indicates the first The initial modulation phase of each LED light-emitting unit can be set by a unified modulation driving module or recorded in a database fingerprint during offline calibration. It should be understood that this invention uses relative phase difference as a positioning feature, eliminating the need to individually solve for the absolute initial phase of each LED during online positioning.

[0046] For example, such as Figure 3 This is a schematic diagram of a single-lamp LED array and its frequency encoding according to the present invention. The initial modulation frequency of the LED light-emitting unit is determined by an odd harmonic frequency allocation method. .

[0047] This invention eliminates the need for a multi-base station geometric positioning system composed of multiple lighting units. Instead, it forms an LED array positioning base station with spatial coding capabilities within a single lighting unit. Since each LED light-emitting unit is located within the same lighting unit, the clock, drive, and phase relationships within the array are more easily maintained stably, reducing the costs associated with deploying multiple lighting units, synchronizing across luminaires, and mapping large-scale anchor points.

[0048] Optionally, the present invention can also impose observability constraints on the phase difference of a single-lighting-unit LED array. Specifically, the observability constraints on the single-lamp LED array are based on a preset fundamental modulation frequency, array size, and coverage area.

[0049] Specifically, a reference light-emitting unit is determined in the LED array. This reference light-emitting unit can be selected from the center of the array, a position with a high signal-to-noise ratio, or a position with high long-term stability. It can be flexibly adjusted according to actual needs, and there are no specific limitations. It is worth noting that the reference light-emitting unit described below is this reference light-emitting unit, and will not be described again hereafter.

[0050] Based on this, the reference light-emitting unit and the first The phase difference amplitude between individual LED light-emitting units is expressed as: Formula 5 In the formula, Indicates reference light-emitting unit With the The phase difference amplitude between individual LED light-emitting units; Indicates the preset basic modulation frequency; Represents the speed of light; This indicates the distance from the light-emitting unit to the receiving end; Indicates the first The distance from each LED light-emitting unit to the receiving end; This indicates the LED light-emitting unit number. To prevent indistinguishable periodic ambiguity in the principal phase value, the present invention sets the following constraints: Formula 6 This constraint limits the relationship between the LED light-emitting unit number, the preset fundamental modulation frequency, and the effective coverage distance. Under engineering conditions where the LED array size is much smaller than the distance from the LED to the receiver, it can be approximated. Then the following index constraints are obtained: Formula 7 In the formula, Indicates the first The number offset of each LED light-emitting unit relative to the reference light-emitting unit; Represents the speed of light; Indicates the preset basic modulation frequency; This represents the distance from the reference light-emitting unit to the receiving end. Based on this, if the maximum coverage radius of the LED array is... The installation height of the light fixtures is Then the reference propagation distance is: Formula 8 In the formula, Indicates the horizontal coverage radius of the LED array; This indicates the height difference between the luminous surface of the lamp and the receiving surface; This represents the approximate maximum propagation distance from the reference light-emitting unit to the receiving end. Based on this, the selectable range for the LED light-emitting unit number is determined as follows: Formula 9 In the formula, Indicates the LED light-emitting unit number; Represents the speed of light; Indicates the preset basic modulation frequency; Indicates the coverage radius; Indicates the height of the light fixture.

[0051] It should be noted that before actual positioning is implemented, when designing, configuring and setting parameters of array units, the reasonable relationship between the preset basic modulation frequency, the number of LED light-emitting units and the effective coverage area is determined by observing the array design and parameter verification. This avoids the appearance of indistinguishable periodic ambiguity in the phase difference within the positioning area, thereby enabling the single lamp array design to take into account phase distinguishability, spectrum separability and positioning coverage capability.

[0052] In step S220, each initial modulation frequency obtained by frequency decomposition of the composite visible light signal is filtered and down-converted to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency.

[0053] In an exemplary embodiment of the present invention, the receiving end of the mobile carrier is provided with a photoelectric receiving module, which includes a photodetector, an analog front-end amplifier circuit, a filter circuit, an analog-to-digital converter, and a signal processing unit. When the mobile carrier is within the coverage area of ​​the LED array, the photoelectric receiving module receives a composite visible light signal formed by the superposition of multiple LED light-emitting units. Wherein, the first... The distance from each LED light-emitting unit to the photoelectric receiving module is expressed as: Formula 10 In the formula, Indicates the first The spatial distance between each LED light-emitting unit and the photoelectric receiving module; This indicates that the photoelectric receiving module is in the navigation coordinate system or the local engineering coordinate system. The lower position; Indicates the first Each LED light-emitting unit in the coordinate system The lower position; Denotes the Euclidean norm; superscript This indicates the navigation coordinate system or the local coordinate system of the project.

[0054] Furthermore, under the condition of direct propagation of visible light, the first The DC channel gain from each LED light-emitting unit to the receiver is expressed as: Formula 11 In the formula, Indicates the first DC channel gain between each LED light-emitting unit and the photoelectric receiving module; Indicates the first The Lambertian radiation order of each LED light-emitting unit can be calculated from the half-power angle in the LED device datasheet, or it can be determined by fitting the light intensity distribution at different angles measured on site. This indicates the effective receiving area of ​​the photoelectric receiving module; Indicates the first The distance between each LED light-emitting unit and the photoelectric receiving module; Indicates the first The radiation angle of each LED light-emitting unit is determined by the angle between the optical axis direction of the LED light-emitting unit and the direction in which the LED points to the photoelectric receiving module. The optical axis direction of the LED light-emitting unit is given in advance by the installation posture of the lamp, and the direction in which the LED points to the photoelectric receiving module is determined by the position of the LED and the position of the photoelectric receiving module. Indicates that the photoelectric receiving module is relative to the first The incident angle of each LED light-emitting unit is determined by the angle between the normal vector of the photoelectric receiving module and the direction of the incident light of the LED. The normal vector of the photoelectric receiving module can be preset by the installation posture of the receiving module or measured by the posture sensor of the moving carrier. Indicates the optical filter at the angle of incidence. The transmittance can be determined from the device datasheet or calibration experiment. When the photoelectric receiving module is not equipped with an optical filter, the transmittance is taken as 1. Indicates the angle of incidence of the optical concentrator. The gain below.

[0055] The composite visible light signal acquired by the photoelectric receiving module at the receiving end can be represented as: Formula 12 In the formula, Indicates the time of the photoelectric receiving module The acquired composite visible light signal; Indicates the responsivity of the photodetector; Indicates the first Channel gain corresponding to each LED light-emitting unit; Indicates the first AC modulated light power amplitude of each LED light-emitting unit; Indicates the first The initial modulation frequency of each LED light-emitting unit; Indicates the propagation distance; Represents the speed of light; Indicates the initial modulation phase; This represents the fixed phase deviation introduced by the driving circuit, receiving circuit, filter, and hardware channel; This indicates received noise.

[0056] In one exemplary embodiment, such as Figure 4 The initial modulation frequency obtained by frequency decomposition of the composite visible light signal is filtered and down-converted to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency, which may include: Step S410: For each LED light-emitting unit, obtain the equivalent amplitude of the initial modulation frequency corresponding to the LED light-emitting unit.

[0057] The equivalent amplitude is used to characterize the received intensity and signal quality corresponding to the LED light-emitting unit. This equivalent amplitude is the comprehensive amplitude observed at the receiver, which is jointly determined by the photodetector responsivity, visible light channel gain, LED modulated light power, and analog front-end gain. In actual calculations, it is not necessary to solve for each physical quantity separately, but to directly estimate the comprehensive amplitude through the received signal of the corresponding frequency channel.

[0058] Step S420: Obtain the equivalent fixed phase and narrowband noise corresponding to the initial modulation frequency.

[0059] The equivalent fixed phase is generated by the fixed delay of the drive channel, receiver channel, filter, and sampling link. It can be obtained through sampling at known reference points during offline calibration, or it can be recorded as a fixed deviation in the database fingerprint library for direct application during actual positioning. Narrowband noise is the random disturbance remaining after filtering of the corresponding frequency channel. It can be obtained through background noise measurement, static sampling statistics, or filtering residual estimation, and no special requirements are placed on it.

[0060] Step S430: Based on the equivalent amplitude, equivalent fixed phase and narrowband noise, perform narrowband filtering on the initial modulation frequency corresponding to the LED light-emitting unit to obtain the narrowband signal corresponding to the LED light-emitting unit.

[0061] After obtaining the equivalent amplitude, equivalent fixed phase, and narrowband noise, the initial modulation frequency corresponding to the LED light-emitting unit can be narrowband filtered using the following formula: Formula 13 In Equation 13, Indicates the first The narrowband signal after filtering each frequency channel, that is, the narrowband signal corresponding to the LED light-emitting unit; Indicates the first The equivalent amplitude of each channel; Indicates the first The equivalent fixed phase of each channel; Indicates the first Narrowband noise in each channel.

[0062] Step S440: Down-convert the narrowband signals corresponding to each LED light-emitting unit to a preset base frequency to obtain the channel signals of each LED light-emitting unit after unification of frequency.

[0063] To facilitate subsequent phase comparisons between different LED light-emitting units, this invention unifies each frequency channel to a preset base adjustment frequency. For example, the first The signal after down-conversion and filtering of each channel is represented as follows: Formula 14 In the formula, Indicates the first The signal of each channel is unified to the base frequency, that is, the channel signal of the LED light-emitting unit after unification of the base frequency; This represents the equivalent amplitude after down-conversion; Indicates the preset basic modulation frequency; Indicates the first The initial modulation frequency of each LED light-emitting unit; Indicates the propagation distance; Represents the speed of light; This represents the equivalent fixed phase after down-conversion; This indicates the noise after down-conversion.

[0064] This invention enables the observation of different LED light-emitting units to be compared under the same phase reference by frequency separation and unified down-conversion, thereby providing a stable signal basis for the construction of phase difference fingerprints within the array.

[0065] In step S230, a reference light-emitting unit is determined among the multiple LED light-emitting units, and a first analytical signal is constructed based on the channel signal of the reference light-emitting unit after unification of frequency. A second analytical signal is constructed based on the channel signal of the remaining light-emitting units in the multiple LED light-emitting units after unification of frequency, so as to determine the phase difference estimate between the multiple LED light-emitting units inside the LED array based on the first analytical signal and the second analytical signal.

[0066] In an exemplary embodiment of the present invention, a reference light-emitting unit is determined among a plurality of LED light-emitting units as described above. The remaining light-emitting units refer to the other light-emitting units among the plurality of LED light-emitting units besides the reference light-emitting unit.

[0067] Optionally, such as Figure 5As shown, a first analytical signal is constructed based on the channel signal of the reference light-emitting unit at a unified frequency, and a second analytical signal is constructed based on the channel signals of the remaining light-emitting units among the multiple LED light-emitting units at a unified frequency, respectively. The phase difference estimate between the multiple LED light-emitting units is then determined based on the first and second analytical signals, including: Step S510: Obtain the first Hilbert transform operator of the channel signal with unified frequency corresponding to the reference light-emitting unit, and obtain the second Hilbert transform operator of the channel signal with unified frequency corresponding to each remaining light-emitting unit.

[0068] Step S520: Determine the first analytical signal based on the channel signal with the unified frequency corresponding to the reference light-emitting unit and the first Hilbert transform operator.

[0069] Step S530: For each remaining light-emitting unit, determine the second analytical signal based on the channel signal with the unified frequency corresponding to the remaining light-emitting unit and the second Hilbert transform operator.

[0070] Step S540: Determine the phase difference estimate between multiple LED light-emitting units based on the first analytical signal and the second analytical signal.

[0071] The first analytical signal can be constructed using the following formula based on the channel signal of the reference light-emitting unit at a unified frequency: Formula 15 In the formula, This represents the first analytical signal corresponding to the reference light-emitting unit; This represents the real-valued signal of the channel signal corresponding to the reference light-emitting unit at a unified frequency. This indicates the number of the reference light-emitting unit.

[0072] The second analytical signal can be constructed based on the channel signals of the remaining light-emitting units in the LED light-emitting unit after unifying the frequency, using the following formula: Formula 16 In the formula, Indicates the first The second analytical signal corresponding to each channel signal is constructed by the channel signals of the remaining light-emitting units after unifying the frequency; Indicates the first The real-valued signal of the channel signal corresponding to each channel after unifying the frequency; Represents the imaginary unit; This represents the Hilbert transform operator.

[0073] After obtaining the first and second analytical signals, the phase difference estimate between multiple LED light-emitting units can be determined accordingly. Specifically, for each LED light-emitting unit, the complex conjugate of the first analytical signal can be obtained. Then, based on the second analytical signal and the complex conjugate, the phase difference estimate of the LED light-emitting unit relative to the reference light-emitting unit can be determined. Finally, based on the phase difference estimate of each LED light-emitting unit relative to the reference light-emitting unit, the phase difference estimate between multiple LED light-emitting units can be determined.

[0074] The phase difference estimate between the LED light-emitting unit and the reference light-emitting unit can be determined by the following formula: Formula 17 In the formula, Indicates the first Phase difference estimation of each LED light-emitting unit relative to a reference light-emitting unit; This represents the complex conjugate of the reference channel analyzed signal; Indicates the first The second analytical signal of each channel; This indicates taking the real part of a complex number; This indicates taking the imaginary part of a complex number; This represents the arctangent function in the four quadrants, used to obtain the value of the function located in the quadrant. The principal phase within the range.

[0075] In determining the phase difference estimation process, the positioning feature in this invention is not the received signal strength of a single LED, nor the angle of arrival or time of arrival between multiple lamps, but rather the relative phase difference between LED light-emitting units within the same lighting unit. Since the aforementioned phase difference estimation is a relative observation feature, it can, to a certain extent, offset the combined effects of slowly varying ambient light, changes in receiver gain, and fluctuations in overall optical power, thereby improving positioning stability in complex lighting environments such as underground spaces and tunnels.

[0076] In step S240, multi-frame phase difference estimates of the location to be located are obtained, time unwrapping processing is performed on the multi-frame phase difference estimates to obtain multiple unwrapped phase differences, and a phase difference fingerprint of the location to be located is constructed based on the multiple unwrapped phase differences.

[0077] In an exemplary embodiment of the present invention, multiple frames of phase difference estimation can be acquired at the location to be located. The method for obtaining the phase difference estimation of each frame is as described above and will not be repeated here. For example, Figure 6 As shown, obtaining multi-frame phase difference estimates of the location to be located, performing time unwrapping processing on the multi-frame phase difference estimates to obtain multiple unwrapped phase differences, and constructing a phase difference fingerprint of the location to be located based on the multiple unwrapped phase differences can include: Step S610: Based on the phase difference estimation of each frame, determine the phase difference vector representation corresponding to the phase difference estimation of each frame.

[0078] After obtaining multi-frame phase difference estimates for the location r to be located, the phase difference vector corresponding to the phase difference estimate of the l-th frame is determined as follows: Formula 18 In the formula, Indicates the first The phase difference vector corresponding to the frame; Indicates the first The first frame Each LED light-emitting unit is relative to the reference light-emitting unit. The phase difference; Indicates the total number of LED light-emitting units; Indicates the sampling frame number.

[0079] Step S620: Perform time unwrapping of consecutive sampled frames according to the phase difference vector representation corresponding to each frame to obtain multiple unwrapped phase differences.

[0080] After obtaining the phase difference vector of each frame, the phase difference vector of each frame can be continuously sampled and unwrapped over time using the following formula: Formula 19 In the formula, This represents the phase difference after untangling; This represents the phase unwrapping operator, used to eliminate the phase entanglement between adjacent sampling frames. Discontinuity caused by periodic jumps.

[0081] Step S630: Based on the preset effective variables for phase difference observation, the observation validity of each unwrapped phase difference is judged, and the phase difference that passes the validity screening is determined based on the judgment result.

[0082] This invention considers that factors such as occlusion, low signal-to-noise ratio, frequency leakage, and sudden noise can easily cause abnormal observations, thereby affecting the robustness of the constructed phase difference fingerprint. Therefore, it is necessary to determine the observation validity of each unwrapped phase difference. The determination of the observation validity of each unwrapped phase difference includes validity determination based on one or more of amplitude, signal-to-noise ratio, frequency, phase, and noise.

[0083] The preset effective variables for phase difference observation are as follows: Formula 20 In the formula, Indicates the first The first frame The validity variable for each LED channel is set to 1, indicating that the observation for that channel is valid, and 0, indicating that the observation for that channel is invalid. Indicates the first The received amplitude of each LED channel; Indicates the lower limit threshold of amplitude; Indicates the first Signal-to-noise ratio of each LED channel; Indicates the signal-to-noise ratio threshold; This indicates the maximum allowed phase transition threshold. The specific settings for these thresholds can be determined based on the needs of the actual scenario or engineering experience.

[0084] Based on Equation 20, the observation validity of each unwrapped phase difference is judged, and the phase difference that passes the validity screening is determined according to the judgment result.

[0085] Step S640: Construct the phase difference fingerprint to be located based on the phase differences that have passed the validity screening.

[0086] The phase differences that pass the validity screening can be averaged or robustly statistically analyzed to construct the phase difference fingerprint of the target location, for example, as shown in the following formula: Formula 21 In the formula, This represents the phase difference fingerprint to be located; Indicates the first The phase difference of each LED channel is determined through effective screening and robust statistical analysis.

[0087] Correspondingly, channels that fail the screening will not participate in subsequent fingerprint matching, or they can be marked as invalid in the mask.

[0088] Based on this, the present invention can address the problems of phase observation being easily wrapped, easily jumping, and easily affected by channel anomalies by establishing a fingerprint from the original phase difference to the stable phase difference, so that the phase difference features can be stably used for engineering spatial positioning, thereby improving the accuracy of positioning estimation.

[0089] In step S250, a set of candidate reference points is determined from the database fingerprint based on the similarity between the phase difference fingerprint to be located and the pre-built database fingerprint. A query center graph structure is then constructed based on the set of candidate reference points, the location to be located, and the phase difference fingerprint to be located, so as to estimate the location result of the location to be located based on the query center graph structure.

[0090] In an exemplary embodiment of the present invention, the pre-built database fingerprint can be a fingerprint in a database built in the offline stage. Multiple preset reference points can be determined in the carrier workspace, and the database fingerprint can be built in advance based on the normalized fingerprint embedding vector and validity mask vector corresponding to each preset reference point and the coordinate information of the preset reference points.

[0091] The preset reference points can be set according to a grid layout, equidistant layout along the line, or denser layout in key areas. For example, for tunnels and engineering corridors, the preset reference points are preferably equidistantly arranged along the direction of travel of the moving vehicle; the density of points is increased in areas where the phase difference fingerprint changes rapidly, such as corners, intersections, and tunnel entrances. This invention can determine multiple preset reference points based on the actual working space of the vehicle. Each preset reference point can be a known two-dimensional or three-dimensional coordinate, which can be selected according to the actual situation. For each of the multiple preset reference points, a multi-frame phase difference estimate of that preset reference point is obtained, and the multi-frame phase difference estimate is subjected to time unwrapping processing to obtain multiple unwrapped phase differences. A phase difference fingerprint is constructed based on the multiple unwrapped phase differences to form a database fingerprint. The process of constructing the phase difference fingerprint is the same as the process of constructing the phase difference fingerprint to be located described above, and will not be described again here.

[0092] For example, a database fingerprint database can be represented by the following formula: Formula 22 In the formula, Represents a database fingerprint database; Indicates the first Sine and cosine embedded fingerprints (i.e., phase difference fingerprints) at preset reference points; Indicates the first A validity mask for each preset reference point; Indicates the first The coordinates of a preset reference point; This indicates the total number of preset reference points. The process of determining the validity mask will be described later.

[0093] In one exemplary embodiment, such as Figure 7 Determining a set of candidate reference points from the database fingerprints based on the similarity between the phase difference fingerprint to be located and a pre-built database fingerprint can include: Step S710: Map the phase difference fingerprint to be located as a pair of sine and cosine to construct the fingerprint embedding feature vector of the location to be located.

[0094] This invention takes into account that the original phase difference has a 2π periodicity. If it is directly used as a feature in Euclidean space for matching, discontinuities will occur near -π and π. Therefore, this invention maps each phase difference to a pair of sine and cosine to construct a phase embedding feature, i.e., a fingerprint embedding feature vector.

[0095] The phase difference fingerprint to be located can be mapped to a pair of sine and cosines as shown in the following formula to construct the fingerprint embedding feature vector of the location to be located: Formula 23 In the formula, The fingerprint embedding feature vector represents the location to be located; Indicates the first Phase difference fingerprint of each LED channel to be located; and These represent the cosine function and the sine function, respectively.

[0096] Step S720: Normalize the fingerprint embedding feature vector to obtain the normalized fingerprint embedding vector.

[0097] After obtaining the fingerprint embedding feature vector, it can be normalized. The normalization process can be performed according to the following formula: Formula 24 In the formula, This means updating the left-hand variable with the result on the right. express The Euclidean norm; This indicates a positive number that prevents the denominator from being zero.

[0098] Step S730: Construct a validity mask vector based on the discrimination result.

[0099] The validity mask vector is used to characterize the availability status of the frequency channel corresponding to each LED light-emitting unit for fingerprint matching at the location to be located. When all LED channels are valid, all elements of the mask vector are 1; when some LED channels are invalid, the corresponding elements are 0. As mentioned above, invalid channels are determined based on conditions such as received amplitude, signal-to-noise ratio, phase jump, channel missing, and abnormal sampling, for example, based on the aforementioned preset effective variables of phase difference observation.

[0100] The following formula is the validity mask vector constructed based on the discrimination result: Formula 25 In the formula, Represents the validity mask vector constructed; Indicates the first Validity markers for each LED channel.

[0101] Step S740: For each database fingerprint, determine the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint based on the normalized fingerprint embedding vector and validity mask vector of the database fingerprint, as well as the normalized fingerprint embedding vector and validity mask vector of the location to be located.

[0102] This invention eliminates the discontinuity problem at the phase period boundary by using sine and cosine embedding, preserves the available state of the LED channel by using an effectiveness mask, and reduces the difference in amplitude scale by normalization, so that the phase difference fingerprint has better retrieval, learnability and cross-time stability, thereby ensuring the accuracy of positioning estimation.

[0103] This can be understood as meaning that when locating an object online, the fingerprint database built in the offline phase can be directly accessed.

[0104] The masked weighted distance between the phase difference fingerprint to be located and the fingerprint in the database can be determined according to the following formula: Formula 26 In the formula, This represents the phase difference fingerprint to be located. With the fingerprint database Mask-weighted distance between database fingerprints; Represents the normalized fingerprint embedding vector of the location to be located; Indicates the first Normalized fingerprint embedding vectors of fingerprints from a database; Represents the validity mask vector of the location to be located. And the validity mask vector of database fingerprints A commonly determined effective channel mask; This represents element-wise multiplication of vectors; This represents the Euclidean norm.

[0105] Step S750: Based on the mask weighted distance and the preset temperature coefficient, determine the similarity between the phase difference fingerprint to be located and the fingerprint in the database, so as to determine the set of candidate reference points from the fingerprint in the database according to the similarity.

[0106] The temperature coefficient is used to adjust the rate at which similarity decreases with distance. The similarity between the phase difference fingerprint to be located and the fingerprint in the database can be determined by the following formula: Formula 27 In the formula, This indicates the phase difference fingerprint to be located and the first Similarity between fingerprints in the database; Indicates the mask-weighted distance; Indicates the temperature coefficient; This represents an exponential function.

[0107] After obtaining the similarity between the phase difference fingerprint to be located and the fingerprints in various databases, a predetermined number of candidate reference points can be obtained from each database fingerprint based on the similarity, thus obtaining a set of candidate reference points. The predetermined number can be set according to actual needs, such as 5, 10, 25, etc., without any special limitation. Each candidate reference point can contain information such as fingerprint, coordinates, and validity mask.

[0108] Based on this, by introducing a channel validity mask in the candidate retrieval stage, occlusion, low signal-to-noise ratio, or malfunctioning LED channels will not directly disrupt the similarity calculation, thereby improving the retrieval reliability under local occlusion and non-uniform lighting conditions, and further improving the accuracy of localization estimation.

[0109] In one exemplary embodiment, such as Figure 8 The query center graph structure, constructed based on the candidate reference point set, the location to be located, and the phase difference fingerprint to be located, may include: Step S810: Determine the graph node set based on the candidate reference point set and the location to be located, and determine the directed edges based on the candidate reference points in the candidate reference point set pointed to by the location to be located, so as to obtain the graph edge set.

[0110] This invention can construct a query center graph structure based on a set of candidate reference points, the location to be located, and related information. Specifically, the set of graph nodes can be determined according to the following formula based on the set of candidate reference points and the location to be located: Formula 28 In the formula, This indicates the query node corresponding to the point to be located, i.e., the location to be located. Represents the set of candidate reference points; This indicates the set union operation.

[0111] Directed edges are determined based on the candidate reference points in the candidate reference point set from the location to be located. The graph edge set can be determined according to the following formula: Formula 29 In the formula, Indicates that by query node Pointing to candidate reference point The directed edges; Indicates the number of the candidate reference point.

[0112] Based on this, for the point to be located and its candidate reference point set The initial query center graph structure is constructed as follows: Formula 30 In the formula, Indicates the location to be located The corresponding initial query center graph structure; Represents a set of graph nodes; This represents the set of edges in the graph.

[0113] Step S820: Determine the first initial feature based on the normalized fingerprint embedding vector and validity mask vector of the candidate reference point, the mask-weighted distance and similarity between the phase difference fingerprint to be located and the normalized fingerprint embedding vector corresponding to the candidate reference point, and determine the second initial feature based on the normalized fingerprint embedding vector and validity mask vector corresponding to the location to be located.

[0114] After obtaining the initial query center graph structure, the graph neural network is further refined by constructing corresponding initial features to obtain the final query center graph structure. The first initial feature can be constructed using the following formula: Formula 31 In the formula, Indicates candidate reference point The first initial feature; This represents the normalized fingerprint embedding vector of the candidate reference point; A mask representing the validity of candidate reference points; This represents the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint corresponding to the candidate reference point; This indicates the degree of similarity between the two.

[0115] The second initial feature can be constructed using the following formula: Formula 32 In the formula, Indicates the query node The second initial feature; This represents the normalized fingerprint embedding vector corresponding to the location to be located. Indicates the validity mask corresponding to the location to be located; symbol This indicates vector concatenation.

[0116] Step S830: Determine the query center graph structure based on the graph node set, graph edge set, first initial feature, and second initial feature.

[0117] After obtaining the graph node set, graph edge set, first initial feature, and second initial feature, the query center graph structure can be constructed accordingly.

[0118] Instead of treating candidate reference points as isolated samples with fixed weights, this invention weaves the location to be located and candidate reference points into a query center graph, enabling the graph neural network to simultaneously perceive fingerprint similarity, channel validity, and local candidate structure. This provides a complete and accurate structured input for subsequent adaptive coordinate fusion, thereby improving the accuracy of localization estimation.

[0119] In one exemplary embodiment, such as Figure 9 The location result estimated based on the query center map structure can include: Step S910: For each candidate reference point, based on a preset number of attention heads, determine the attention score between the location to be located and the candidate reference point according to the first initial feature and the second initial feature in the query center map structure.

[0120] In the query center graph structure, the contribution weights of candidate reference points to the positioning structure can be learned using graph attention message passing. The preset number of attention heads serves as a parameter of the graph neural network structure and can be preset or tuned using a validation set. For example, it can be determined based on the number of candidate reference points, fingerprint dimension, positioning accuracy requirements, and computational resources. In practical implementation, 2, 4, or 8 attention heads can be used.

[0121] Based on this, and using a preset number of attention heads, the attention score between the location to be located and the candidate reference point is determined according to the first and second initial features in the query center map structure, as follows: Formula 33 In the formula, Indicates the first Query node in each attention head With candidate reference points The unnormalized attention score between them, i.e., the attention score; This represents the activation function for a leaky linear rectified circuit. Indicates the first The learnable attention parameter vector of each attention head; Indicates the first A learnable linear transformation matrix for each attention head; Indicates the second initial feature; The first initial feature representing the candidate reference point; This indicates vector concatenation.

[0122] Step S920: Normalize the attention scores corresponding to the candidate reference points to obtain the attention coefficients.

[0123] After obtaining the attention scores corresponding to the candidate reference points, normalization can be performed using the following formula: Formula 34 In the formula, Indicates the first Candidate reference points in each attention focus For query nodes The normalized attention coefficient, i.e., the attention coefficient; This indicates an unnormalized attention score; Represents the set of candidate reference points Any node number in the array; This represents an exponential function.

[0124] Step S930: Perform feature aggregation based on the preset quantity, attention coefficient, and first initial feature to obtain query context features.

[0125] After obtaining the attention coefficient, feature aggregation can be performed based on the preset quantity, attention coefficient, and first initial feature to obtain the query context features according to the following formula: Formula 35 In the formula, Indicates query context characteristics; Indicates the number of attention heads; Indicates will The outputs of each attention head are spliced ​​together; Represents a nonlinear activation function; Indicates the attention coefficient; Represents a learnable linear transformation matrix; This represents the first initial feature of the candidate reference node.

[0126] Step S940: Calculate the fusion score of the candidate reference point based on the query context features, the first initial features, and the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint corresponding to the candidate reference point, and determine the fusion weight of the candidate reference point based on the fusion score.

[0127] After obtaining the query context features, the fusion score of the candidate reference point can be calculated using the following formula: based on the query context features, the first initial features, and the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint corresponding to the candidate reference point. Formula 36 In the formula, This represents the fusion score of the candidate reference point; This represents a score prediction function composed of a multilayer perceptron or a lightweight regression network. Indicates query context characteristics; The first initial feature representing the candidate reference point; This represents the mask-weighted distance between the candidate reference point and the query point.

[0128] Furthermore, after obtaining the fusion scores of the candidate reference points, the fusion weight of each candidate reference point r can be determined based on the fusion scores using the following formula: Formula 37 In the formula, Indicates candidate reference point The fusion weights; Indicates candidate reference point The fusion score; Indicates candidate reference point The fusion score.

[0129] Step S950: Determine the positioning result of the location to be located based on the fusion weight and coordinate information of each candidate reference point.

[0130] After obtaining the fusion weights of each candidate reference point, the positioning result of the location to be located can be determined according to the following formula based on the fusion weights and coordinate information of each candidate reference point: Formula 38 In the formula, This represents the positioning result of the location to be located, q. Indicates candidate reference point The fusion weights; Indicates candidate reference point The known coordinates.

[0131] For example, such as Figure 10 The diagram illustrates a localization method based on a query center graph structure. For each candidate reference point (candidate reference fingerprint), based on a preset number of attention heads, the attention score between the location to be located (fingerprint to be located) and the candidate reference point is determined according to the first and second initial features in the query center graph structure. Then, the attention scores corresponding to the candidate reference points are normalized to obtain attention coefficients. Next, feature aggregation is performed based on the preset number of attention heads, the attention coefficients, and the first initial features to obtain query context features. Based on the query context features, the first initial features, and the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint corresponding to the candidate reference point, the fusion score of the candidate reference point is calculated, and the fusion weight of the candidate reference point is determined based on the fusion score. Finally, based on the fusion weights and coordinate information of each candidate reference point, the localization result of the location to be located is determined. .

[0132] This invention learns adaptive weights for candidate reference points through graph neural networks. Compared with the fixed distance metric and fixed weight rules in traditional K-nearest neighbors or weighted K-nearest neighbors, the positioning process can automatically suppress mismatched candidate reference points based on the local fingerprint structure, the relationship between candidate reference points and physical similarity. This improves the robustness of positioning in complex layouts such as corridors, corners, intersections and underground caverns, and improves the positioning accuracy in these environmental scenarios.

[0133] Furthermore, it should be noted that this invention can also use the collected reference points or verification trajectory samples to supervise the training of the graph neural network during the training phase. In actual online deployment, the photoelectric receiving module collects composite visible light signals in real time, extracts the phase difference fingerprint to be located, performs retrieval, constructs a query center map, executes the graph neural network forward inference, and outputs the location of the mobile carrier to be located. This process does not require real-time synchronization between multiple illumination units, nor does it require complex multi-base station geometric calculations during the online phase. For example, the following loss function can be used during the training phase: Formula 39 In the formula, Represents the training loss function; Indicates the location to be located The estimated location; Indicates the location to be located The calibration position; Indicates the squared Euclidean error; This represents the weight distribution adjustment coefficient; Indicates candidate reference point The fusion weights; Represents the natural logarithm function; This prevents the logarithmic function from receiving a positive number that is zero. Represents the weight of similarity consistency constraints; This represents the mask-weighted distance between the fingerprint to be located and the fingerprints of the candidate reference points.

[0134] In the loss function, the first term is used to constrain the positioning accuracy, the second term is used to adjust the distribution of candidate weights to form an interpretable contribution relationship of candidate reference points, and the third term is used to suppress excessive weighting of fingerprints of candidate reference points that are far away or have low similarity.

[0135] The LED array-based positioning method in the exemplary embodiments of the present invention, on the one hand, utilizes the LED array within a single lighting unit to construct the phase difference fingerprint to be located. This eliminates the need for strict synchronization and complex geometric coordination among multiple luminaires, reducing the workload of multi-base station mapping, calibration, synchronization, and maintenance, lowering deployment costs and engineering maintenance complexity, and making it more suitable for large-scale deployment in long-distance scenarios such as tunnels, engineering corridors, and integrated utility tunnels. Furthermore, by setting distinguishable modulation frequencies for different LED light-emitting units within the same lighting unit and extracting the phase difference estimate within the LED array, a single lighting unit is no longer just a regular lighting point, but forms an array positioning base station with spatial coding capabilities, thereby enhancing the observability of single-lamp positioning. On the other hand, the phase difference fingerprint to be located belongs to the relative phase characteristics between different light-emitting units within the LED array, which can, to a certain extent, weaken the impact of overall light source brightness changes, receiver gain changes, slowly varying ambient light interference, and some receiver intensity fluctuations on the positioning results, thus improving positioning stability in complex lighting environments such as underground areas and tunnels. Furthermore, by unwrapping the phase and constructing the phase difference fingerprint to be located, the impact of the original phase period jump on subsequent fingerprint matching is reduced, enabling the phase difference fingerprint to be stored and matched in a more stable and continuous form, thereby improving the accuracy of positioning.

[0136] In an exemplary embodiment of the present invention, a positioning system based on an LED array is also provided. (See reference...) Figure 11 As shown, the LED array-based positioning system 1100 may include a photoelectric receiving module 1110, a signal processing module 1120, a phase difference estimation module 1130, a phase difference fingerprint construction module 1140, and a positioning module 1150. Specifically: The photoelectric receiving module 1110 is used to acquire the composite visible light signal emitted by the LED array received by the mobile carrier at the location to be positioned. The LED array is obtained by integrating multiple LED light-emitting units from the same illumination unit, and the initial modulation frequencies corresponding to each LED light-emitting unit are mutually exclusive. The signal processing module 1120 is used to filter and down-convert each initial modulation frequency obtained by frequency decomposition of the composite visible light signal to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency. The phase difference estimation module 1130 is used to determine a reference light-emitting unit among the multiple LED light-emitting units, construct a first analytical signal based on the channel signal with the unified frequency corresponding to the reference light-emitting unit, and construct a second analytical signal based on the unified frequency corresponding to the remaining light-emitting units among the multiple LED light-emitting units. The channel signals are used to construct a second analytical signal, and the phase difference estimate between multiple LED light-emitting units inside the LED array is determined based on the first analytical signal and the second analytical signal. The phase difference fingerprint construction module 1140 is used to obtain multi-frame phase difference estimates of the position to be located, perform time unwrapping processing on the multi-frame phase difference estimates to obtain multiple unwrapped phase differences, and construct the phase difference fingerprint to be located based on the multiple unwrapped phase differences. The positioning module 1150 is used to determine a set of candidate reference points from the database fingerprint based on the similarity between the phase difference fingerprint to be located and the pre-constructed database fingerprint, and construct a query center graph structure based on the set of candidate reference points, the position to be located and the phase difference fingerprint to be located, so as to estimate the positioning result of the position to be located based on the query center graph structure.

[0137] For example, such as Figure 12 This is a schematic diagram of another LED array-based positioning system. In this system, the LED array and modulation driving module constitute a single-lamp LED positioning base station 1210. The mobile carrier (i.e., positioning terminal) 1220 includes a photoelectric receiving module, a signal processing module, and a data communication module. The positioning calculation and application platform 1230 may include a phase difference fingerprint construction module, a positioning module, an application interface, and a display module. The data communication module allows the mobile carrier to communicate with the single-lamp LED array positioning base station and the positioning calculation and application platform. The application interface and display module provide visualization functionality.

[0138] Since the details of each functional module of the LED array-based positioning system of the exemplary embodiment of the present invention have been described in the exemplary embodiment of the LED array-based positioning method described above, they will not be repeated here.

[0139] It should be noted that although several modules or units of the LED array-based positioning system have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0140] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. The invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

Claims

1. A method of positioning based on an array of LEDs, characterized by, include: The composite visible light signal emitted by the LED array received by the mobile carrier at the location to be positioned is obtained. The LED array is obtained by integrating multiple LED light-emitting units from the same lighting unit, and the initial modulation frequencies corresponding to each LED light-emitting unit do not overlap. The composite visible light signal is frequency decomposed to obtain each initial modulation frequency, which is then filtered and down-converted to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency. A reference light-emitting unit is determined among the plurality of LED light-emitting units, and a first analytical signal is constructed based on the channel signal of the reference light-emitting unit after unification of frequency. A second analytical signal is constructed based on the channel signal of the remaining light-emitting units among the plurality of LED light-emitting units after unification of frequency, so as to determine the phase difference estimate between the plurality of LED light-emitting units inside the LED array based on the first analytical signal and the second analytical signal. The phase difference estimates of the location to be located are obtained in multiple frames. The phase difference estimates of the multiple frames are subjected to time unwrapping processing to obtain multiple unwrapped phase differences. The phase difference fingerprint of the location to be located is constructed based on the multiple unwrapped phase differences. Based on the similarity between the phase difference fingerprint to be located and the pre-constructed database fingerprint, a set of candidate reference points is determined from the database fingerprint. A query center graph structure is constructed based on the set of candidate reference points, the location to be located, and the phase difference fingerprint to be located, so as to estimate the location result of the location to be located based on the query center graph structure.

2. The method of claim 1, wherein, The plurality of LED light-emitting units are controlled by a unified clock from the same driving control module. Each LED light-emitting unit corresponds to a frequency channel of a receiver on the mobile carrier, and the receiver is used to receive the composite visible light signal. The method further includes: For each LED light-emitting unit, the initial modulation frequency of the LED light-emitting unit is determined according to the preset basic modulation frequency and the odd frequency sequence number corresponding to the LED light-emitting unit; The odd frequency sequence number is predetermined based on the number of the LED light-emitting unit in the plurality of LED light-emitting units or the number of the driving channel, and the odd frequency sequence number does not change with the position of the moving carrier.

3. The method of claim 2, wherein, The process of filtering and down-converting each initial modulation frequency obtained by frequency decomposition of the composite visible light signal to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency includes: For each LED light-emitting unit, the equivalent amplitude of the initial modulation frequency corresponding to the LED light-emitting unit is obtained; wherein, the equivalent amplitude is used to characterize the received strength and signal quality corresponding to the LED light-emitting unit; Obtain the equivalent fixed phase and narrowband noise corresponding to the initial modulation frequency; Based on the equivalent amplitude, the equivalent fixed phase, and the narrowband noise, the initial modulation frequency corresponding to the LED light-emitting unit is narrowband filtered to obtain the narrowband signal corresponding to the LED light-emitting unit. The narrowband signals corresponding to each LED light-emitting unit are down-converted to a preset base frequency to obtain the channel signals of each LED light-emitting unit after unification of frequency.

4. The method of claim 1, wherein, The step of constructing a first analytical signal based on the channel signal of the reference light-emitting unit at a unified frequency, and constructing second analytical signals based on the channel signals of the remaining light-emitting units among the plurality of LED light-emitting units at a unified frequency, to determine the phase difference estimate between the plurality of LED light-emitting units based on the first analytical signal and the second analytical signal, includes: Obtain the first Hilbert transform operator of the channel signal with a unified frequency corresponding to the reference light-emitting unit, and obtain the second Hilbert transform operator of the channel signal with a unified frequency corresponding to each of the remaining light-emitting units; The first analytical signal is determined based on the channel signal of the reference light-emitting unit after unification of frequency and the first Hilbert transform operator; For each of the remaining light-emitting units, the second analytical signal is determined based on the channel signal with the unified frequency corresponding to the remaining light-emitting unit and the second Hilbert transform operator; Based on the first analytical signal and the second analytical signal, the phase difference estimate between the plurality of LED light-emitting units is determined.

5. The method of claim 4, wherein, The step of determining the phase difference estimate between the plurality of LED light-emitting units based on the first analytical signal and the second analytical signal includes: For each of the LED light-emitting units, obtain the complex conjugate of the first analytical signal; Based on the second analytical signal and the complex conjugate, the phase difference estimate of the LED light-emitting unit relative to the reference light-emitting unit is determined; Based on the phase difference estimate of each LED light-emitting unit relative to the reference light-emitting unit, the phase difference estimate between the plurality of LED light-emitting units is determined.

6. The method of claim 1, wherein, The process of obtaining multiple frames of phase difference estimates for the location to be located, performing time unwrapping processing on the multiple frames of phase difference estimates to obtain multiple unwrapped phase differences, and constructing a phase difference fingerprint for the location to be located based on the multiple unwrapped phase differences includes: Based on the phase difference estimation of each frame, determine the phase difference vector representation corresponding to the phase difference estimation of each frame; Based on the phase difference vector representation corresponding to each frame, the temporal unwrapping of consecutive sampled frames is performed to obtain the multiple unwrapped phase differences; Based on the preset effective variables for phase difference observation, the observation effectiveness of each unwrapped phase difference is judged, and the phase difference that passes the effectiveness screening is determined based on the judgment result. Based on the phase differences that have passed the validity screening, construct the phase difference fingerprint of the target location; The validity determination of the observed phase difference after unwrapping includes validity determination based on one or more of amplitude, signal-to-noise ratio, frequency, phase and noise.

7. The method of claim 6, wherein, The step of determining a set of candidate reference points from the database fingerprint based on the similarity between the phase difference fingerprint to be located and a pre-built database fingerprint includes: The phase difference fingerprint to be located is mapped into a pair of sine and cosine to construct the fingerprint embedding feature vector of the location to be located. The fingerprint embedding feature vector is normalized to obtain the normalized fingerprint embedding vector; A validity mask vector is constructed based on the discrimination result. The validity mask vector is used to characterize whether the frequency channel corresponding to each LED light-emitting unit is available for fingerprint matching at the location to be located. For each of the database fingerprints, the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint is determined based on the normalized fingerprint embedding vector and validity mask vector of the database fingerprint, as well as the normalized fingerprint embedding vector and validity mask vector of the location to be located. Based on the mask weighted distance and the preset temperature coefficient, the similarity between the phase difference fingerprint to be located and the fingerprint in the database is determined, so as to determine a set of candidate reference points from the fingerprint in the database according to the similarity. The temperature coefficient is used to adjust the rate at which similarity decreases with distance.

8. The method of claim 7, wherein, The step of constructing a query center graph structure based on the candidate reference point set, the location to be located, and the phase difference fingerprint to be located includes: Based on the set of candidate reference points and the location to be located, a set of graph nodes is determined, and directed edges are determined based on the location to be located pointing to candidate reference points in the set of candidate reference points, so as to obtain a set of graph edges. Based on the normalized fingerprint embedding vector and validity mask vector of the candidate reference point, the mask-weighted distance and similarity between the phase difference fingerprint to be located and the normalized fingerprint embedding vector corresponding to the candidate reference point, a first initial feature is determined, and a second initial feature is determined based on the normalized fingerprint embedding vector and validity mask vector corresponding to the location to be located. The query center graph structure is determined based on the graph node set, the graph edge set, the first initial feature, and the second initial feature.

9. The method of claim 8, wherein, The estimation of the location result of the position to be located based on the query center map structure includes: For each candidate reference point, based on a preset number of attention heads, the attention score between the location to be located and the candidate reference point is determined according to the first initial feature and the second initial feature in the query center graph structure. The attention scores corresponding to the candidate reference points are normalized to obtain attention coefficients; Based on the preset quantity, the attention coefficient, and the first initial feature, feature aggregation is performed to obtain query context features; Based on the query context features, the first initial features, and the mask-weighted distance between the phase difference fingerprint to be located and the database fingerprint corresponding to the candidate reference point, the fusion score of the candidate reference point is calculated, and the fusion weight of the candidate reference point is determined based on the fusion score. Based on the fusion weights and coordinate information of each candidate reference point, the positioning result of the location to be located is determined.

10. The method according to any one of claims 1 to 9, characterized in that, The method further includes: Multiple preset reference points are determined within the carrier's workspace; The database fingerprint is constructed in advance based on the normalized fingerprint embedding vector and validity mask vector corresponding to each preset reference point and the coordinate information of the preset reference point.

11. An LED array based positioning system, characterized by include: The photoelectric receiving module is used to acquire the composite visible light signal emitted by the LED array received by the mobile carrier at the location to be positioned. The LED array is obtained by integrating multiple LED light-emitting units from the same lighting unit, and the initial modulation frequencies corresponding to each LED light-emitting unit do not overlap. The signal processing module is used to filter and down-convert each initial modulation frequency obtained by frequency decomposition of the composite visible light signal to obtain a channel signal with a unified frequency corresponding to each initial modulation frequency. A phase difference estimation module is used to determine a reference light-emitting unit among the plurality of LED light-emitting units, and construct a first analytical signal based on the channel signal of the reference light-emitting unit after unification of frequency, and construct a second analytical signal based on the channel signal of the remaining light-emitting units among the plurality of LED light-emitting units after unification of frequency, so as to determine the phase difference estimate between the plurality of LED light-emitting units inside the LED array based on the first analytical signal and the second analytical signal. A phase difference fingerprint construction module is used to obtain multiple frames of phase difference estimates of the location to be located, perform time unwrapping processing on the multiple frames of phase difference estimates to obtain multiple unwrapped phase differences, and construct a phase difference fingerprint of the location to be located based on the multiple unwrapped phase differences. The localization module is used to determine a set of candidate reference points from the database fingerprint based on the similarity between the phase difference fingerprint to be located and the pre-built database fingerprint, and to construct a query center graph structure based on the set of candidate reference points, the location to be located, and the phase difference fingerprint to be located, so as to estimate the localization result of the location to be located based on the query center graph structure.

Images