Vehicle positioning method and device based on fingerprint positioning, and electronic equipment

Abstract

The invention discloses a vehicle positioning method and device based on fingerprint positioning and electronic equipment, and relates to the 5G technical field or other related technical fields, and the method comprises the steps: employing a plurality of receiving antennas of a base station to detect the characteristics of an uplink signal transmitted by a to-be-positioned cash truck, generating a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-time delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, and matching the extracted angle-time delay channel frequency power fingerprint matrix ADCFP with fingerprints in a pre-established fingerprint database to obtain a matching result; and according to a matching result, positioning the position coordinates of the to-be-positioned cash truck by adopting an improved power WKNN matching algorithm. The technical problem that in the prior art, an armor cash carrier is influenced by different environmental factors, the positioning precision is low, and the vehicle safety is reduced is solved.

Description

Technical Field

[0001] The present invention relates to the field of 5G technology, financial technology or other related fields, and specifically to a vehicle positioning method and device based on fingerprint positioning, and electronic equipment. Background Art

[0002] With the rapid development of the market economy and the continuous growth of the financial industry, the increase in cash flow and the expansion of financial institutions' outlets have put forward higher requirements for the safety management of cash transport vehicles. The safety of cash transport vehicles on the way, especially the real-time monitoring of their precise location, has become a key point for the security precautions of financial institutions and banks. Traditional cash transport vehicle positioning systems mostly rely on the GPS global positioning system, but because the GPS signal is easily affected by environmental factors such as occlusion, multipath effect and urban canyons, its positioning accuracy is difficult to meet high security requirements in some cases.

[0003] In the related technology, a positioning system based on signal fingerprints is proposed, which uses the received signal features to compare with the pre-established fingerprint database to achieve accurate positioning of the target. However, the current fingerprint positioning technology still has some drawbacks in practical applications. In the fingerprint positioning process, the matching of fingerprint information is a key step that affects the positioning speed. Traditional fingerprint matching algorithms, such as the k-nearest neighbor (KNN) algorithm, although simple and intuitive, have low search efficiency in large-scale databases and are not able to distinguish between near and far fingerprints, affecting positioning accuracy. In addition, in complex environments such as multipath propagation and non-line-of-sight (NLOS), the extraction and matching of signal features becomes more difficult, which requires the positioning algorithm to effectively process multipath information and reduce the impact of environmental factors on positioning accuracy.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a vehicle positioning method and device based on fingerprint positioning, and electronic equipment to at least solve the technical problem in the related art that the cash transport vehicle is affected by different environmental factors, the positioning accuracy is low, and the vehicle safety is reduced.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a vehicle positioning method based on fingerprint positioning is provided, comprising: using multiple receiving antennas of a base station to detect the uplink signal characteristics emitted by a cash transport vehicle to be positioned, generating a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized as the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP contains the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be positioned to reach the receiving antenna of the base station, which is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation; matching the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprint in the pre-established fingerprint database to obtain a matching result; and locating the position coordinates of the cash transport vehicle to be positioned using the improved power WKNN matching algorithm according to the matching result.

[0007] Optionally, the step of using multiple receiving antennas of a base station to detect uplink signal characteristics emitted by the cash transport vehicle to be located includes: using multiple receiving antennas in the base station to detect signals emitted by the cash transport vehicle to be located, and obtaining uplink signal characteristics of the cash transport vehicle to be located, wherein the base stations are arranged according to a uniform linear array model, the number of receiving antennas in each base station is greater than the number of cash transport vehicles to be located distributed in the area to which the base station belongs, and the fingerprint information of the user to be located on the cash transport vehicle to be located is transmitted to the receiving antenna along multiple paths.

[0008] Optionally, the fingerprint database is constructed in the following manner: collecting historical signal features at known positions, wherein the historical signal features include: a historical channel impulse response fingerprint matrix CIR, a historical channel frequency response fingerprint matrix CFR, and a historical angle-delay channel frequency power fingerprint matrix ADCFP improved based on the CFR fingerprint; extracting fingerprint information from the preprocessed historical signal features to generate a fingerprint matrix, wherein the fingerprint matrix includes the historical angle-delay channel frequency power fingerprint matrix ADCFP and the corresponding relationship between the coordinates of the corresponding known position; storing the fingerprint matrix in a preset database to obtain the fingerprint database, wherein the fingerprint database stores the corresponding relationship between each reference fingerprint and the corresponding position coordinates.

[0009] Optionally, the step of collecting historical signal features at a known position includes: performing channel estimation on historical uplink signals received in the historical process to obtain the historical channel impulse response fingerprint matrix CIR, wherein the matrix size of the historical channel impulse response fingerprint matrix CIR is determined by the number of receiving antennas of the base station and the length of the channel impulse response; extracting arrival angle AOA and arrival time TOA information from the historical channel impulse response fingerprint matrix CIR; converting the historical channel impulse response fingerprint matrix CIR into the historical channel frequency response fingerprint matrix CFR through Fourier transform to form frequency domain signal features, wherein the matrix size of the historical channel frequency response fingerprint matrix CFR is determined by the number of receiving antennas and the number of transmitting antennas of the base station; using a phase shift matrix to transform the historical channel frequency response fingerprint matrix CFR from the frequency domain to the angle domain to obtain a historical angle-delay channel frequency response matrix ADCFR, wherein the phase shift matrix is ​​obtained based on a uniform linear array model corresponding to the antenna array in the base station; performing a preset inner product operation on the historical angle-delay channel frequency response matrix ADCFR to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP.

[0010] Optionally, the step of performing a preset inner product operation on the historical angle-delay channel frequency response ADCFR to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP includes: performing multiplication operations on each row and each column element in the historical angle-delay channel frequency response matrix ADCFR with corresponding elements of the phase shift matrix, and summing the product values ​​to obtain each row and each column element of the historical angle-delay channel frequency power fingerprint matrix ADCFP; performing a Hadamard inner product operation on the historical angle-delay channel frequency power fingerprint matrix ADCFP to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP, wherein the historical angle-delay channel frequency response matrix ADCFR is a complex matrix, and the historical angle-delay channel frequency power fingerprint matrix ADCFP is a real matrix ADCFP.

[0011] Optionally, after collecting historical signal features at known positions, it also includes: compressing the ADCFP fingerprint matrix row by row to obtain a horizontal channel frequency power matrix HCFP, and compressing the ADCFP fingerprint matrix column by column to obtain a vertical channel frequency power matrix VCFP; combining the horizontal channel frequency power matrix HCFP and the vertical channel frequency power matrix VCFP to generate a horizontal vertical channel frequency fingerprint matrix HVCFF.

[0012] Optionally, after storing the fingerprint matrix in a preset database to obtain the fingerprint database, it also includes: using a central arrival angle clustering algorithm to divide the fingerprint database into intervals to obtain multiple cluster intervals; calculating the central arrival angle CAOA of each reference fingerprint in the fingerprint database; and dividing the reference fingerprint into corresponding cluster intervals according to the size of the central arrival angle CAOA of each reference fingerprint.

[0013] Optionally, the step of matching the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprints in a pre-established fingerprint database to obtain a matching result includes: calculating the similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database, and obtaining the matching result based on the similarity, wherein Manhattan distance or Euclidean distance is used as the similarity in the process of calculating the similarity.

[0014] Optionally, based on the matching results, the step of using an improved power WKNN matching algorithm to locate the position coordinates of the armored car to be located includes: sorting the similarities in the matching results, and selecting K fingerprints with the highest similarity as candidate matching fingerprints, wherein the K value is an integer greater than 1; using an improved power WKNN matching algorithm, assigning different weighting coefficients to each candidate matching fingerprint based on the distance between the candidate matching fingerprint and the fingerprint of the point to be located; and calculating the weighted average of the position coordinates corresponding to the candidate matching fingerprints based on the weighting coefficients as the position coordinates of the armored car to be located.

[0015] According to another aspect of an embodiment of the present invention, a vehicle positioning device based on fingerprint positioning is also provided, comprising: a fingerprint matrix generating unit, which is used to use multiple receiving antennas of a base station to detect the uplink signal characteristics emitted by the cash transport vehicle to be positioned, and generate a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, and the channel frequency response fingerprint matrix ADCFP is obtained by performing Fourier transform on the CIR fingerprint. The fingerprint matrix CFR is characterized by the sum of the CIRs of each path under different delays. The angle-delay channel frequency power fingerprint matrix ADCFP contains the channel power information and arrival angle AOA information of the signal sent by the to-be-located cash transport vehicle to the base station receiving antenna, which is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation; the fingerprint matching unit is used to match the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprint in the pre-established fingerprint database to obtain a matching result; the vehicle positioning unit is used to locate the position coordinates of the to-be-located cash transport vehicle according to the matching result by using the improved power WKNN matching algorithm.

[0016] Optionally, the fingerprint matrix generation unit includes: a vehicle signal detection module, which is used to use multiple receiving antennas in the base station to detect the signal emitted by the money transport vehicle to be located, and obtain the uplink signal characteristics of the money transport vehicle to be located, wherein the base stations are arranged according to a uniform linear array model, and the number of receiving antennas in each base station is greater than the number of money transport vehicles to be located distributed in the area to which the base station belongs, and the fingerprint information of the user to be located on the money transport vehicle to be located is transmitted to the receiving antenna along multiple paths.

[0017] Optionally, when constructing the fingerprint database, the vehicle positioning device based on fingerprint positioning further includes: a signal feature acquisition unit, used to collect historical signal features at known positions, wherein the historical signal features include: a historical channel impulse response fingerprint matrix CIR, a historical channel frequency response fingerprint matrix CFR, and a historical angle-delay channel frequency power fingerprint matrix ADCFP improved based on the CFR fingerprint; a fingerprint information extraction unit, used to extract fingerprint information from the pre-processed historical signal features to generate a fingerprint matrix, wherein the fingerprint matrix includes the historical angle-delay channel frequency power fingerprint matrix ADCFP and the corresponding relationship between the coordinates of the known position; a fingerprint matrix storage unit, used to store the fingerprint matrix in a preset database to obtain the fingerprint database, wherein the fingerprint database stores the corresponding relationship between each reference fingerprint and the corresponding position coordinates.

[0018] Optionally, the signal feature acquisition unit includes: a channel estimation module, which is used to perform channel estimation on the historical uplink signals received in the historical process to obtain the historical channel impulse response fingerprint matrix CIR, wherein the matrix size of the historical channel impulse response fingerprint matrix CIR is determined by the number of receiving antennas of the base station and the length of the channel impulse response; a matrix information extraction module, which is used to extract the arrival angle AOA and arrival time TOA information from the historical channel impulse response fingerprint matrix CIR; a matrix conversion module, which is used to convert the historical channel impulse response fingerprint matrix CIR into the historical channel frequency response fingerprint matrix CFR through Fourier transform, forming The invention relates to a frequency domain signal feature, wherein the matrix size of the historical channel frequency response fingerprint matrix CFR is determined by the number of receiving antennas and the number of transmitting antennas of the base station; a matrix frequency conversion module is used to transform the historical channel frequency response fingerprint matrix CFR from the frequency domain to the angle domain by using a phase shift matrix to obtain a historical angle-delay channel frequency response matrix ADCFR, wherein the phase shift matrix is ​​obtained based on a uniform linear array model corresponding to the antenna array in the base station; an inner product module is used to perform a preset inner product operation on the historical angle-delay channel frequency response matrix ADCFR to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP.

[0019] Optionally, the inner product module includes: a product submodule, used to multiply each row and each column element in the historical angle-delay channel frequency response matrix ADCFR with the corresponding element of the phase shift matrix, and sum each product value to obtain each row and each column element of the historical angle-delay channel frequency power fingerprint matrix ADCFP; an inner product submodule, used to perform a Hadamard inner product operation on the historical angle-delay channel frequency power fingerprint matrix ADCFP to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP, wherein the historical angle-delay channel frequency response matrix ADCFR is a complex matrix, and the historical angle-delay channel frequency power fingerprint matrix ADCFP is a real matrix ADCFP.

[0020] Optionally, the vehicle positioning device based on fingerprint positioning also includes: a fingerprint matrix compression unit, which is used to compress the ADCFP fingerprint matrix row by row to obtain the horizontal channel frequency power matrix HCFP after collecting historical signal features at a known position, and compress the ADCFP fingerprint matrix column by column to obtain the vertical channel frequency power matrix VCFP; a matrix combination unit, which is used to combine the horizontal channel frequency power matrix HCFP and the vertical channel frequency power matrix VCFP to generate a horizontal and vertical channel frequency fingerprint matrix HVCFF.

[0021] Optionally, the vehicle positioning device based on fingerprint positioning also includes: an interval division unit, which is used to store the fingerprint matrix in a preset database to obtain the fingerprint database, and then use a central arrival angle clustering algorithm to divide the fingerprint database into intervals to obtain multiple cluster intervals; a fingerprint parameter calculation unit, which is used to calculate the central arrival angle CAOA of each reference fingerprint in the fingerprint database; a fingerprint interval division unit, which is used to divide the reference fingerprint into corresponding cluster intervals according to the size of the central arrival angle CAOA of each reference fingerprint.

[0022] Optionally, the fingerprint matching unit includes: a similarity calculation module, used to calculate the similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database, and obtain the matching result based on the similarity, wherein Manhattan distance or Euclidean distance is used as the similarity in the process of calculating the similarity.

[0023] Optionally, the vehicle positioning unit includes: a similarity sorting module, which is used to sort the similarities in the matching results and select the K fingerprints with the highest similarity as candidate matching fingerprints, wherein the K value is an integer greater than 1; a weighting module, which is used to adopt an improved power WKNN matching algorithm to assign different weighting coefficients to each candidate matching fingerprint according to the distance between the candidate matching fingerprint and the fingerprint of the point to be positioned; and a positioning module, which is used to calculate the weighted average of the position coordinates corresponding to the candidate matching fingerprints according to the weighting coefficients as the position coordinates of the armored car to be positioned.

[0024] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned vehicle positioning methods based on fingerprint positioning.

[0025] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-mentioned vehicle positioning methods based on fingerprint positioning.

[0026] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of any one of the above-mentioned vehicle positioning methods based on fingerprint positioning are implemented.

[0027] In the present disclosure, multiple receiving antennas of a base station are used to detect the characteristics of an uplink signal emitted by a cash transport vehicle to be located, and a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP are generated based on the characteristics of the uplink signal, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized as the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP includes the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be located to the receiving antenna of the base station, which is obtained by the CFR fingerprint through a phase shift matrix and an inner product transformation, and the extracted angle-delay channel frequency power fingerprint matrix ADCFP is matched with the fingerprint in a pre-established fingerprint database to obtain a matching result, and according to the matching result, an improved power WKNN matching algorithm is used to locate the position coordinates of the cash transport vehicle to be located.

[0028] According to the above-mentioned contents of the present disclosure, the ADCFP fingerprint matrix can be used to utilize the concentrated characteristics of the channel power in the angle domain and the multipath propagation information to more accurately describe the position characteristics of the money transporter on the way, thereby realizing fingerprint matching and obtaining the real-time and accurate positioning of the money transporter. At the same time, the remote monitoring center can also understand the accurate position and driving conditions of the money transporter in real time. In complex multipath and non-line-of-sight environments, the fingerprint based on angle and power can effectively utilize the multipath effect, further enhance the accuracy of positioning, reduce unexpected yaw or delay, and reduce the risk of property loss caused by inaccurate position information of the money transporter, thereby solving the technical problem in the related technology that the money transporter is affected by different environmental factors, the positioning accuracy is low, and the vehicle safety is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0030] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing a vehicle positioning method based on fingerprint positioning is shown;

[0031] Figure 2 is a flow chart of an optional vehicle positioning method based on fingerprint positioning according to an embodiment of the present invention;

[0032] Figure 3 It is a schematic diagram of an optional model for building a 5G massive MIMO-OFDM fingerprint positioning system according to an embodiment of the present invention;

[0033] Figure 4 is a schematic diagram of a cash transport vehicle positioning method based on fingerprint positioning according to an embodiment of the present invention;

[0034] Figure 5 is a schematic diagram of an optional transformation from a CFR fingerprint to an ADCFP fingerprint matrix according to an embodiment of the present invention;

[0035] Figure 6 is a schematic diagram of an optional compression process for a fingerprint matrix according to an embodiment of the present invention;

[0036] Figure 7 is a schematic diagram of an optional vehicle positioning device based on fingerprint positioning according to an embodiment of the present invention;

[0037] Figure 8 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0039] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0040] To facilitate those skilled in the art to understand the present invention, some terms or nouns involved in the embodiments of the present invention are explained below:

[0041] Multiple Input Multiple Output (MIMO) is a system that uses multiple antennas at the transmitting and receiving ends to improve the rate and reliability of data transmission.

[0042] Orthogonal Frequency Division Multiplexing, referred to as OFDM, is a technology that decomposes a high-speed data stream into multiple low-speed sub-streams and transmits them simultaneously on multiple orthogonal sub-carriers, thereby improving spectrum utilization and resistance to multipath propagation.

[0043] The MIMO-OFDM system model integrates large-scale multiple-input multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) technology. Large-scale MIMO technology deploys a large number of receiving antennas at the base station to collect richer multipath information, while OFDM technology allows for more sophisticated signal analysis in the frequency domain. In the present invention, the positioning system based on signal fingerprints can use the received signal characteristics to compare with the pre-established fingerprint database to achieve accurate positioning of the target. In the 5G communication environment, the large-scale MIMO-OFDM system model can significantly improve the system throughput and positioning accuracy.

[0044] CIR fingerprint, Channel Impulse Response, channel impulse response fingerprint, is a time delay and strength characteristic of a wireless signal when it reaches the receiver through multipath propagation. The channel impulse response is a direct reflection of the multipath propagation effect of the signal caused by environmental factors (such as buildings, terrain, etc.) when the signal is transmitted in the channel. In the MIMO-OFDM system provided by the present invention, by extracting and analyzing CIR, detailed information about the signal propagation path can be obtained, including arrival time (TOA), arrival angle (AOA), etc., and this information can be used as the basic fingerprint for positioning.

[0045] CFR fingerprint, where CFR is the abbreviation of Channel Frequency Response, is a frequency domain feature converted from CIR fingerprint. In the present invention, it is obtained by Fourier transforming the CIR in the time domain, reflecting the response characteristics of the channel to signals of different frequencies. The extraction of CFR fingerprint can further enhance the description ability of signal characteristics, especially in a multipath environment, where signals of different frequencies will show different attenuation and phase changes due to the multipath effect, which provides additional information for the positioning algorithm and helps to improve positioning accuracy.

[0046] ADCFP fingerprint matrix, where ADCFP is the abbreviation of Angle Delay Channel Frequency Power, is a fingerprint matrix that combines the arrival angle, delay and frequency domain channel power. In the present invention, based on the CFR fingerprint, a matrix containing channel power and arrival angle information is generated by transforming to the angle domain and performing the Hadamard inner product. Compared with CFR and CIR fingerprints, the ADCFP fingerprint matrix provides richer environmental features and can more accurately describe the propagation characteristics of the signal at a specific location, especially in a multipath propagation environment. This matrix can make full use of multipath information and improve the robustness and accuracy of positioning.

[0047] CAOA clustering algorithm, where CAOA is the abbreviation of Central Angle of Arrival. It is a clustering method based on the signal arrival angle characteristics. The fingerprints in the fingerprint database are grouped according to the size of the central arrival angle to form multiple cluster intervals. This grouping method utilizes the concentrated characteristics of the signal in the angle domain in a multipath environment, which can effectively reduce the fingerprint matching range during online positioning, thereby accelerating the positioning process and reducing the consumption of computing resources. The CAOA algorithm can quickly screen out the most likely matching fingerprint candidate set during positioning by pre-classifying the fingerprints, thereby improving the efficiency of the positioning algorithm.

[0048] The KNN algorithm is based on the traditional K-nearest neighbor algorithm and is the abbreviation of K-Nearest Neighbors. It is a non-parametric classification and regression method. Its basic idea is to select the K points with the closest distance as its neighbors by calculating the distance between the unknown sample point and all the points in the training sample, and then classify or regress the unknown sample point according to the label information of these K neighbors. In fingerprint positioning, the unknown sample point is the armored car to be located, and the training sample point is the fingerprint information of the known position in the fingerprint database. The KNN algorithm is simple and intuitive, but its biggest drawback is that it does not distinguish the distance of the neighboring points, that is, all K neighboring points contribute equally to the classification or regression of the unknown sample point. This may lead to large errors in the positioning results when a distant sample point is selected as a neighbor.

[0049] The WKNN algorithm, Weighted K-Nearest Neighbors, assigns a weight to each neighbor point, which is usually inversely proportional to the distance between the point and the unknown sample point. The closer the point is, the greater its weight is, and the greater its impact on the classification or regression of the unknown sample point is. This can effectively reduce the negative impact of distant points on the estimation results and improve positioning accuracy.

[0050] The improved WKNN algorithm improves the positioning accuracy of the cash transport vehicle positioning system based on fingerprint positioning by adjusting the weight calculation method. By increasing the weight coefficient of the distance to the power form, that is, the weight coefficient is inversely proportional to the nth power of the distance (n>1), the influence of distant fingerprint points is further reduced, and the weight of nearby fingerprint points is enhanced, thereby improving the accuracy and robustness of positioning. This algorithm design takes into account the signal characteristics in a multipath environment. By dynamically adjusting the weight of the fingerprint point, the location of the signal source can be estimated more accurately, especially in a complex urban environment, and can provide more stable and reliable positioning results.

[0051] It should be noted that the vehicle positioning method and device based on fingerprint positioning in the present disclosure can be used in the field of 5G technology. When the accurate positioning of the armored car is achieved based on fingerprint positioning, it can also be used in any field except the field of 5G technology. When the accurate positioning of the armored car is achieved based on fingerprint positioning, the application field of the vehicle positioning method and device based on fingerprint positioning in the present disclosure is not limited.

[0052] It should be noted that in the present disclosure, when collecting and analyzing customer information, corresponding operation entrances are provided for users to choose to agree or reject the automated decision results; if the user chooses to reject, the expert decision-making process is entered.

[0053] The following embodiments of the present invention can be applied to various systems / applications / devices for vehicle positioning based on fingerprint positioning. The present invention introduces an improved ADCFP fingerprint matrix and utilizes the concentrated characteristics of channel power in the angle domain and multipath propagation information to more accurately describe the position characteristics of the money transporter on the way. Compared with traditional CIR or CFR fingerprints, the ADCFP fingerprint matrix provides richer environmental information, thereby improving the recognition and positioning accuracy of the positioning system. In complex multipath and non-line-of-sight environments, this fingerprint based on angle and power can effectively utilize the multipath effect and further enhance the accuracy of positioning. By achieving real-time and accurate positioning of money transporters on the way, the present invention can effectively monitor the running trajectory of money transporters, reduce unexpected deviations or delays, and reduce the risk of property loss caused by inaccurate position information of money transporters.

[0054] At the same time, the present invention is based on the improved quadratic WKNN matching algorithm. On the basis of the traditional WKNN algorithm, the present invention increases the weight of fingerprints with higher similarity. This enhanced weight distribution mechanism ensures that fingerprints with closer distances have a greater impact on the positioning results, and can provide stable and reliable positioning results even in multipath environments or signal interference conditions, thereby improving the robustness of the positioning system.

[0055] The present invention is described in detail below in conjunction with various embodiments.

[0056] Embodiment 1

[0057] According to an embodiment of the present invention, an embodiment of a vehicle positioning method based on fingerprint positioning is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0058] The vehicle positioning method embodiment based on fingerprint positioning provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing a vehicle positioning method based on fingerprint positioning is shown. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more ( Figure 1 The computer system includes a processor 102 (the processor 102 may include but is not limited to a microcontroller unit (MCU) or a programmable logic device (FPGA) or a processing device such as a field programmable gate array (FPGA)), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, the computer system may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. A person skilled in the art can understand that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0059] It should be noted that the one or more processors 102 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10 (or mobile device). As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0060] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the vehicle positioning method based on fingerprint positioning in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the above-mentioned vehicle positioning method based on fingerprint positioning. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0061] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0062] The display may be, for example, a touch screen liquid crystal display (LCD), which enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0063] It should be noted that the embodiments of the present invention can be applied to the cash transporter positioning system of financial institutions. The cash transporter precise positioning system based on fingerprint positioning can provide real-time, high-precision vehicle location information. By installing a dedicated signal transmitter on each cash transporter in the fleet, the system can continuously collect and analyze the characteristics of the wireless signals sent by the cash transporter. These signal characteristics, including the channel impulse response (CIR), channel frequency response (CFR) and angle-delay channel frequency power (ADCFP) fingerprint, will be transmitted to the central monitoring center for real-time processing and matching.

[0064] Furthermore, the positioning method of this embodiment can combine the real-time position and route of the cash transport vehicle to provide dynamic route planning suggestions for the control center. Through the pre-established fingerprint database and CAOA clustering algorithm, the system can issue a timely warning when the cash transport vehicle deviates from the predetermined route or enters a high-risk area, prompting the escort personnel to take necessary safety measures, such as increasing the escort force or changing the route.

[0065] Under the above operating environment, this application provides Figure 2 The vehicle positioning method based on fingerprint positioning is shown. Figure 2 is a flow chart of an optional vehicle positioning method based on fingerprint positioning according to an embodiment of the present invention, such as Figure 2 As shown, the method comprises the following steps:

[0066] Step S201, using multiple receiving antennas of the base station to detect the characteristics of the uplink signal transmitted by the cash transport vehicle to be located, and generating a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized by the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP includes the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be located arriving at the receiving antenna of the base station, which is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation.

[0067] In step S201, the present invention first focuses on the extraction of signal features. In the 5G massive MIMO-OFDM environment, multiple receiving antennas equipped with the base station can capture rich and complex signal features, especially when these antennas are arranged according to the uniform linear array (ULA) model, the angle of arrival (AOA) information of the signal can be collected more effectively. In addition, by designing that the number of antennas exceeds the number of money transporters in the area, it can be ensured that the signal characteristics of each money transporter can be fully captured and distinguished even in a multi-target scenario.

[0068] Optionally, the step of using multiple receiving antennas of a base station to detect uplink signal characteristics emitted by the cash transport vehicle to be located includes: using multiple receiving antennas in the base station to detect signals emitted by the cash transport vehicle to be located, and obtaining uplink signal characteristics of the cash transport vehicle to be located, wherein the base stations are arranged according to a uniform linear array model, the number of receiving antennas in each base station is greater than the number of cash transport vehicles to be located distributed in the area to which the base station belongs, and the fingerprint information of the user to be located on the cash transport vehicle to be located is transmitted to the receiving antenna along multiple paths.

[0069] When the cash transport vehicle to be located transmits an uplink signal, the signal propagates through multiple paths to reach the base station, and the signal on each path carries different time of arrival (TOA) and angle of arrival (AOA) information. Multiple receiving antennas can detect the signal at the same time, thereby obtaining the uplink signal characteristics containing multipath information.

[0070] Figure 3 is a schematic diagram of an optional model for building a 5G massive MIMO-OFDM fingerprint positioning system according to an embodiment of the present invention, such as Figure 3 As shown, the system model is simplified and configured as follows:

[0071] 1. The base station is a uniform linear array (ULA) model, and the number of receiving antennas is N. t , the number of transmitting antennas is 1, and the distance between two adjacent antennas is d. It should be noted that in the ULA model for signal feature capture here, when multiple receiving antennas of the base station are used to detect the uplink signal characteristics emitted by the armored car to be located, the antennas between the base stations are arranged according to the ULA model. The ULA model means that the antenna array is evenly distributed along a straight line, and the distance between the antennas is a fixed value, usually half or an integer multiple of the signal wavelength. This arrangement enables the receiving antenna to accurately measure the AOA of the signal because the signal will experience phase differences when it reaches different antennas, and the antenna array under the ULA model can convert these phase differences into angle information.

[0072] 2. There are K cash transport vehicles to be located randomly distributed in the area, and N t >>K.

[0073] 3. There are many scatterers randomly and evenly distributed in this area, so that the fingerprint information of the user in the cash transport vehicle to be located can be transmitted to the receiving end along P paths. Each multipath transmission path is represented by d p,k , the incident angle is θ p,k .

[0074] 4. The receiving end obtains the channel state from the cash transport vehicle to the base station by performing channel estimation on the received signal. The two most common channel state information are the channel impulse response CIR and the channel frequency response CFR.

[0075] It should be noted that the CIR fingerprint matrix is ​​the result obtained after channel estimation of the uplink signal, recording the strength and time delay characteristics of the signal when it reaches the receiving antenna after multipath propagation. Through the channel estimation strategy, the base station in this embodiment can accurately measure the signal strength and arrival time on each path, and then generate a CIR fingerprint matrix. The channel frequency response CFR fingerprint matrix is ​​the result of Fourier transform of the CIR fingerprint. It converts the time domain signal characteristics to the frequency domain, thereby characterizing the frequency domain expression of the sum of the CIRs of each path under different delays. The generation of the CFR fingerprint matrix can not only reduce the volatility of the CIR fingerprint in the time domain and improve the stability of the characteristics, but also utilize the characteristics of the frequency domain, such as frequency selective fading, to further enrich the fingerprint information. The CFR fingerprint matrix can effectively capture the spectral characteristics of the channel, which is crucial for fingerprint matching and positioning in the frequency domain.

[0076] It should also be noted that the angle-delay channel frequency power (ADCFP) fingerprint matrix is ​​an innovation in the present invention, which is obtained by transforming the CFR fingerprint through the phase shift matrix and the inner product. The ADCFP fingerprint matrix not only contains the frequency domain information in the CFR, but also integrates the signal arrival angle information and the power distribution on each path. By converting the CFR fingerprint to the angle domain, the present invention makes full use of the advantages of MIMO technology in angle resolution and can more accurately identify the multipath effect in the channel.

[0077] In this embodiment, multipath information such as angle and delay is used to form ADCFP fingerprint, which includes the channel power and arrival angle AOA information of the user reaching the base station. By receiving the uplink signal sent by the cash transport vehicle in real time, CIR is extracted, and then ADCFP fingerprint is obtained through transformation, and the location is obtained by matching with the fingerprint in the fingerprint library.

[0078] Furthermore, in the embodiment of the present invention, the number of receiving antennas in each base station is greater than the number of money transporters in the area to which the base station belongs, ensuring that even in a multi-target environment, the system has sufficient signal discrimination and receiving capability to accurately capture the signal characteristics of each money transporter. In addition, through the design of the above-mentioned ULA model, the present invention can utilize the characteristics of the antenna array to effectively extract the multipath information of the signal and further improve the positioning accuracy. In the multi-target tracking (MOT) system, the money transporter is the target object, and its fingerprint information (including but not limited to CIR, CFR and ADCFP) can be transmitted to the receiving antenna of the base station along multiple paths, and the signal characteristics on each path will be captured and recorded by the receiving antenna for building a fingerprint database.

[0079] Step S202, matching the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprints in a pre-established fingerprint database to obtain a matching result.

[0080] In step S202, based on the extracted angle-delay channel frequency power (ADCFP) fingerprint matrix, it is matched with a pre-established fingerprint database. The fingerprint database stores a large number of historical signal features at known locations, which are collected and pre-processed in the offline stage to generate a specific fingerprint matrix. The matching process determines the possible location of the cash transport vehicle to be located by comparing the similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database.

[0081] Fingerprint database construction is an important guarantee for the accuracy of online positioning, and its construction process includes collection, preprocessing, fingerprint extraction and storage. Optionally, the fingerprint database is constructed in the following manner: collect historical signal features at known locations, wherein the historical signal features include: historical channel impulse response fingerprint matrix CIR, historical channel frequency response fingerprint matrix CFR, and historical angle-delay channel frequency power fingerprint matrix ADCFP improved based on CFR fingerprint; extract fingerprint information from preprocessed historical signal features to generate a fingerprint matrix, the fingerprint matrix contains the historical angle-delay channel frequency power fingerprint matrix ADCFP, and the corresponding relationship between the coordinates of the corresponding known position; store the fingerprint matrix in a preset database to obtain a fingerprint database, wherein the fingerprint database stores the corresponding relationship between each reference fingerprint and the corresponding position coordinates.

[0082] The historical signal features are collected at known locations, including the historical channel impulse response (CIR), the historical channel frequency response (CFR), and the historical ADCFP fingerprint matrix. AOA and TOA information are extracted from the historical CIR, and the historical CFR is obtained by Fourier transform. The historical CFR is further transformed into the angle domain using the phase shift matrix to form the historical ADCFR. The Hadamard inner product operation is performed on the historical ADCFR to generate the historical ADCFP fingerprint matrix. The Hadamard inner product can convert the complex matrix into a real matrix, simplifying the amount of calculation in the online matching process.

[0083] Among them, the historical CIR fingerprint matrix describes the time delay and strength characteristics of the signal in a multipath environment, and reflects the impulse response of the signal when it reaches the receiving antenna; the historical CFR fingerprint matrix is ​​obtained from the CIR fingerprint through Fourier transform. It reflects the response characteristics of the channel to signals of different frequencies, can eliminate the volatility of the signal in the time domain, and provide more stable spectrum characteristics. The historical ADCFP fingerprint matrix is ​​improved based on the CFR fingerprint. It contains the angle, delay and channel frequency power information of the signal arrival. It is a fingerprint matrix that integrates multi-dimensional information, which helps to improve the accuracy and robustness of positioning.

[0084] After obtaining the historical signal features, preprocessing is required. Here, preprocessing may include but is not limited to: data cleaning, signal enhancement, feature extraction and other steps to eliminate the influence of noise, interference and other non-signal factors. Next, fingerprint information is extracted from the preprocessed historical signal features to generate a fingerprint matrix containing ADCFP fingerprints. This process converts the historical signal features into a specific fingerprint format for subsequent storage and matching. The fingerprint matrix establishes a direct correspondence with the coordinates of the known location. This correspondence is based on the intrinsic connection between signal features and geographic locations and is the core of the positioning system. The extracted fingerprint matrix is ​​stored in a preset database to obtain a fingerprint database. The fingerprint positioning reference fingerprint in the fingerprint database, each reference fingerprint is associated with a specific location coordinate, and there is a one-to-one correspondence between the stored fingerprint matrix and the corresponding location coordinate. For large-scale MIMO-OFDM systems, the rich multipath information contained in the fingerprint matrix can be used to describe the propagation of signals in complex environments. By associating and storing each fingerprint matrix with its actual location coordinates when it was collected, the constructed fingerprint database becomes a reference for comparison and positioning in the online positioning stage.

[0085] Optionally, the step of collecting historical signal features at a known position includes: performing channel estimation on historical uplink signals received in the historical process to obtain a historical channel impulse response fingerprint matrix CIR, wherein the matrix size of the historical channel impulse response fingerprint matrix CIR is determined by the number of receiving antennas of the base station and the length of the channel impulse response; extracting arrival angle AOA and arrival time TOA information from the historical channel impulse response fingerprint matrix CIR; converting the historical channel impulse response fingerprint matrix CIR into a historical channel frequency response fingerprint matrix CFR through Fourier transform to form frequency domain signal features, wherein the matrix size of the historical channel frequency response fingerprint matrix CFR is determined by the number of receiving antennas and the number of transmitting antennas of the base station; using a phase shift matrix to transform the historical channel frequency response fingerprint matrix CFR from the frequency domain to the angle domain to obtain a historical angle-delay channel frequency response matrix ADCFR, wherein the phase shift matrix is ​​obtained based on a uniform linear array model corresponding to the antenna array in the base station; performing a preset inner product operation on the historical angle-delay channel frequency response matrix ADCFR to generate a historical angle-delay channel frequency power fingerprint matrix ADCFP.

[0086] In the process of collecting historical signal features, the system first performs channel estimation on the received historical uplink signals to obtain the historical channel impulse response (CIR) fingerprint matrix. CIR is the time domain representation of the signal after multipath propagation. It contains information about the arrival time (TOA) of the signal and the change of signal strength over time. The size of the historical CIR matrix depends on the number of receiving antennas of the base station and the length of the channel impulse response, which reflects the richness of multipath information that the system can extract from the signals received by multiple antennas. For example, the matrix size is L×M, where L represents the length of the channel impulse response and M represents the number of receiving antennas of the base station. From the acquired historical CIR fingerprint matrix, the signal's angle of arrival (AOA) and time of arrival (TOA) information can be further extracted. This information is crucial for describing the signal propagation path, especially for MIMO (Multiple Input Multiple Output) systems. AOA and TOA can reflect the propagation characteristics of the signal in space and provide key parameter support for subsequent signal feature conversion and fingerprint library construction. Then, the historical CIR fingerprint matrix is ​​converted into the historical channel frequency response (CFR) fingerprint matrix through Fourier transform (FFT), realizing the conversion of signal characteristics from time domain to frequency domain. This conversion helps to capture the frequency selective fading characteristics of the channel, which is of great value for signal analysis and processing in the frequency domain. The size of the historical CFR matrix is ​​determined by the number of receiving antennas M and the number of transmitting antennas N of the base station, which is an M×N complex matrix, which reflects the information of multi-antenna signals in the frequency domain that the system can process simultaneously.

[0087] After that, the angle domain conversion from CFR to ADCFR is realized. The historical CFR fingerprint matrix is ​​transformed from the frequency domain to the angle domain using the phase shift matrix to obtain the historical angle-delay channel frequency response (ADCFR) matrix. The design of the phase shift matrix is ​​based on the uniform linear array model (ULA) of the antenna array in the base station. This model takes into account the impact of the antenna spacing in the antenna array on signal reception and can combine the signal frequency information with the arrival angle information. The generation of the ADCFR matrix provides a basis for subsequent angle domain signal analysis and processing, and can effectively separate and identify signal features at different arrival angles.

[0088] Finally, by performing a preset inner product operation (such as Hadamard inner product) on the historical angle-delay channel frequency response matrix ADCFR, the historical angle-delay channel frequency power (ADCFP) fingerprint matrix is ​​generated, and the complex information in the ADCFR matrix is ​​converted into real information. The generated ADCFP matrix contains channel power and arrival angle information, while also retaining the signal delay characteristics. Through channel estimation, arrival angle and time information extraction, Fourier transform from CIR to CFR, angle domain conversion from CFR to ADCFR, and inner product operation from ADCFR to ADCFP, a comprehensive fingerprint matrix containing angle, delay and power information is constructed. This series of transformations and processing not only fully explores the spatial and frequency characteristics of the signal, but also provides multi-dimensional information support for the construction of the fingerprint database, ultimately improving the positioning efficiency and accuracy of the cash transporter precision positioning system based on fingerprint positioning.

[0089] Optionally, the step of performing a preset inner product operation on the historical angle-delay channel frequency response ADCFR to generate a historical angle-delay channel frequency power fingerprint matrix ADCFP includes: performing multiplication operations on each row and each column element in the historical angle-delay channel frequency response matrix ADCFR with the corresponding elements of the phase shift matrix, and summing the product values ​​to obtain each row and each column element of the historical angle-delay channel frequency power fingerprint matrix ADCFP; performing a Hadamard inner product operation on the historical angle-delay channel frequency power fingerprint matrix ADCFP to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP, wherein the historical angle-delay channel frequency response matrix ADCFR is a complex matrix, and the historical angle-delay channel frequency power fingerprint matrix ADCFP is a real matrix ADCFP.

[0090] Among them, the key is to use the phase shift matrix to perform the inner product operation on the historical angle-delay channel frequency response (ADCFR) to generate the historical angle-delay channel frequency power fingerprint matrix (ADCFP) in order to extract the power characteristics of the signal in the angle domain. First, each row and column element of the historical ADCFR matrix (complex matrix) is multiplied with the corresponding element of the phase shift matrix. The phase shift matrix is ​​designed based on the ULA model (Uniform linear array), and its elements reflect the phase difference between the signals received by different antennas. By multiplying with the elements of the ADCFR matrix, the signal can be converted from the frequency domain to the angle domain while retaining the power information of the signal. Then, the product values ​​of each row and column are summed to obtain each row and column element of the ADCFP matrix. This inner product operation combines the frequency and angle characteristics of the signal, which helps to use multipath information for positioning in a multipath environment, especially when the channel power is relatively concentrated in the angle domain. Then, a Hadamard inner product operation is performed on the historical ADCFR to generate the historical ADCFP. The Hadamard inner product, also known as element-by-element multiplication, can multiply the corresponding elements of two matrices of the same size and then sum them to generate a new matrix. In the present invention, this operation converts the historical ADCFR matrix into a real number matrix ADCFP, which simplifies the calculation of the online matching stage while retaining the key features of the angle and power information. The generation of the ADCFP matrix not only improves the positioning accuracy, but also reduces the complexity of data processing and improves the overall efficiency of the system.

[0091] After collecting the historical signal features at the known position, in order to reduce the amount of calculation in the online matching stage and save the storage space of the fingerprint database, the present invention further compresses the ADCFP fingerprint matrix to generate a more concise fingerprint representation. Optionally, after collecting the historical signal features at the known position, it also includes: compressing the ADCFP fingerprint matrix by row to obtain the horizontal channel frequency power matrix HCFP, and compressing the ADCFP fingerprint matrix by column to obtain the vertical channel frequency power matrix VCFP; combining the horizontal channel frequency power matrix HCFP and the vertical channel frequency power matrix VCFP to generate the horizontal vertical channel frequency fingerprint matrix HVCFF.

[0092] During the compression process, the ADCFP fingerprint matrix is ​​first compressed row by row, that is, the elements of each row are summed to generate the horizontal channel frequency power matrix HCFP. This method makes full use of the distribution characteristics of the signal in the frequency domain. By integrating the power information at the same frequency, the dimension of the matrix is ​​reduced, thereby reducing the complexity of storage and calculation. Next, the ADCFP fingerprint matrix is ​​compressed column by column, that is, the elements of each column are summed to generate the vertical channel frequency power matrix VCFP. This step is opposite to HCFP. It focuses on the characteristics of the signal in the angle domain, integrates the signal power information at different angles, and further compresses the size of the fingerprint database. Finally, by combining HCFP and VCFP, a simplified fingerprint matrix HVCFF is generated. The matrix is ​​represented by a one-dimensional vector and contains the channel frequency power information in both horizontal and vertical dimensions. This combination not only makes full use of the information in two dimensions, but also greatly reduces the data dimension, realizes the effective compression of the fingerprint matrix, improves the positioning speed, and maintains the positioning accuracy.

[0093] Through compression, not only the data dimension is reduced, but also key channel features such as channel power distribution are retained. Finally, the HCFP and VCFP matrices are combined to generate the horizontal vertical channel frequency fingerprint (HVCFF) matrix, which integrates the information of horizontal and vertical dimensions and provides a more comprehensive positioning reference.

[0094] Optionally, after the fingerprint matrix is ​​stored in a preset database to obtain the fingerprint database, it also includes: using a central angle of arrival clustering algorithm to divide the fingerprint database into intervals to obtain multiple cluster intervals; calculating the central angle of arrival CAOA of each reference fingerprint in the fingerprint database; and dividing the reference fingerprint into corresponding cluster intervals according to the size of the central angle of arrival CAOA of each reference fingerprint.

[0095] After the fingerprint database is built, in order to further improve the speed and efficiency of online matching, the present invention introduces a center angle of arrival (CAOA) clustering algorithm, and divides the data in the fingerprint database into multiple cluster intervals according to the distribution characteristics of the channel power in the angle domain. First, a series of CAOA interval ranges are defined according to the concentration of the channel power. These intervals classify the fingerprints in the fingerprint database according to the central value of the angle of arrival, reducing the number of fingerprints that need to be compared during online matching. Then, the center angle of arrival CAOA of each fingerprint in the fingerprint database is calculated. The calculation of CAOA depends on the ADCFP fingerprint matrix. By analyzing the distribution of signal power at different angles, the center value of the angle of arrival of each fingerprint can be determined. Finally, according to the CAOA value of each reference fingerprint, it is assigned to the corresponding cluster interval. In the online positioning stage, the system only needs to search for fingerprints in the cluster interval that matches the CAOA of the point to be positioned, thereby reducing the search range and improving the real-time and accuracy of positioning.

[0096] It should be noted that this embodiment illustrates the center angle of arrival (CAOA) clustering algorithm, which divides the fingerprints in the database into intervals according to the size of CAOA based on the concentration of channel power in the angle domain in the historical signal characteristics, forming multiple cluster intervals. In the online matching stage, the ADCFP fingerprint of the point to be located is first determined to which cluster interval its CAOA belongs, and then only matched with the fingerprints in the interval, which effectively reduces the scope of fingerprint matching, speeds up the positioning speed, and maintains a high positioning accuracy.

[0097] Optionally, the extracted angle-delay channel frequency power fingerprint matrix ADCFP is matched with the fingerprints in a pre-established fingerprint database to obtain a matching result, comprising: calculating the similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database, and obtaining a matching result based on the similarity, wherein Manhattan distance or Euclidean distance is used as the similarity in the process of calculating the similarity.

[0098] In the accurate positioning system of the cash transport vehicle based on fingerprint positioning, the extracted ADCFP fingerprint matrix is ​​matched with the fingerprints in the pre-established fingerprint database. Specifically, in the embodiment of the present invention, Manhattan distance or Euclidean distance is used as the calculation standard of similarity to determine the most likely position of the cash transport vehicle to be located. In the online positioning stage, the system needs to calculate the similarity between the ADCFP fingerprint matrix of the cash transport vehicle to be located and each reference fingerprint in the fingerprint database. The similarity measures the degree of difference between two fingerprints. The smaller the difference, the more similar the fingerprints are, that is, the vehicle to be located may be at the position corresponding to the reference fingerprint.

[0099] It should be noted that in the similarity calculation process, Manhattan distance or Euclidean distance can be used as a measure of similarity. Manhattan distance refers to the total distance between two points along the coordinate axis in a multidimensional data space, while Euclidean distance refers to the straight-line distance between two points in a multidimensional space. In the present invention, Manhattan distance is preferred because of its simple calculation and low sensitivity to large values, especially when the dimension of fingerprint data is too high, it can significantly improve the matching speed while maintaining a high positioning accuracy. However, Euclidean distance may provide a more intuitive distance measurement in some scenarios, and is suitable for situations where fingerprint features are densely distributed, because it can more accurately reflect the distance relationship between fingerprints.

[0100] Step S203: According to the matching result, the improved square WKNN matching algorithm is used to locate the position coordinates of the cash transport vehicle to be located.

[0101] Based on KNN, the WKNN algorithm assigns a weight to each reference fingerprint. The weight is inversely proportional to the distance between the reference fingerprint and the fingerprint to be located. That is, the closer the reference fingerprint is, the greater the impact on the position estimation, which can improve the positioning accuracy, especially when there are multiple location fingerprints in the fingerprint database that have a high similarity with the fingerprint to be located.

[0102] In order to further optimize the fingerprint matching process, the embodiment of the present invention also proposes an improved power WKNN matching algorithm. Based on the WKNN algorithm, the algorithm increases the weight of fingerprints with higher similarity and slightly reduces the weight of fingerprints with lower similarity by adjusting the power form of the distance in the weighting coefficient. This adjustment helps to improve positioning accuracy in complex environments. By more carefully considering fingerprint information at a closer distance, positioning errors can be reduced. In particular, when there are multiple similar fingerprints, the current position of the money transporter can be more accurately located.

[0103] In the improved power WKNN matching algorithm, the weighted coefficient of the fingerprint is proportional to the nth power of the fingerprint similarity, and the position coordinates of the cash transport vehicle to be located are obtained by calculating the weighted average value.

[0104] Optionally, step S203 includes: sorting the similarities in the matching results, selecting K fingerprints with the highest similarity as candidate matching fingerprints, where the K value is an integer greater than 1; using an improved power WKNN matching algorithm, assigning different weighting coefficients to each candidate matching fingerprint according to the distance between the candidate matching fingerprint and the fingerprint of the point to be located; and calculating the weighted average of the position coordinates corresponding to the candidate matching fingerprints according to the weighting coefficients as the position coordinates of the armored car to be located.

[0105] By calculating the similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database, the system can obtain a set of reference fingerprints that are closest to the fingerprint to be located. Based on the position information of this set of reference fingerprints, the estimated position of the armored car to be located can be obtained through weighted averaging or other position estimation methods. This process not only depends on the accuracy of the fingerprint data, but is also closely related to the selection of the matching algorithm. The correct algorithm can significantly improve the efficiency and accuracy of positioning.

[0106] The acquisition of matching results provides the remote monitoring center with the ability to understand the exact location of the cash transporter in real time. Combined with the driving route and surrounding environment information of the cash transporter, the monitoring center can monitor the driving status of the vehicle in real time, promptly detect safety risks such as deviation from the route and abnormal stop, and quickly activate the early warning response mechanism to provide all-round protection for the safety of the cash transporter on the way.

[0107] Through the above steps, multiple receiving antennas of the base station can be used to detect the uplink signal characteristics emitted by the cash transport vehicle to be located, and a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP are generated based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized as the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP includes the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be located to the receiving antenna of the base station, which is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation, and the extracted angle-delay channel frequency power fingerprint matrix ADCFP is matched with the fingerprint in the pre-established fingerprint database to obtain a matching result, and according to the matching result, the position coordinates of the cash transport vehicle to be located are located by using the improved power WKNN matching algorithm. In this embodiment, the ADCFP fingerprint matrix can be used to utilize the concentrated characteristics of channel power in the angle domain and multipath propagation information to more accurately describe the position characteristics of the money transporter on the way, thereby realizing fingerprint matching and obtaining real-time and accurate positioning of the money transporter. At the same time, the remote monitoring center can also understand the exact position and driving conditions of the money transporter in real time. In complex multipath and non-line-of-sight environments, the angle- and power-based fingerprint can effectively utilize the multipath effect, further enhance the accuracy of positioning, reduce unexpected yaw or delay, and reduce the risk of property loss caused by inaccurate position information of the money transporter, thereby solving the technical problem in related technologies that the money transporter is affected by different environmental factors, has low positioning accuracy, and reduces vehicle safety.

[0108] Another optional specific implementation is described in detail below.

[0109] Figure 4 is a schematic diagram of a cash transport vehicle positioning method based on fingerprint positioning according to an embodiment of the present invention, such as Figure 4 As shown, it includes an offline stage and an online stage, wherein the offline training stage builds a fingerprint database, and the online positioning stage completes the matching of the received information and the fingerprint database.

[0110] In the offline stage, each monitoring station first captures the signal characteristics at a known location, extracts the fingerprint into a specific fingerprint matrix, and then compresses the fingerprint matrix and performs other related preprocessing to form a fingerprint database containing many fingerprints. The position coordinates of the signal with the base station as the reference point should have a one-to-one correspondence with the fingerprint that stores the signal characteristics. In order to alleviate the pressure of fingerprint database storage and search and improve positioning speed, clustering algorithms can also be used to compress the fingerprint database.

[0111] When in the online stage, the signal characteristics of the terminal to be located are first collected, and the extracted fingerprint and fingerprint database are mapped using the fingerprint similarity criterion to build a mapping relationship. Finally, the matched fingerprint is used to estimate the position through a matching algorithm.

[0112] Specifically, the cash transport vehicle positioning method based on fingerprint positioning includes the following steps:

[0113] Step 1: Build a 5G massive MIMO-OFDM fingerprint positioning system model.

[0114] like Figure 3 As shown, the embodiment of the present invention first constructs a 5G massive MIMO-OFDM fingerprint positioning system model.

[0115] As one of the key technologies of 5G, massive MIMO can collect more abundant multipath fingerprint information due to the deployment of a large number of receiving antennas at the base station, thereby improving the accuracy of fingerprint positioning.

[0116] In order to facilitate the subsequent description of the system model, the MIMO-OFDM fingerprint positioning system model is configured as follows in this embodiment:

[0117] 1. The base station is a uniform linear array (ULA) model, and the number of receiving antennas is N. t , the number of transmitting antennas is 1, and the distance between two adjacent antennas is d.

[0118] 2. There are K users to be located randomly distributed in the area, and N t >>K.

[0119] 3. There are many scatterers randomly and evenly distributed in the area, so that the fingerprint information of the user to be located can be transmitted to the receiving end along P paths. Each multipath transmission path is represented by dp,k , the incident angle is θ p,k .

[0120] 4. The receiving end obtains the channel state from the user to be located to the base station by performing channel estimation on the received signal. The two most common channel state information are the channel impulse response CIR and the channel frequency response CFR.

[0121] Step 2: Extraction of fingerprint matrix.

[0122] The fingerprint database of reference points in the offline phase and the fingerprint matrix of the points to be located in the online phase are constructed to generate CIR fingerprints and their improved ADCFP fingerprints.

[0123] step1: CIR fingerprint.

[0124] The CIR fingerprint can be obtained by estimating the channel of the received uplink signal, and the CIR fingerprint can be extracted by ray tracing. The channel impulse response qp of the kth user on path p is k , the specific formula is as follows:

[0125]

[0126] Among them, a p,k is the complex channel gain, d p,k is the distance from the transmitter to the first receiving antenna, λ c is the wavelength of the carrier, It is the signal AOA, is the array response vector.

[0127] Since the TOA of different paths is different, the CIR of the kth user can be expressed as q on each path p,k The specific formula is as follows:

[0128]

[0129] Among them, τ p,k is the arrival time. The CIR fingerprint matrix is ​​represented by Q k , is of size N t ×N q A matrix where N q is the length of the channel impulse response.

[0130] step2: CFR fingerprint.

[0131] The CFR fingerprint in this embodiment is obtained by Fourier transforming the CIR fingerprint. That is, for the l-th subcarrier, the CFR is the sum of the CIRs of the paths extended at different times. The specific formula is as follows:

[0132]

[0133] in, represents the propagation delay, Refers to the integer closest to x. According to the sampling interval T s The channel output is sampled, and the OFDM symbol interval is represented by T c =N c T s , where N c is the number of OFDM symbols.

[0134] CFR Matrix H k The h of each subcarrier k,l Composition, N t ×N l The complex matrix of is as follows:

[0135]

[0136] step3: ADCFP fingerprint.

[0137] The CIR fingerprint is improved and a new fingerprint matrix is ​​proposed: Angle Delay Channel Frequency Power (ADCFP). This fingerprint matrix contains the channel power and arrival angle AOA information of the user reaching the base station.

[0138] Figure 5 FIG. 1 is a schematic diagram of an optional transformation from a CFR fingerprint to an ADCFP fingerprint matrix according to an embodiment of the present invention. First, the CFR fingerprint matrix H k Transform it from the frequency domain to the angle domain to obtain the angle-delay channel frequency response B k (ADCFR, Angle Delay Channel Frequency Response), is N t ×N l The complex matrix of is as follows:

[0139]

[0140] in, is the phase shift matrix, and the specific formula is as follows:

[0141]

[0142] Because B k It is a complex number, which is more complicated when calculating the distance. Therefore, the ADCFR fingerprint is taken as the Hadamard inner product to obtain the ADCFP fingerprint C k , the specific formula is as follows:

[0143]

[0144] Through the Hadamard inner product transformation, the number of elements of the ADCFP fingerprint remains unchanged compared to the ADCFR fingerprint, but the ADCFR fingerprint is a complex matrix, while the ADCFP is a matrix of size N. t ×N l Therefore, compared with ADCFR fingerprint, ADCFP fingerprint can have a faster processing speed in the online matching stage, reducing the storage pressure of the fingerprint library, and is also conducive to fingerprint compression, clustering and other processing.

[0145] The ADCFP fingerprint is constructed using multipath information such as angle and delay, which includes the channel power and arrival angle AOA information of the user reaching the base station. By receiving the uplink signal sent by the cash transporter in real time, the CIR is extracted, and then the ADCFP fingerprint is obtained through transformation, and the location is obtained by matching it with the fingerprint in the fingerprint library.

[0146] By receiving the uplink signal sent by the cash transporter in real time, the CIR is extracted, and then the CFR and ADCFP fingerprints are obtained through transformation. In the offline stage, each monitoring station has captured and collected the signal characteristics at a known location, extracted the fingerprint into a specific fingerprint matrix, and constructed a fingerprint library. In the online positioning stage, the location can be obtained by simply matching the extracted fingerprint information of the cash transporter with the fingerprint in the fingerprint library.

[0147] Step 3: Compress the generated fingerprint matrix.

[0148] In order to reduce the storage and search pressure of the fingerprint library in the offline training stage and speed up the matching speed in the online stage, this embodiment proposes a storage method of row and column compression and applies it to the compressed fingerprint extraction of ADCFP. t ×N l The fingerprint matrix is ​​converted into a simple one-dimensional vector.

[0149] Figure 6 is a schematic diagram of an optional compression process for a fingerprint matrix according to an embodiment of the present invention. Figure 6 Describe the compression process.

[0150] step1: HCFP fingerprint.

[0151] First, the ADCFP fingerprint is compressed row by row, and the compressed fingerprint is named as the horizontal channel frequency power (HCFP). That is, the elements of each row of the ADCFP fingerprint matrix are summed up, and the obtained value is used as the element of each row of HCFP. The specific formula is as follows:

[0152]

[0153] Where C is the ADCFP fingerprint, C H is the HCFP fingerprint, which is a size of N t ×1 vector, its fingerprint compression rate is:

[0154]

[0155] step2: VCFP fingerprint.

[0156] Then the ADCFP fingerprint is compressed by column, and the compressed fingerprint is named as the vertical channel frequency power (VCFP). That is, the elements of each column of the ADCFP fingerprint matrix are summed up, and the obtained value is used as the element of each column of VCFP. The specific formula is as follows:

[0157]

[0158] Where C is the ADCFP fingerprint, C V is a VCFP fingerprint with a size of 1×N l The fingerprint compression rate of a vector is:

[0159]

[0160] step3: HVCFF fingerprint.

[0161] Finally, the VCFP fingerprint is spliced ​​after the HCFP fingerprint, making full use of the information of the two dimensions. The resulting compressed fingerprint is named the horizontal and vertical channel frequency fingerprint (HVCFF, Horizontal and Vertical Channel Frequency Fingerprint). The specific formula is as follows:

[0162]

[0163] Where C is the ADCFP fingerprint, C HV is the HVCFF fingerprint, which is a 1×(N t +N l ), its fingerprint compression rate is:

[0164]

[0165] From the compression ratios of the three compressed fingerprints, we can see that due to N t 、N lThe values ​​of are all large, so this method can achieve a good compression effect, greatly save the storage space of the fingerprint library, and effectively improve the positioning speed in the online stage.

[0166] Step 4: Use clustering algorithm on the fingerprint library in the offline stage.

[0167] According to the characteristic that the channel power of ADCFP fingerprint is relatively concentrated in the angle domain, this implementation method introduces the central angle of arrival (CAOA) clustering algorithm to make full use of the characteristics of its angle spectrum. The basic principle of the algorithm is to first divide the range of several CAOA intervals, then calculate the CAOA of each fingerprint in the fingerprint library, and divide it into specified intervals according to the size of its CAOA. In the online positioning stage, determine the cluster to which the CAOA of the point to be located belongs, and compare it with the fingerprint in this class to complete the fingerprint positioning. This method greatly reduces the number of fingerprint comparisons and effectively improves the positioning speed.

[0168] When the angle interval is divided in the CAOA algorithm, the value range of CAOA depends on the size of the ADCFP fingerprint matrix, which can be divided into Nt clusters at most, where Nt is the number of antennas. First, the solution method of CAOA needs to be defined. Taking ADCFP as an example, the specific formula is as follows:

[0169]

[0170] Among them, F i is the ADCFP fingerprint, α i is the CAOA corresponding to the fingerprint, α i The value range is 1 to N t , and can only be an integer. Since there will be angular diffusion in a scattering environment, each cluster should be an angular range, as follows:

[0171] [α i -△α,α i +△α]

[0172] Among them, △α is the maximum angular offset.

[0173] Step 5: Calculate fingerprint similarity using similarity criteria.

[0174] Compare the fingerprint data of the point to be located with all the reference fingerprints of the cluster center to which it belongs, select an appropriate similarity criterion to calculate the similarity of all fingerprints, and then select K points with the highest fingerprint similarity. The distance between two fingerprints is usually used to indicate the similarity between fingerprints. The closer the distance, the higher the fingerprint similarity.

[0175] Minkowski distance is a general definition of a type of distance relationship, including Euclidean distance and Manhattan distance. The specific formula is as follows:

[0176]

[0177] When m=2, the Minkowski distance is converted into the Euclidean distance, which represents the absolute distance between two points. The specific formula is as follows:

[0178]

[0179] When m=1, Minkowski distance is converted into Manhattan distance. The specific formula is as follows:

[0180]

[0181] For the above two distance criteria, it can be seen from the complexity of the formula that the Manhattan distance has a simpler structure than the Euclidean distance formula, so the Manhattan distance has a faster matching speed in the fingerprint positioning stage.

[0182] Step 6: Use matching algorithm to estimate the coordinates of the points to be located.

[0183] step1: KNN algorithm.

[0184] The KNN algorithm is an improvement of the nearest neighbor algorithm. It uses the fingerprint similarity criterion to calculate the distance relationship between the fingerprint of the point to be located and the fingerprint in the fingerprint library. Then, K fingerprints with the highest similarity are taken out and the average value of these fingerprint coordinates is calculated. The result is the coordinate position of the target user. The specific formula is as follows:

[0185]

[0186] in, is the coordinate position of the target user, is the coordinate position of the kth fingerprint. However, this algorithm simply averages the positions of the K fingerprints with the highest similarity, without distinguishing the specific distance. If there are points far away in these K fingerprints, the estimated position of the target signal will have a large deviation.

[0187] step2: WKNN algorithm.

[0188] Based on the KNN algorithm, the WKNN algorithm assigns a different weight to each of the selected K fingerprints. The more similar the fingerprint is to the target signal fingerprint, the greater the value of the weight coefficient is, and the greater the influence on the estimated position of the target signal is. The specific formula is as follows:

[0189]

[0190] Among them, w k is the weight coefficient of the kth fingerprint, and d k is the distance between the kth fingerprint and the target fingerprint. △d is a small positive number. The purpose is to avoid the denominator being 0. Its significance can be ignored.

[0191] step3: Improved WKNN algorithm.

[0192] On this basis, this embodiment proposes a new matching algorithm based on the WKNN algorithm, by adjusting the distance d in the weighted coefficient k In the form of a square, the square WKNN algorithm is introduced. The specific formula is as follows:

[0193]

[0194] in, is the distance d k The nth power, n value is greater than 1 (n value greater than 1 ensures that the algorithm can increase the weight of fingerprints with closer distance, which meets the performance requirements of the algorithm.). Based on the WKNN algorithm, this algorithm reduces the weight of fingerprints with lower similarity and increases the weight of fingerprints with higher similarity. In theory, the performance of this algorithm should be better than the WKNN algorithm. In addition, the larger the value of n, the better. When the value of n is too large, the weight of fingerprints with closer distance is too large, which can almost directly affect the positioning result. The effect is equivalent to the K value of 1 in the K nearest neighbor algorithm, which will also cause a large positioning error.

[0195] Through the above implementation, in the large-scale MIMO-OFDM system model under the 5G single-station environment, an improved ADCFP fingerprint matrix based on CIR is proposed, which can realize the real-time and accurate positioning of the cash transport vehicle in transit, and at the same time enable the remote monitoring center to understand the exact location and driving conditions of the cash transport vehicle in real time, realize real-time monitoring, flexibly change the driving route, provide timely rescue guarantee, etc. It improves the safety of the cash transport vehicle during driving, reduces property losses, and improves the ability of cash transport vehicle in transit supervision to a certain extent.

[0196] At the same time, for the offline training stage, the embodiment of the present invention proposes a storage method that compresses rows and columns, making full use of the information in two dimensions, converting the fingerprint matrix into a one-dimensional vector, and obtaining three compressed fingerprints of HCFP, VCFP and HVCFF. At the same time, the CAOA clustering algorithm is introduced to make full use of the characteristics that the channel power is more concentrated in the angle domain. For the method of improving the positioning accuracy in the online matching stage, the embodiment of the present invention proposes an improved power WKNN matching algorithm, which reduces the weight of fingerprints with lower similarity based on the WKNN algorithm.

[0197] The following is a detailed description in conjunction with another embodiment.

[0198] Embodiment 2

[0199] A vehicle positioning device based on fingerprint positioning provided in this embodiment includes multiple implementation units, each implementation unit corresponds to each implementation step in the above-mentioned embodiment 1. Its specific implementation method and beneficial effects can refer to the above-mentioned method embodiment, which will not be repeated here.

[0200] Figure 7 is a schematic diagram of an optional vehicle positioning device based on fingerprint positioning according to an embodiment of the present invention, such as Figure 7 As shown, the vehicle positioning device based on fingerprint positioning may include: a fingerprint matrix generating unit 71, a fingerprint matching unit 72, and a vehicle positioning unit 73.

[0201] Among them, the fingerprint matrix generation unit 71 is used to use multiple receiving antennas of the base station to detect the uplink signal characteristics emitted by the cash transport vehicle to be located, and generate a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized by the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP includes the channel power information and arrival angle AOA information of the signal sent by the cash transport vehicle to be located arriving at the receiving antenna of the base station, which is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation.

[0202] The fingerprint matching unit 72 is used to match the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprints in the pre-established fingerprint database to obtain a matching result.

[0203] The vehicle positioning unit 73 is used to locate the position coordinates of the cash transport vehicle to be located based on the matching results by using the improved square WKNN matching algorithm.

[0204] The above-mentioned vehicle positioning device based on fingerprint positioning can use multiple receiving antennas of the base station to detect the uplink signal characteristics emitted by the cash transport vehicle to be positioned through the fingerprint matrix generation unit 71, and generate the channel impulse response fingerprint matrix CIR, the channel frequency response fingerprint matrix CFR and the angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics. The extracted angle-delay channel frequency power fingerprint matrix ADCFP is matched with the fingerprint in the pre-established fingerprint database through the fingerprint matching unit 72 to obtain the matching result. The vehicle positioning unit 73 uses the improved power WKNN matching algorithm to locate the position coordinates of the cash transport vehicle to be positioned according to the matching result. In this embodiment, the ADCFP fingerprint matrix can be used to utilize the concentrated characteristics of channel power in the angle domain and multipath propagation information to more accurately describe the position characteristics of the money transporter on the way, thereby realizing fingerprint matching and obtaining real-time and accurate positioning of the money transporter. At the same time, the remote monitoring center can also understand the exact position and driving conditions of the money transporter in real time. In complex multipath and non-line-of-sight environments, the angle- and power-based fingerprint can effectively utilize the multipath effect, further enhance the accuracy of positioning, reduce unexpected yaw or delay, and reduce the risk of property loss caused by inaccurate position information of the money transporter, thereby solving the technical problem in related technologies that the money transporter is affected by different environmental factors, has low positioning accuracy, and reduces vehicle safety.

[0205] Optionally, the fingerprint matrix generation unit includes: a vehicle signal detection module, which is used to use multiple receiving antennas in the base station to detect the signal emitted by the money transport vehicle to be located, and obtain the uplink signal characteristics of the money transport vehicle to be located, wherein the base stations are arranged according to a uniform linear array model, and the number of receiving antennas in each base station is greater than the number of money transport vehicles to be located distributed in the area to which the base station belongs, and the fingerprint information of the user to be located on the money transport vehicle to be located is transmitted to the receiving antenna along multiple paths.

[0206] Optionally, when constructing a fingerprint database, the vehicle positioning device based on fingerprint positioning further includes: a signal feature acquisition unit, used to collect historical signal features at known positions, wherein the historical signal features include: a historical channel impulse response fingerprint matrix CIR, a historical channel frequency response fingerprint matrix CFR, and a historical angle-delay channel frequency power fingerprint matrix ADCFP improved based on the CFR fingerprint; a fingerprint information extraction unit, used to extract fingerprint information from the pre-processed historical signal features to generate a fingerprint matrix, wherein the fingerprint matrix includes the corresponding relationship between the historical angle-delay channel frequency power fingerprint matrix ADCFP and the coordinates of the corresponding known position; a fingerprint matrix storage unit, used to store the fingerprint matrix in a preset database to obtain a fingerprint database, wherein the fingerprint database stores the corresponding relationship between each reference fingerprint and the corresponding position coordinates.

[0207] Optionally, the signal feature acquisition unit includes: a channel estimation module, which is used to perform channel estimation on the historical uplink signals received in the historical process to obtain a historical channel impulse response fingerprint matrix CIR, wherein the matrix size of the historical channel impulse response fingerprint matrix CIR is determined by the number of receiving antennas of the base station and the length of the channel impulse response; a matrix information extraction module, which is used to extract the arrival angle AOA and arrival time TOA information from the historical channel impulse response fingerprint matrix CIR; a matrix conversion module, which is used to convert the historical channel impulse response fingerprint matrix CIR into a historical channel frequency response fingerprint matrix CFR through Fourier transform to form a frequency domain signal feature, wherein the matrix size of the historical channel frequency response fingerprint matrix CFR is determined by the number of receiving antennas and the number of transmitting antennas of the base station; a matrix frequency conversion module, which is used to transform the historical channel frequency response fingerprint matrix CFR from the frequency domain to the angle domain using a phase shift matrix to obtain a historical angle-delay channel frequency response matrix ADCFR, wherein the phase shift matrix is ​​obtained based on a uniform linear array model corresponding to the antenna array in the base station; an inner product module, which is used to perform a preset inner product operation on the historical angle-delay channel frequency response matrix ADCFR to generate a historical angle-delay channel frequency power fingerprint matrix ADCFP.

[0208] Optionally, the inner product module includes: a product submodule, used to multiply each row and each column element in the historical angle-delay channel frequency response matrix ADCFR with the corresponding element of the phase shift matrix, and sum each product value to obtain each row and each column element of the historical angle-delay channel frequency power fingerprint matrix ADCFP; an inner product submodule, used to perform a Hadamard inner product operation on the historical angle-delay channel frequency power fingerprint matrix ADCFP to generate a historical angle-delay channel frequency power fingerprint matrix ADCFP, wherein the historical angle-delay channel frequency response matrix ADCFR is a complex matrix, and the historical angle-delay channel frequency power fingerprint matrix ADCFP is a real matrix ADCFP.

[0209] Optionally, the vehicle positioning device based on fingerprint positioning also includes: a fingerprint matrix compression unit, which is used to compress the ADCFP fingerprint matrix row by row to obtain the horizontal channel frequency power matrix HCFP after collecting historical signal features at a known position, and compress the ADCFP fingerprint matrix column by column to obtain the vertical channel frequency power matrix VCFP; a matrix combination unit, which is used to combine the horizontal channel frequency power matrix HCFP and the vertical channel frequency power matrix VCFP to generate a horizontal and vertical channel frequency fingerprint matrix HVCFF.

[0210] Optionally, the vehicle positioning device based on fingerprint positioning also includes: an interval division unit, which is used to store the fingerprint matrix in a preset database to obtain the fingerprint database, and then use the center arrival angle clustering algorithm to divide the fingerprint database into intervals to obtain multiple cluster intervals; a fingerprint parameter calculation unit, which is used to calculate the center arrival angle CAOA of each reference fingerprint in the fingerprint database; a fingerprint interval division unit, which is used to divide the reference fingerprint into corresponding cluster intervals according to the size of the center arrival angle CAOA of each reference fingerprint.

[0211] Optionally, the fingerprint matching unit includes: a similarity calculation module, which is used to calculate the similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database, and obtain a matching result based on the similarity, wherein Manhattan distance or Euclidean distance is used as the similarity in the process of calculating the similarity.

[0212] Optionally, the vehicle positioning unit includes: a similarity sorting module, which is used to sort the similarities in the matching results and select the K fingerprints with the highest similarity as candidate matching fingerprints, wherein the K value is an integer greater than 1; a weighting module, which is used to adopt an improved power WKNN matching algorithm to assign different weighting coefficients to each candidate matching fingerprint according to the distance between the candidate matching fingerprint and the fingerprint of the point to be positioned; and a positioning module, which is used to calculate the weighted average of the position coordinates corresponding to the candidate matching fingerprints according to the weighting coefficients as the position coordinates of the armored car to be positioned.

[0213] The above-mentioned vehicle positioning device based on fingerprint positioning may also include a processor and a memory. The above-mentioned fingerprint matrix generation unit 71, fingerprint matching unit 72, vehicle positioning unit 73, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0214] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the cash transport vehicle positioning based on the fingerprint matching result can be achieved by adjusting the kernel parameters.

[0215] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0216] Embodiment 3

[0217] An embodiment of the present application may provide an electronic device, Figure 8 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 8 As shown, the electronic device may include: one or more ( Figure 8 (only one is shown) processor 802, memory 804, storage controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0218] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the vehicle positioning method and device based on fingerprint positioning in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned vehicle positioning method based on fingerprint positioning. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0219] The processor can call the information and application program stored in the memory through the transmission device to perform the following steps: using multiple receiving antennas of the base station to detect the uplink signal characteristics emitted by the cash transport vehicle to be located, generating a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized as the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP includes the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be located to the receiving antenna of the base station, which is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation; matching the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprint in the pre-established fingerprint database to obtain a matching result; and locating the position coordinates of the cash transport vehicle to be located using the improved power WKNN matching algorithm according to the matching result.

[0220] It can be understood by those skilled in the art that Figure 8 The structure shown is for illustration only, and the electronic device may also be a smart phone, a tablet computer, a PDA, a mobile Internet device (MID), a PAD or other terminal device. Figure 8 The structure of the electronic device is not limited. Figure 8 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 8Different configurations shown.

[0221] A person skilled in the art can understand that all or part of the steps in the various fingerprint-based vehicle positioning methods of the above-mentioned embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0222] Embodiment 4

[0223] The embodiment of the present application further provides a storage medium. Optionally, in this embodiment, the storage medium can be used to store the program code executed by the vehicle positioning method based on fingerprint positioning provided in the first embodiment.

[0224] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the vehicle positioning methods based on fingerprint positioning in the above-mentioned embodiment 1.

[0225] Optionally, in this embodiment, the above storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0226] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the vehicle positioning method based on fingerprint positioning described in each embodiment of the present application.

[0227] The present application also provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vehicle positioning method based on fingerprint positioning described in each embodiment of the present application are implemented.

[0228] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0229] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0230] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0231] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0232] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A vehicle positioning method based on fingerprint positioning, characterized in that: include: Multiple receiving antennas of a base station are used to detect the characteristics of the uplink signal emitted by the cash transport vehicle to be located, and a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP are generated based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized as the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP contains the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be located arriving at the receiving antenna of the base station, and is obtained from the CFR fingerprint through the phase shift matrix and the inner product transformation; The extracted angle-delay channel frequency power fingerprint matrix ADCFP is matched with the fingerprints in the pre-established fingerprint database to obtain a matching result; According to the matching results, the improved square WKNN matching algorithm is used to locate the position coordinates of the cash transport vehicle to be located.

2. The vehicle positioning method according to claim 1, characterized in that: The steps of using multiple receiving antennas of a base station to detect the characteristics of an uplink signal emitted by the cash transport vehicle to be located include: Multiple receiving antennas in the base station are used to detect the signal emitted by the money transport vehicle to be located, and the uplink signal characteristics of the money transport vehicle to be located are obtained, wherein the base stations are arranged according to a uniform linear array model, and the number of receiving antennas in each base station is greater than the number of money transport vehicles to be located distributed in the area to which the base station belongs, and the fingerprint information of the user to be located on the money transport vehicle to be located is transmitted to the receiving antenna along multiple paths.

3. The vehicle positioning method according to claim 1, characterized in that: The fingerprint database is constructed in the following way: Collecting historical signal features at a known position, wherein the historical signal features include: a historical channel impulse response fingerprint matrix CIR, a historical channel frequency response fingerprint matrix CFR, and a historical angle-delay channel frequency power fingerprint matrix ADCFP improved based on the CFR fingerprint; Extracting fingerprint information from the preprocessed historical signal features to generate a fingerprint matrix, wherein the fingerprint matrix includes the corresponding relationship between the historical angle-delay channel frequency power fingerprint matrix ADCFP and the coordinates of the corresponding known positions; The fingerprint matrix is ​​stored in a preset database to obtain the fingerprint database, wherein the fingerprint database stores a corresponding relationship between each reference fingerprint and a corresponding position coordinate.

4. The vehicle positioning method according to claim 3, characterized in that: The steps of collecting historical signal characteristics at a known location include: Perform channel estimation on historical uplink signals received in the historical process to obtain the historical channel impulse response fingerprint matrix CIR, wherein the matrix size of the historical channel impulse response fingerprint matrix CIR is determined by the number of receiving antennas of the base station and the length of the channel impulse response; Extracting angle of arrival AOA and time of arrival TOA information from the historical channel impulse response fingerprint matrix CIR; The historical channel impulse response fingerprint matrix CIR is converted into the historical channel frequency response fingerprint matrix CFR by Fourier transform to form a frequency domain signal feature, wherein the matrix size of the historical channel frequency response fingerprint matrix CFR is determined by the number of receiving antennas and the number of transmitting antennas of the base station; The historical channel frequency response fingerprint matrix CFR is transformed from the frequency domain to the angle domain by using a phase shift matrix to obtain a historical angle-delay channel frequency response matrix ADCFR, wherein the phase shift matrix is ​​obtained based on a uniform linear array model corresponding to the antenna array in the base station; A preset inner product operation is performed on the historical angle-delay channel frequency response matrix ADCFR to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP.

5. The vehicle positioning method according to claim 4, characterized in that: The step of performing a preset inner product operation on the historical angle-delay channel frequency response ADCFR to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP includes: Each row and column of the historical angle-delay channel frequency response matrix ADCFR is multiplied by the corresponding element of the phase shift matrix, and each product value is summed to obtain each row and column of the historical angle-delay channel frequency power fingerprint matrix ADCFP; A Hadamard inner product operation is performed on the historical angle-delay channel frequency power fingerprint matrix ADCFP to generate the historical angle-delay channel frequency power fingerprint matrix ADCFP, wherein the historical angle-delay channel frequency response matrix ADCFR is a complex matrix, and the historical angle-delay channel frequency power fingerprint matrix ADCFP is a real matrix ADCFP.

6. The vehicle positioning method according to claim 3, characterized in that: After collecting historical signal characteristics at known locations, it also includes: The ADCFP fingerprint matrix is ​​compressed row by row to obtain the horizontal channel frequency power matrix HCFP, and the ADCFP fingerprint matrix is ​​compressed column by column to obtain the vertical channel frequency power matrix VCFP; The horizontal channel frequency power matrix HCFP and the vertical channel frequency power matrix VCFP are combined to generate a horizontal vertical channel frequency fingerprint matrix HVCFF.

7. The vehicle positioning method according to claim 3, characterized in that: After storing the fingerprint matrix in a preset database to obtain the fingerprint database, the method further includes: Using a center arrival angle clustering algorithm to partition the fingerprint database into intervals to obtain a plurality of clustering intervals; Calculate the central angle of arrival CAOA of each reference fingerprint in the fingerprint database; According to the size of the central arrival angle CAOA of each reference fingerprint, the reference fingerprint is divided into corresponding cluster intervals.

8. The vehicle positioning method according to claim 1, characterized in that: The step of matching the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprint in the pre-established fingerprint database to obtain the matching result includes: The similarity between the ADCFP fingerprint matrix and each reference fingerprint in the fingerprint database is calculated, and the matching result is obtained based on the similarity, wherein the Manhattan distance or the Euclidean distance is used as the similarity in the process of calculating the similarity.

9. The vehicle positioning method according to claim 1, characterized in that: According to the matching result, the step of locating the position coordinates of the to-be-located cash transport vehicle using an improved square WKNN matching algorithm comprises: Sorting the similarities in the matching results, and selecting K fingerprints with the highest similarity as candidate matching fingerprints, where the K value is an integer greater than 1; Using the improved square WKNN matching algorithm, different weighting coefficients are assigned to each candidate matching fingerprint according to the distance between the candidate matching fingerprint and the fingerprint of the point to be located; According to the weighting coefficient, the weighted average value of the position coordinates corresponding to the candidate matching fingerprints is calculated as the position coordinates of the money transport vehicle to be located.

10. A vehicle positioning device based on fingerprint positioning, characterized in that: include: The fingerprint matrix generating unit is used to detect the uplink signal characteristics emitted by the cash transport vehicle to be located by using multiple receiving antennas of the base station, and generate a channel impulse response fingerprint matrix CIR, a channel frequency response fingerprint matrix CFR and an angle-delay channel frequency power fingerprint matrix ADCFP based on the uplink signal characteristics, wherein the channel impulse response fingerprint matrix CIR is obtained by performing channel estimation on the received uplink signal, the channel frequency response fingerprint matrix CFR is obtained by performing Fourier transform on the CIR fingerprint, the channel frequency response fingerprint matrix CFR is characterized as the sum of the CIRs of each path under different delays, and the angle-delay channel frequency power fingerprint matrix ADCFP contains the channel power information and the arrival angle AOA information of the signal sent by the cash transport vehicle to be located arriving at the receiving antenna of the base station, and is obtained by the CFR fingerprint through the phase shift matrix and the inner product transformation; A fingerprint matching unit is used to match the extracted angle-delay channel frequency power fingerprint matrix ADCFP with the fingerprints in a pre-established fingerprint database to obtain a matching result; The vehicle positioning unit is used to locate the position coordinates of the cash transport vehicle to be located by using an improved square WKNN matching algorithm according to the matching results.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the vehicle positioning method based on fingerprint positioning as described in any one of claims 1 to 9.

12. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle positioning method based on fingerprint positioning as described in any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the vehicle positioning method based on fingerprint positioning as described in any one of claims 1 to 9 are implemented.

Images