Integrated positioning method based on spatial spectrum estimation and map matching

Abstract

The invention discloses an integrated positioning method based on spatial spectrum estimation and map matching, and provides a two-stage communication sensing integrated positioning method based on spatial spectrum estimation and map matching to construct a communication sensing integrated cellular system model based on orthogonal frequency division multiplexing. The devices in the system can send and receive communication signal waveforms at the same time so as to complete single-target device positioning. In particular, it is considered that an orthogonal frequency division multiplexing scheme is used for communication signal transmission, and thus the result is compatible with 5G and later cellular networks. In this arrangement, a two-stage positioning frame is proposed. In the first stage, the parallelism, global optimization and simplicity of a genetic algorithm are utilized, a traditional multi-signal classification algorithm is improved, and finally joint estimation of the distance, the direction of arrival (AOA) and the pitch angle of arrival (EOA) of a target is achieved.

Description

Technical Field

[0001] The present invention relates to the field of wireless communications, and in particular to an integrated positioning method based on spatial spectrum estimation and map matching. Background Art

[0002] The development of the sixth generation of mobile communication technology has spawned the emergence of many emerging location-aware applications, such as virtual reality, drone inspections, and autonomous driving. To this end, people have integrated radar sensing and communication functions to achieve shared spectrum, hardware, signaling, protocols, networks and other resources to meet the stringent requirements of emerging applications for the communication and perception performance of today's wireless systems, such as extremely high data rates, seamless coverage, high reliability, extremely low latency, and high-precision sensing. ISAC technology can not only bring significant benefits to future communication systems, such as using sensor information in millimeter wave systems to design effective beam selection and alignment, but also play a key role in sensing systems. How to use ISAC technology to perceive the environment and transmit perception data between vehicles to improve traffic efficiency and positioning accuracy is a technical problem that needs to be solved at this stage. Summary of the invention

[0003] The purpose of the present invention is to solve the above problems and to design an integrated positioning method based on spatial spectrum estimation and map matching.

[0004] To achieve the above object, the technical solution of the present invention is that, further, in the above-mentioned integrated positioning method based on spatial spectrum estimation and map matching, the crowd flow monitoring method comprises the following steps:

[0005] Constructing a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing, wherein the synaesthesia integrated cellular system model at least includes a BS base station, an MS mobile station and a UAV drone;

[0006] According to the initial location information obtained by the BS base station and the MS mobile station, the target location information is estimated by using a genetic multi-signal classification algorithm;

[0007] A map matching method is used to obtain multiple sensing location maps, and the maps are redrawn by overlapping coordinates on the coordinate axes of the corresponding passive targets. The map information from various locations in the cellular network is transmitted to the cloud server through the fronthaul link.

[0008] Furthermore, in the above-mentioned integrated positioning method based on spatial spectrum estimation and map matching, the construction of a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing includes:

[0009] M≥1 BS base stations, the BS base station is used for sensing and communication, and is one of the ISAC devices. The antennas on the base station are deployed in an L×P uniform rectangular array;

[0010] I ≥ 1 mobile stations, which consist of vehicle users and are one of the ISAC devices. They also have sensing and communication functions and are equipped with a L × P uniform rectangular array;

[0011] K ≥ 1 suspended UAVs, as passive targets in the air, have no communication and positioning capabilities and are considered as passive sensing targets in 3D positioning;

[0012] The synaesthesia integrated cellular system model also has a base station and a mobile station of the same URA for transmitting and receiving OFDM waves.

[0013] Furthermore, in the above-mentioned integrated positioning method based on spatial spectrum estimation and map matching, the target position information is estimated using a genetic multi-signal classification algorithm according to the initial position information obtained by the BS base station and the MS mobile station, including:

[0014] The three-dimensional positioning problem of suspended target UAVs is solved from three aspects: angle estimation, spatial filter design, and time delay estimation.

[0015] Furthermore, in the above-mentioned integrated positioning method based on spatial spectrum estimation and map matching, the angle estimation includes:

[0016] When the base station senses the surrounding environment, if there are K target drones, the antenna will receive K reflected signals. Let N and Δf (Hz) represent the number of subcarriers and subcarrier spacing of the downlink OFDM signal, respectively. The channel bandwidth is B = NΔf (Hz). Therefore, the OFDM signal transmitted by each antenna can be defined as:

[0017]

[0018] where d i Represents user data;

[0019] Before the start of each OFDM symbol, a cyclic prefix (CP) consisting of Q < N OFDM samples is inserted. The discrete transmission channel impulse response (CIR) can be expressed as:

[0020]

[0021] Where K is the number of target drones in the environment, α k is the complex gain of each path, τ k is the arrival path time delay;

[0022] In n OFDM sampling periods, the signal received at the base station sensor receiving antenna is expressed as:

[0023]

[0024] in, is an independent Gaussian white noise, is the incident signal vector of the uniform matrix array, θ k is the horizontal azimuth, is the pitch angle, here

[0025] The incident signal vector can be expressed as:

[0026]

[0027] in,

[0028] a L (μ)=[1,e jξ ,…,e j(L-1)μ ] T ,

[0029] a P (ξ)=[1,e jξ ,…,e j(P-1)ξ ] T , d x d y are the parallel distances between the antenna element and the x-axis and y-axis, respectively, is the carrier wavelength, [] T Represents the transpose symbol.

[0030] Assume that from The echo reflected back to the array from the direction satisfies the far-field assumption, and the signal received by the array can be written as:

[0031] Y(n)=A(θ)S(n)+ω(n)

[0032] in, Is The spatial signal direction matrix composed of is the spatial signal vector reflected by the K beam, is Gaussian white noise;

[0033] Assuming that these K echoes are statistically independent signals and have nothing to do with noise, the correlation matrix R of the antenna array receiving data is YY ={Y(n)Y(n) H}Use eigenvalue decomposition to obtain:

[0034]

[0035] Among them, R SS =E{S(n)S(n) H} is the correlation matrix of the spatial signal, and the variance is Gaussian white noise, I is a diagonal matrix; Us is the signal subspace, U v is the noise subspace, ∑ s is a diagonal matrix consisting of the eigenvalues ​​corresponding to the signal space, ∑ v is a diagonal matrix consisting of the eigenvalues ​​corresponding to the noise space; ideally, the noise subspace and the signal subspace are orthogonal:

[0036]

[0037] In practical applications, through N Snap The received data Y(n) obtained by the snapshot is time averaged:

[0038]

[0039] U obtained by its eigenvalue decomposition v and It may not be completely orthogonal. In order to estimate the actual direction (DOA) of the K-beam reflection echo, it is completed through the minimum optimization search:

[0040]

[0041] Furthermore, in the above-mentioned integrated positioning method based on spatial spectrum estimation and map matching, the spatial domain filtering includes:

[0042] The arrival angle in each direction is filtered using the MVDR spatial domain filter, that is, When filtering is performed in the direction, the reflected echo in that direction is retained, while to The reflected waves in the direction are suppressed, and the output of the MVDR spatial domain filter is:

[0043]

[0044] where w MVDR is the weight vector of the spatial filter;

[0045] In the MVDR spatial domain filter, from The echo incident in the direction is separated from all the echoes arriving at the array together, and the calculation of the weight vector should satisfy the following formula:

[0046]

[0047] The output of this filtering is:

[0048]

[0049] Repeat the filtering steps for other directions and finally get the filtered output:

[0050]

[0051] Furthermore, in the above-mentioned integrated positioning method based on spatial spectrum estimation and map matching, the delay estimation includes:

[0052] The spatial domain filtering method is used to separate the reflected echoes from different directions, and then the distance measurement method is used to estimate the time delay information in each beam of echoes. Based on the relationship between the propagation time and propagation distance of electromagnetic waves, the distance information of the target in the direction of these echoes can be calculated;

[0053] The correlation matrix of the received data after filtering output Using eigenvalue decomposition, the MUSIC estimated spectral profile can be calculated:

[0054]

[0055] It is then used as a fitness function to optimize the spectrum peak, and the time coordinate corresponding to the output spectrum peak is the TOA estimate;

[0056] Since the distance between antennas is significantly smaller than the distance from the target to the BS, the distance to the target can be calculated using the following method:

[0057]

[0058] Here, c represents the speed of light.

[0059] Further, a system for an integrated positioning method based on spatial spectrum estimation and map matching is implemented, characterized in that the map matching method is used to obtain multiple perception location maps, the maps are redrawn by overlapping coordinates on the coordinate axes of corresponding passive targets, and the map information from each location of the cellular network is transmitted to the cloud server through a fronthaul link, including:

[0060] The centroids of all common target drones are designated as the overlapping points of the map, and the average coordinates of the target drones sensed by the base station are expressed as Right now:

[0061]

[0062] Among them, x k ,y k and z k are the three-dimensional coordinates of the target drone perceived by the kth base station, K com is the total number of public target drones;

[0063] Regarding the three-dimensional coordinate calculation of the target drone, assuming that the coordinates of the base station are (x BS ,y BS ), then we have:

[0064]

[0065] The centroid position vector perceived by the BS can be written as:

[0066]

[0067] For the mobile station, calculate the centroid of the target drone and obtain the centroid position vector of the mobile station:

[0068]

[0069] Get the position vector of the mobile station relative to the base station, which is:

[0070]

[0071] The base station transmits map information from various locations in the cellular network to the cloud server through the fronthaul link to complete the mapping of the entire city's traffic.

[0072] The beneficial effect is that, by providing a two-stage communication-perception integrated positioning method based on spatial spectrum estimation and map matching, a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing is constructed, in which all devices in the system can simultaneously send and receive communication signal waveforms to complete the positioning of single-target devices. In particular, an orthogonal frequency division multiplexing scheme is considered for communication signal transmission, so the result is compatible with 5G and later cellular networks. A two-stage positioning framework is proposed under this setting. In the first stage, the traditional multi-signal classification algorithm is improved by utilizing the parallelism, global optimization and simplicity of the genetic algorithm, and finally the joint estimation of the distance, arrival direction (AOA) and arrival pitch angle (EOA) of the target is achieved. In the second stage, a map matching-based ISAC device positioning method is proposed, in which the base station / mobile station receives the location map of the passive target from the mobile station / base station through the communication function, and redraws the perception positioning map by overlapping the common target coordinates.

[0073] (1) Compared with traditional multi-signal classification methods, the designed genetic multi-signal classification algorithm has the advantages of strong global search capability, high adaptability and flexibility, high parallel processing efficiency, ability to optimize multi-objective problems and good robustness. It is suitable for processing classification problems with complex or unknown data distribution.

[0074] (2) The MVDR spatial domain filter used can suppress interference signals from undesired directions to the greatest extent while ensuring that the integrity of the desired signal is not affected. This feature greatly improves the signal-to-noise ratio while maintaining the quality of the desired signal, making it particularly suitable for scenarios that require high-precision signal processing. In addition, the MVDR filter can be adaptively adjusted based on the received signal, enhancing the system's ability to adapt to dynamic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0076] Figure 1 It is a schematic diagram of a first embodiment of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0077] Figure 2 A synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing and an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0078] Figure 3 It is a flow chart of a genetic multi-signal classification algorithm of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0079] Figure 4 It is a working principle diagram of an MVDR spatial domain filter of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0080] Figure 5 A cumulative distribution function diagram of MAE under different signal-to-noise ratios of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0081] Figure 6 A GA-MUSIC angle estimation diagram of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0082] Figure 7 A comparison diagram of different positioning algorithms of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0083] Figure 8 It is a GA-MUSIC delay estimation diagram of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention;

[0084] Fig. 9It is a perception map of an integrated positioning method based on spatial spectrum estimation and map matching in an embodiment of the present invention. DETAILED DESCRIPTION

[0085] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0086] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0087] Embodiment 1:

[0088] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1 As shown, an integrated positioning method based on spatial spectrum estimation and map matching, the integrated positioning method comprises the following steps:

[0089] Step 101: construct a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing, where the synaesthesia integrated cellular system model at least includes a BS base station, an MS mobile station and a UAV drone;

[0090] Specifically, in this embodiment:

[0091] A model of an orthogonal frequency division multiplexing-based interaceptive integrated cellular system is constructed, which consists of three entities, including base stations (BS), mobile stations (MS) and unmanned aerial vehicles (UAVs). There are M≥1 fixed base stations, which are used for both perception and communication and are one of the ISAC devices. The antennas on the base stations are deployed in an L×P uniform rectangular array. There are I≥1 mobile stations, which consist of vehicle users, as one of the ISAC devices, they also have perception and communication functions and are equipped with the same antenna array. There are also K≥1 suspended UAVs, which are passive targets in the air. They have no communication positioning capabilities and are only regarded as passive sensing targets in three-dimensional positioning. In this system, base stations and mobile stations with the same URA can send and receive OFDM waves to communicate with each other, that is, the base station sends downlink communication signals to the user. In addition, the base station and mobile station can also analyze and process the echo signals reflected by the surrounding target UAVs to accurately estimate the location of the target, thereby realizing the radar perception function. Using the target location information obtained by the base station and mobile station, a location map of the area around the target can be drawn. By uploading all the drawn maps to the cloud server, perception of the entire city can be achieved.

[0092] Step 102: Estimate the target location information using a genetic multi-signal classification algorithm based on the initial location information obtained by the BS base station and the MS mobile station;

[0093] Specifically, in this embodiment

[0094] The genetic multi-signal classification algorithm is used to accurately estimate the position of the target in step 1. Here, the three-dimensional positioning problem of the suspended target UAV will be solved from three aspects: angle estimation, spatial filter design, and time delay estimation.

[0095] (1) DOA estimation

[0096] When the base station senses the surrounding environment, if there are K target drones, the antenna will receive K reflected signals. Since OFDM cellular communication technology is quite mature, in the rest of this article, we mainly focus on how to use OFDM communication signals to perceive targets in ISAC cellular systems. Let N and Δf (Hz) represent the number of subcarriers and subcarrier spacing of the downlink OFDM signal, respectively, then the channel bandwidth is B = NΔf (Hz). Therefore, the OFDM signal transmitted by each antenna can be defined as:

[0097]

[0098] where d i Represents user data.

[0099] Before the start of each OFDM symbol, a cyclic prefix (CP) consisting of Q < N OFDM samples is inserted to eliminate inter-symbol interference (ISI). Due to the serious multipath phenomenon in the actual electromagnetic wave propagation environment, a multipath channel will automatically be formed between the base station transmitting antenna and the sensor receiving antenna. Each target acts as a scatterer, resulting in different signal delays and intensities. Therefore, the signal received at the sensor receiving antenna of each base station is the superposition of the receiver noise and the downlink OFDM signal sent by the base station and reflected by all K targets. We assume that the channel is slowly fading, that is, the cyclic prefix length of the OFDM signal is greater than the channel impulse response length, so that the channel remains unchanged within the symbol period. Therefore, the discrete transmission channel impulse response (CIR) can be expressed as:

[0100]

[0101] Where K is the number of target drones in the environment, α k is the complex gain of each path, τ k is the arrival path time delay.

[0102] Since the target drone is suspended in the air, accurate angular positioning information needs to consider both the direction of arrival (AOA) and the elevation angle of arrival (EOA). By analyzing the echo information collected by the uniform matrix antenna array, the signal received at the base station sensor receiving antenna in n OFDM sampling cycles is expressed as:

[0103]

[0104] in, is an independent Gaussian white noise, is the incident signal vector of the uniform matrix array, θ k is the horizontal azimuth, is the pitch angle, here

[0105] The incident signal vector can be expressed as:

[0106]

[0107] in, a L (μ)=[1,e jξ ,…,e j(L-1)μ ] T , a P (ξ)=[1,e jξ ,…,e j(P-1)ξ ] T , d x ,d yare the parallel spacings between the antenna element and the x-axis and y-axis, respectively, is the carrier wavelength, and []T represents the transposed sign.

[0108] Assume that from The echo reflected back to the array from the direction satisfies the far-field assumption, and the signal received by the array can be written as:

[0109] Y(n)=A(θ)S(n)+ω(n)

[0110] in, Is The spatial signal direction matrix composed of is the spatial signal vector reflected by the K beam, is Gaussian white noise.

[0111] Assuming that these K echoes are statistically independent signals and have nothing to do with noise, the correlation matrix R of the antenna array receiving data is YY ={Y(n)Y(n) H}Use eigenvalue decomposition to obtain:

[0112]

[0113] Among them, R SS =E{S(n)S(n) H} is the correlation matrix of the spatial signal, and the variance is Gaussian white noise, I is a diagonal matrix. U s is the signal subspace, U v is the noise subspace, ∑ s is a diagonal matrix consisting of the eigenvalues ​​corresponding to the signal space, Σ v is a diagonal matrix consisting of the eigenvalues ​​corresponding to the noise space. Ideally, the noise subspace and the signal subspace are orthogonal:

[0114]

[0115] In practical applications, through N Snap The received data Y(n) obtained by the snapshot is time averaged:

[0116]

[0117] U obtained by its eigenvalue decomposition v and It may not be completely orthogonal. In order to estimate the actual direction (DOA) of the K-beam reflection echo, it can be done through the minimum optimization search:

[0118]

[0119] Since the MUSIC algorithm is limited in real-time performance due to its large computational complexity and is prone to falling into local optimal solutions and premature convergence, a method combining genetic algorithm (GA) and MUSIC algorithm is introduced to optimize the search process of spectrum peaks.

[0120] (2) Spatial filtering

[0121] In the previous section, the angle estimation of K-beam reflected echoes has been obtained. However, although direct distance estimation can obtain the corresponding K distance values, they cannot correspond one-to-one with the K angles estimated previously. To solve this problem, the arrival angle in each direction is filtered using an MVDR (Minimum Variance Distortionless Response) spatial domain filter. The MVDR filter can enhance the signal in a specific direction while suppressing the signal, interference and noise in other directions. By repeating this step, each direction is filtered. That is, in the When filtering is performed in the direction, the reflected echo in that direction is retained, while to Finally, after the reflected echoes in each direction are stripped off one by one, the distance is estimated.

[0122] The output of the MVDR spatial domain filter is:

[0123]

[0124] where w MVDR is the weight vector of the spatial filter.

[0125] In the MVDR spatial filter, the weight vector is used to perform spatial filtering on the received signal to achieve minimum variance and maximum enhancement of the target signal of interest. The echo incident in the direction is separated from all the echoes arriving at the array together, and the calculation of the weight vector should satisfy the following formula:

[0126]

[0127] The output of this filtering is:

[0128]

[0129] Repeat the filtering steps for other directions and finally get the filtered output:

[0130]

[0131] (3)TOA estimation

[0132] Relying on the spatial filtering method in the previous section, the reflected echoes from different directions can be separated one by one. Then, the distance measurement method is used to estimate the time delay information in each beam of echoes, and the distance information of the target in the direction of these echoes can be calculated based on the relationship between the propagation time and propagation distance of electromagnetic waves.

[0133] Similarly, the correlation matrix of the received data after filtering output Using eigenvalue decomposition, the MUSIC estimated spectral profile can be calculated:

[0134]

[0135] It is then used as a fitness function to optimize the spectrum peak, and the time coordinate corresponding to the output spectrum peak is the TOA estimate.

[0136] Since the distance between antennas is significantly smaller than the distance from the target to the BS, the distance to the target can be calculated using the following method:

[0137]

[0138] Here, c represents the speed of light (in m / s).

[0139] Step 103: Acquire multiple sensing location maps using a map matching method, redraw the map by overlapping coordinates on the coordinate axes of the corresponding passive targets, and transmit the map information from each location of the cellular network to the cloud server via a fronthaul link.

[0140] Specifically, in this embodiment

[0141] The map matching method is used to obtain multiple perception location maps, and the map is redrawn by overlapping the coordinates on the coordinate axes of the corresponding passive targets. Finally, the map information from various locations in the cellular network is transmitted to the cloud server through the fronthaul link, thereby completing the mapping of the entire urban traffic.

[0142] By using the angle optimal estimation and distance estimation output by GA-MUSIC, a UAV point map near the base station can be generated. Similarly, following the same steps, the mobile station can also perceive a similar UAV point map.

[0143] The base station transmits the perceived map to the mobile station via a downlink communication signal. The mobile station then compares the two maps and identifies common points, indicating the location of the target drone, and then overlays them. Ultimately, the comprehensive map is redrawn to extend the perception range.

[0144] Considering that there may be multiple common target drones, we specify the centroids (coordinate centers) of all common target drones as the overlapping points of the map. The average coordinates of the target drones sensed by the base station are expressed as Right now:

[0145]

[0146] Among them, x k ,y k and z k are the three-dimensional coordinates of the target drone perceived by the kth base station, K com is the total number of public target drones.

[0147] Regarding the three-dimensional coordinate calculation of the target drone, assuming that the coordinates of the base station are (x BS ,y BS ), then we have:

[0148]

[0149] The centroid position vector perceived by the BS can be written as:

[0150]

[0151] So far, the map matching process is completed. The biggest advantage of doing this is that only one device is needed to locate the target.

[0152] For the mobile station, the same method can be used to calculate the centroid of the target drone. Similarly, the centroid position vector of the mobile station can be obtained:

[0153]

[0154] Therefore, the position vector of the mobile station relative to the base station is finally obtained, which is:

[0155]

[0156] Finally, the base station transmits the map information from various locations in the cellular network to the cloud server through the fronthaul link, thus completing the mapping of the entire city's traffic.

[0157] The beneficial effect is that, by providing a two-stage communication-perception integrated positioning method based on spatial spectrum estimation and map matching, a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing is constructed, in which all devices in the system can simultaneously send and receive communication signal waveforms to complete the positioning of single-target devices. In particular, an orthogonal frequency division multiplexing scheme is considered for communication signal transmission, so the result is compatible with 5G and later cellular networks. A two-stage positioning framework is proposed under this setting. In the first stage, the traditional multi-signal classification algorithm is improved by utilizing the parallelism, global optimization and simplicity of the genetic algorithm, and finally the joint estimation of the distance, arrival direction (AOA) and arrival pitch angle (EOA) of the target is achieved. In the second stage, a map matching-based ISAC device positioning method is proposed, in which the base station / mobile station receives the location map of the passive target from the mobile station / base station through the communication function, and redraws the perception positioning map by overlapping the common target coordinates.

[0158] (3) Compared with traditional multi-signal classification methods, the designed genetic multi-signal classification algorithm has the advantages of strong global search capability, high adaptability and flexibility, high parallel processing efficiency, ability to optimize multi-objective problems and good robustness. It is suitable for processing classification problems with complex or unknown data distribution.

[0159] (4) The MVDR spatial domain filter used can suppress interference signals from undesired directions to the greatest extent while ensuring that the integrity of the desired signal is not affected. This feature greatly improves the signal-to-noise ratio while maintaining the quality of the desired signal, making it particularly suitable for scenarios that require high-precision signal processing. In addition, the MVDR filter can be adaptively adjusted based on the received signal, enhancing the system's ability to adapt to dynamic environments.

[0160] Embodiment 2:

[0161] Step 1: Select a suitable array model. First, determine the array model, and then calculate the covariance matrix based on the received signal data. Next, perform eigenvalue decomposition to obtain the spectrum estimation expression of the MUSIC algorithm. This expression will be used as the basis for calculating individual fitness in the genetic algorithm.

[0162] Step 2: Determine the initial population. Initialize the key parameters of the genetic algorithm, including the size of the population, selection strategy, crossover probability, and mutation probability.

[0163] Step 3: Evaluate the fitness of each individual. Use the spectrum estimation expression of the MUSIC algorithm as the fitness function to calculate the fitness of each individual in the population.

[0164] Step 4: Determine the accuracy of target positioning. The individual advantage is determined based on the fitness. The higher the fitness, the more accurate the spectrum peak positioning.

[0165] Step 5: Construct the next generation population through selection, crossover and mutation operations.

[0166] Step 6: Determine the stopping condition. Stop the iteration when the number of iterations reaches the preset maximum value or the optimal solution of consecutive generations does not change much. Otherwise, repeat steps (4) and (5);

[0167] Step 7: Output the optimal solution. When condition (6) is met, the iteration stops and the individual with the highest fitness value is output as the final solution, i.e., the optimal spectrum peak location.

[0168] The embodiments of the present invention are described in detail as follows. The system simulation adopts Python's Pytorch framework and Matlab platform. The system uses the same uniform matrix array configuration of base stations and mobile stations, and uses the same parameters to transmit OFDM signals. For symbol modulation and OFDM subcarrier decomposition, we use a binary phase shift keying (BPSK) modulation scheme. The simulation also takes into account various factors, including electromagnetic attenuation, noise interference, and radar reflection cross section. The specific parameter configuration based on 5G New Radio (NR) is shown in the table.

[0169]

[0170] Figure 5 It is a cumulative distribution function (CDF) diagram of the mean absolute error (MAE) of distance measurement under different signal-to-noise ratios after multiple iterations in an embodiment of the present invention. It can be seen that when the signal-to-noise ratio is higher, the curve of the CDF of the ranging error usually shows a left-biased feature, that is, the center point of the error distribution is biased to one side of the true value, and the density of the tail is low, indicating that the situation with smaller errors is more common. From the figure, when SNR=35dB, the average ranging error estimated by the GA-MUSIC algorithm is 0.25439m, that is, the proportion of all data points less than this value is about 50%.

[0171] Figure 6 This is a two-dimensional angle estimation diagram based on the GA-MUSIC algorithm. The results show that the algorithm accurately estimates the horizontal azimuth and elevation angle information of the two source directions with an error of no more than 0.5°.

[0172] Figure 7This is a comparison chart of different positioning algorithms in the embodiment of the present invention. It can be seen from the figure that when the signal interval is very close, we place the drone with the azimuth angle fixed at 10° and the pitch angle at 20° and 23° respectively. It can be seen that because the pitch angles are too close, the MVDR algorithm cannot separate the angle information of the two received echoes, while the GA-MUSIC algorithm can still correctly resolve the signals with close intervals. This is because in the GA-MUSIC algorithm, the signal subspace is independent of the noise subspace, so even if the signal interval is very close, their eigenvectors in the subspace will not overlap. This allows the MUSIC algorithm to distinguish between different signals, thereby achieving higher-precision positioning, especially in dense signal environments. In contrast, the MVDR algorithm relies on the inverse matrix of the covariance matrix. When the signal interval is very close, the covariance matrix may become ill-conditioned, resulting in a decrease in accuracy.

[0173] Figure 8 This is a G-MUSIC estimated delay curve in an embodiment of the present invention, where the blue circle represents the real delay of the target drone, and the orange curve represents the estimated spectrum of the GA-MUSIC algorithm. Due to the existence of multipath effects, there may be multiple signal components from different paths in the received signal, and these signal components will overlap in the time domain, causing the received signal to present multiple peaks. Under such conditions, the first arrival path (FAP) in the received signal is used as the delay estimate of the target drone in the line-of-sight link (LOS), and the remaining peaks are the effects of the multipath effect. In addition, it can be seen that the error between the real delay and the TOA estimated by the GA-MUSIC algorithm does not exceed 0.7ns. Of course, the existence of other peaks under the multipath effect will also interfere with the positioning of the signal and reduce the estimation accuracy. In order to reduce the impact of the multipath effect, the signal is usually preprocessed before using the algorithm for delay estimation, such as using spatial smoothing technology.

[0174] Fig. 9 The target perception map between the base station and the mobile station is shown in the embodiment of the present invention, and the position vector from the mobile station to the base station is calculated. For comparison of the results, the relative true position is also shown The estimation error obtained is within 0.2 m. In addition, if the speed of the mobile station is known in advance and remains constant, the model can also make certain predictions about the movement trajectory of the mobile station.

[0175] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An integrated positioning method based on spatial spectrum estimation and map matching, characterized in that: The integrated positioning method comprises the following steps: Constructing a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing, wherein the synaesthesia integrated cellular system model at least includes a BS base station, an MS mobile station and a UAV drone; According to the initial location information obtained by the BS base station and the MS mobile station, the target location information is estimated by using a genetic multi-signal classification algorithm; A map matching method is used to obtain multiple sensing location maps, and the maps are redrawn by overlapping coordinates on the coordinate axes of the corresponding passive targets. The map information from various locations in the cellular network is transmitted to the cloud server through the fronthaul link.

2. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 1, characterized in that: The construction of a synaesthesia integrated cellular system model based on orthogonal frequency division multiplexing includes: M≥1 BS base stations, the BS base station is used for sensing and communication, and is one of the ISAC devices. The antennas on the base station are deployed in an L×P uniform rectangular array; I ≥ 1 mobile stations, which consist of vehicle users and are one of the ISAC devices. They also have sensing and communication functions and are equipped with a L × P uniform rectangular array; K ≥ 1 suspended UAVs, as passive targets in the air, have no communication and positioning capabilities and are considered as passive sensing targets in 3D positioning; The synaesthesia integrated cellular system model also has a base station and a mobile station of the same URA for transmitting and receiving OFDM waves.

3. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 1, characterized in that: The method of estimating the target location information by using a genetic multi-signal classification algorithm according to the initial location information acquired by the BS base station and the MS mobile station includes: The three-dimensional positioning problem of suspended target UAVs is solved from three aspects: angle estimation, spatial filter design, and time delay estimation.

4. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 2, characterized in that: The angle estimation comprises: When the base station senses the surrounding environment, if there are K target drones, the antenna will receive K reflected signals. Let N and Δf (Hz) represent the number of subcarriers and subcarrier spacing of the downlink OFDM signal, respectively. The channel bandwidth is B = NΔf (Hz). Therefore, the OFDM signal transmitted by each antenna can be defined as: where d i Represents user data; Before the start of each OFDM symbol, a cyclic prefix (CP) consisting of Q < N OFDM samples is inserted. The discrete transmission channel impulse response (CIR) can be expressed as: Where K is the number of target drones in the environment, α k is the complex gain of each path, τ k is the arrival path time delay; In n OFDM sampling periods, the signal received at the base station sensor receiving antenna is expressed as: in, is an independent Gaussian white noise, is the incident signal vector of the uniform matrix array, θ k is the horizontal azimuth, is the pitch angle, where 0≤θ k ≤2π, The incident signal vector can be expressed as: in, to L (μ)=[1,e jξ ,…,And j(L-1)μ ] T , a P (ξ)=[1,e jξ ,…,e j(P-1)ξ ] T , d x ,d y are the parallel distances between the antenna element and the x-axis and y-axis, respectively, is the carrier wavelength, [] T Represents the transpose symbol. Assume that from The echo reflected back to the array from the direction satisfies the far-field assumption, and the signal received by the array can be written as: Y(n)=A(θ)S(n)+ω(n) in, Is The spatial signal direction matrix composed of is the spatial signal vector reflected by the K beam, is Gaussian white noise; Assuming that these K echoes are statistically independent signals and have nothing to do with noise, the correlation matrix R of the antenna array receiving data is YY ={Y(n)Y(n) H }Use eigenvalue decomposition to obtain: Among them, R SS =E{S(n)S(n) H } is the correlation matrix of the spatial signal, and the variance is Gaussian white noise, I is a diagonal matrix; U s is the signal subspace, U v is the noise subspace, ∑ s is a diagonal matrix consisting of the eigenvalues ​​corresponding to the signal space, ∑ v is a diagonal matrix consisting of the eigenvalues ​​corresponding to the noise space; ideally, the noise subspace and the signal subspace are orthogonal: In practical applications, through N Snap The received data Y(n) obtained by the snapshot is time averaged: U obtained by its eigenvalue decomposition v and It may not be completely orthogonal. In order to estimate the actual direction (DOA) of the K-beam reflection echo, it is completed through the minimum optimization search:

5. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 2, characterized in that: The spatial domain filtering comprises: The arrival angle in each direction is filtered using the MVDR spatial domain filter, that is, When filtering is performed in the direction, the reflected echo in that direction is retained, while to The reflected waves in the direction are suppressed, and the output of the MVDR spatial domain filter is: where w MVDR is the weight vector of the spatial filter; In the MVDR spatial domain filter, from The echo incident in the direction is separated from all the echoes arriving at the array together, and the calculation of the weight vector should satisfy the following formula: The output of this filtering is: Repeat the filtering steps for other directions and finally get the filtered output:

6. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 2, characterized in that: The delay estimation includes: The spatial domain filtering method is used to separate the reflected echoes from different directions, and then the distance measurement method is used to estimate the time delay information in each beam of echoes. Based on the relationship between the propagation time and propagation distance of electromagnetic waves, the distance information of the target in the direction of these echoes can be calculated; The correlation matrix of the received data after filtering output Using eigenvalue decomposition, the MUSIC estimated spectral profile can be calculated: It is then used as a fitness function to optimize the spectrum peak, and the time coordinate corresponding to the output spectrum peak is the TOA estimate; Since the distance between antennas is significantly smaller than the distance from the target to the BS, the distance to the target can be calculated using the following method: Here, c represents the speed of light.

7. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 1, characterized in that: The method of obtaining multiple sensing location maps by using a map matching method, redrawing the maps by overlapping coordinates on the coordinate axes of corresponding passive targets, and transmitting map information from various locations of the cellular network to a cloud server via a fronthaul link includes: The centroids of all common target drones are designated as the overlapping points of the map, and the average coordinates of the target drones sensed by the base station are expressed as Right now: Among them, x k ,y k and z k are the three-dimensional coordinates of the target drone perceived by the kth base station, K com is the total number of public target drones; Regarding the three-dimensional coordinate calculation of the target drone, assuming that the coordinates of the base station are (x BS ,y BS ), then we have: The centroid position vector perceived by the BS can be written as: For the mobile station, calculate the centroid of the target drone and obtain the centroid position vector of the mobile station: Get the position vector of the mobile station relative to the base station, which is: The base station transmits map information from various locations in the cellular network to the cloud server through the fronthaul link to complete the mapping of the entire city's traffic.

8. The integrated positioning method based on spatial spectrum estimation and map matching according to claim 1, characterized in that: The integrated positioning method comprises the following steps: First, the array model is determined, and the covariance matrix is ​​calculated based on the received signal data. The spectrum estimation expression of the MUSIC algorithm is obtained by performing eigenvalue decomposition, which is used as the basis for calculating the individual fitness in the genetic algorithm. Initialize the key parameters of the genetic algorithm, including population size, selection strategy, crossover probability, and mutation probability; The spectral estimation expression of the MUSIC algorithm is used as the fitness function to calculate the fitness of each individual in the population. The advantage of an individual is determined by its fitness. The higher the fitness, the more accurate the location of the spectrum peak. Construct the next generation population through selection, crossover and mutation operations; When the number of iterations reaches the preset maximum value or the optimal solution of consecutive generations does not change much, the iteration is stopped; When the conditions are met, the iteration stops and the individual with the highest fitness value is output as the final solution, that is, the optimal spectrum peak location.

Images