Low-altitude aircraft trajectory detection and channel tracking method based on super-large-scale MIMO

Abstract

The invention discloses a low-altitude aircraft trajectory detection and channel tracking method based on super-large-scale MIMO, and the method comprises the steps: firstly transmitting a pilot signal to a base station by a user aircraft, designing a beam forming algorithm under a full-connection architecture, and then designing a beam forming algorithm capable of obtaining a channel image with obvious key point features under low pilot overhead, obtaining a channel image based on the hybrid beam forming architecture; the method comprises the following steps: designing an image key point detection algorithm, obtaining position information of an aircraft in an observation area at a first moment and performing fine adjustment, designing an efficient channel tracking module at a subsequent moment, then completing channel reconstruction and trajectory matching based on the estimated position information of the aircraft, and finally designing a trajectory-based low-altitude aircraft supervision algorithm. And track error detection, aircraft signal interruption detection, unauthenticated intrusion aircraft detection and the like of the aircraft in the authentication range of the base station are realized. According to the method, the uplink channel can be quickly estimated, and the real-time performance and reliability of a communication link between the aircraft and the base station are guaranteed.

Description

Technical Field

[0001] The invention relates to a low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO, and belongs to the technical field of wireless communications. Background Art

[0002] The sixth generation of mobile communication technology (6G) has been widely studied around the world. As the next generation of communication technology, it will provide higher network speed, connection density and intelligent capabilities, greatly promoting the development potential of the digital economy. With the continuous advancement of technology and the expansion of application scenarios, the development of the low-altitude economy has gradually become one of the important directions to promote the development of the next generation of digital economy. The growth of market demand, breakthroughs in technological innovation and the improvement of the industrial chain have promoted the rapid development of the low-altitude economy. The application of low-altitude aircraft (such as drones, electric vertical take-off and landing aircraft, etc.) in logistics distribution, agricultural monitoring, urban air travel and other fields is changing the traditional industrial landscape. At the same time, the improvement of relevant policies and regulations has laid a solid foundation for the healthy and sustainable development of the low-altitude economy. Through the development of low-altitude airspace, the low-altitude economy is becoming an important engine to promote economic growth, improve industrial efficiency and promote technological innovation.

[0003] The development of the low-altitude economy has not only promoted the opening and management of low-altitude airspace, but also spawned the growth of demand for low-altitude flight communications. Especially with the support of 6G networks, the communication demand of low-altitude aircraft is showing explosive growth. Low-altitude aircraft have an extremely urgent need for high-speed and low-latency communications. As one of the key technologies of 6G, the application of ultra-large-scale MIMO technology provides a wealth of scenarios. Ultra-large-scale MIMO technology combined with the application of low-altitude economy can provide more efficient communication support and positioning applications for low-altitude aircraft. Through space division multiplexing technology, ultra-large-scale antenna arrays can effectively improve the communication and positioning capabilities of low-altitude aircraft. However, with the rapid increase in array size, the traditional all-digital hardware architecture faces huge hardware cost pressure, especially in ultra-large-scale MIMO systems, the hardware overhead of transceivers has risen sharply, becoming a major obstacle to its technology popularization and practical application. At the same time, with the increase in array size and the generation of near-field effects, the complexity of user positioning and channel estimation increases, which further increases the challenges of ultra-large-scale MIMO in practical applications. In order to reduce hardware costs and computational complexity while ensuring user positioning and channel estimation accuracy, hybrid beamforming architecture has become a potential solution. Hybrid beamforming technology reduces the number of RF links and the hardware cost of the system through a layered design of signal processing between RF and baseband. At the same time, the dynamic characteristics of low-altitude aircraft make ultra-large-scale MIMO systems face the challenge of time-varying channels. It is necessary to combine advanced technologies such as deep learning and image processing to reduce the computational complexity of time-varying channels and improve spectrum efficiency.

[0004] The combination of 6G ultra-large-scale MIMO technology and low-altitude economy will bring huge innovation potential to future communication systems. Through technological innovation and reasonable architecture design, it is expected to overcome the existing bottlenecks of hardware cost and computational complexity, promote the widespread deployment of ultra-large-scale MIMO technology in practical applications, and provide strong communication guarantee for the low-altitude economy in the 6G era. Therefore, in view of the demand for ultra-large-scale MIMO in low-altitude economic applications, it is particularly important to study low-cost hardware architecture, hybrid beamforming algorithms, and low-complexity aircraft trajectory and channel tracking solutions. These technologies will help improve the communication and positioning quality of low-altitude aircraft, promote the rapid development of the low-altitude economy, and further promote the application of 6G technology in the low-altitude economy, providing support for future communication needs. Summary of the invention

[0005] The technical problem to be solved by the present invention is: to provide a low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO, to design a beamforming matrix that can restore a channel image with obvious characteristics, and to restore the channel image from a base station received signal, to design an image-based key point detection network, to obtain the position of the aircraft, to design a channel extrapolation scheme, and to design the trajectory tracking of the aircraft based on the position information at all times, so as to achieve low-cost, low-complexity aircraft identification and trajectory tracking.

[0006] The present invention adopts the following technical solutions to solve the above technical problems:

[0007] A low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO, the ultra-large-scale MIMO system in a preset observation area includes a base station, an authenticated user aircraft and S scatterer aircraft, the S scatterer aircraft include authenticated aircraft and unauthenticated aircraft; the base station is equipped with M uniform linear array antennas and adopts a hybrid beamforming architecture; the base station antenna is divided into Q sub-arrays, the number of antennas in each sub-array is M / Q, and the number of pilots is M / 2Q;

[0008] The trajectory detection and channel tracking method comprises the following steps:

[0009] Step 1: The user aircraft sends a pilot signal to the base station. The base station designs a beamforming matrix under the hybrid beamforming architecture to recover the channel image from the base station received signal.

[0010] Step 2: Use the key point detection network to perform network inference on the channel image recovered from the base station received signal to obtain the rough estimated position of the user aircraft and all scatterer aircraft;

[0011] Step 3: For the channel estimation at the first moment, the rough estimated position obtained in step 2 is optimized by using orthogonal matching pursuit optimization to obtain the estimated position; the path gain is obtained by using the least squares algorithm for the base station received signal to achieve channel estimation;

[0012] Step 4: For the channel tracking at the subsequent time, determine whether there is a path missed detection at the current time according to the received signal. If yes, perform channel estimation at the current time according to the channel estimation method at the first time in step 3. Otherwise, perform channel extrapolation using orthogonal matching pursuit optimization based on the aircraft position estimated at the previous time and the movable range of the aircraft to achieve channel tracking.

[0013] Step 5: Based on the estimated positions of each aircraft at all times, the Hungarian algorithm is used to match and associate the positions at different times to obtain the movement trajectory of each aircraft;

[0014] Step 6: The base station compares the movement trajectory of each aircraft with the paths of the authenticated user aircraft and the authenticated aircraft to correct the trajectory of each aircraft.

[0015] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0016] 1. The present invention can adapt low-complexity aircraft position estimation and channel estimation based on image key point detection with low hardware cost and pilot overhead by designing a hybrid beamforming matrix.

[0017] 2. The present invention uses high-precision image key point target detection to estimate the positions of all aircraft at one time, effectively reducing the complexity of the method while achieving channel estimation accuracy that exceeds that of an orthogonal matching scheme under a codebook of the same size.

[0018] 3. The present invention uses the Hungarian algorithm to match the estimated positions at different times, thereby effectively completing the trajectory tracking of the aircraft.

[0019] 4. The present invention can effectively and accurately implement trajectory error detection, signal interruption detection and intrusion detection for low-altitude aircraft by designing an aircraft monitoring algorithm based on estimated trajectory. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of a low-altitude network under a very large-scale MIMO system provided by the present invention;

[0021] Figure 2 It is a schematic diagram of an aircraft position detection solution based on a key point detection network provided by the present invention;

[0022] Figure 3It is a flow chart of a low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO provided by the present invention. DETAILED DESCRIPTION

[0023] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0024] In the present invention, the ultra-large-scale MIMO system in the preset observation area is composed of a base station, an authenticated user aircraft and S scatterer aircraft, such as Figure 1 As shown. Among them, scatterer aircraft include certified aircraft and uncertified aircraft. The base station is equipped with M uniform linear array antennas and adopts a hybrid beamforming architecture. The base station antenna is divided into Q sub-arrays, the number of antennas in each sub-array is M / Q, and the number of pilots is M / 2Q.

[0025] The channel between the base station and the user can be expressed as:

[0026]

[0027] Where L(t) represents the number of paths at time t, that is, the number of aircraft in the observation area, g l (t) represents the path complex gain of the lth path at time t, (z l (t),x l (t)) represents the position of the lth aircraft at time t in the Cartesian coordinate system, a(z l (t),x l (t)) and d(z l (t),x l (t)) represent the steering vector and Doppler frequency shift vector respectively. The mth element in the steering vector can be modeled as:

[0028]

[0029] Among them, k c represents the wave number, D m (z,x) represents the distance between the aircraft and the mth antenna, which can be expressed as:

[0030]

[0031] Where d is the antenna spacing, i.e. half a wavelength. The mth element in the Doppler shift vector can be modeled as:

[0032]

[0033] Among them, v(t) represents the moving speed, Nsf Represents the number of subframes, T sf represents the subframe duration, φ m Represents the angle between the moving direction and the x-axis. The motion model of the aircraft in the time-varying scenario is defined as observing in the near-field area within a certain range, with 1 user aircraft and S scatterer aircraft, some of which can move during the observation time, and the other scatterer aircraft are stationary during the observation time. The paths at adjacent moments are respectively expressed with probability P a , P d 、(1-P a -P d ) There are new paths, annihilation, and the number of paths remains unchanged. The position relationship between the same aircraft at adjacent times is expressed as:

[0034] (z t+1 ,x t+1 )=(z t +v t T s ·sinα0(t+1),x t +v t T s ·cosα0(t+1))

[0035] Among them, α0 represents the angle between the motion direction and the x-axis. During initialization, the motion direction α0 is randomly initialized in the range of [0,2π]. At subsequent moments, the change of the motion direction remains within a certain range, that is, it satisfies

[0036] α l (t) = α l (t-1)(1+δ α (t)γ α,max )

[0037] Among them, γ α,max Represents the maximum range of the difference in motion direction between adjacent moments, δ α (t) represents the rate of change.

[0038] Based on the above background, Figure 3As shown, the present invention proposes a low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO, designs a beamforming matrix that can restore a channel image with obvious characteristics, and restores the channel image from the base station received signal, designs an image-based key point detection network, obtains the position of the aircraft, designs a channel extrapolation scheme, and designs the trajectory tracking of the aircraft based on all time position information, which can achieve low-cost, low-complexity aircraft identification and trajectory tracking, thereby accurately realizing trajectory error detection, signal interruption detection, and intrusion detection of low-altitude aircraft. In addition, it can achieve fast estimation of the uplink channel to ensure the real-time and reliability of the communication link between the user aircraft and the base station. Specifically, when the number of subarrays Q is 2, analog beamforming is designed and used. When the number of subarrays is larger, hybrid beamforming includes analog beamforming and digital beamforming. The received signal can be written as y(t), and the converted received image can be written as y img (t), the key point detection network is used to detect the positions of the user aircraft and the scatterer aircraft from the received image, and the least squares algorithm is used to obtain the path complex gain after optimization. The channel reconstruction can be completed by substituting it into the channel model. The channel tracking module uses the received signal and the estimated position of the aircraft at the previous moment to perform channel extrapolation. Then, based on all the estimated positions, the Hungarian algorithm is used to match the positions at different times and form the motion trajectories of all aircraft. Finally, the base station compares the known trajectory with the estimated trajectory to achieve trajectory error detection, signal interruption detection, and intrusion detection of low-altitude aircraft.

[0039] The specific steps include:

[0040] Step 1, beamforming stage: The ultra-large-scale MIMO system model is as follows Figure 1 As shown, the user aircraft sends a pilot signal to the base station. Under the partially connected molecular array hybrid beamforming architecture, when the number of subarrays Q is 2, the base station designs the simulated beamforming matrix W A is a diagonal matrix, which enables the channel image with obvious propagation path characteristics to be recovered from the received signal; that is,

[0041]

[0042] Q = 2, where are the first M / 4 rows and the last M / 4 rows in the M / 2-point discrete Fourier transform (DFT) matrix. The received signal can be expressed as in

[0043]

[0044] h1(t) and h2(t) are the first M / 2 elements and the last M / 2 elements of channel h, respectively.

[0045] When the number of subarrays Q is greater than 2, hybrid beamforming includes digital beamforming and analog beamforming, so that the channel image with the same quality as when the number of subarrays Q is 2 can be restored from the received signal. The received signal can be expressed as where F B It can be expressed as:

[0046]

[0047] It can be expressed as:

[0048]

[0049] M p,q It can be expressed as:

[0050]

[0051] W B It can be expressed as:

[0052]

[0053] When q is less than or equal to Q / 2, the analog beamforming matrix of the qth subarray can be expressed as:

[0054]

[0055] When q is greater than Q / 2, the analog beamforming matrix of the qth subarray can be expressed as:

[0056]

[0057] in, and They are respectively the first subarray and the second subarray of the simulated beamforming matrix when Q = 2. Under the simulated beamforming algorithm, Q (Q>2) subarrays can obtain the same received pilot signal as Q = 2 subarrays.

[0058] Under the hybrid beamforming matrix architecture, the near-field conversion domain code used can be expressed as in, It is a near-field conversion domain codebook under a fully digital architecture, in which each element is a steering vector at the sampling point position.

[0059] The channel image can be generated by the following formula:

[0060]

[0061] Step 2: Aircraft position estimation phase: Use Figure 2The key point detection network shown is trained on a pre-labeled training set. During testing, the channel image recovered from the base station received signal is input to obtain a rough estimate of the aircraft's position.

[0062] The network first extracts the general features of the image through the skeleton network ResNet-50, and then passes through three identical detection heads consisting of a 3*3 size, 64-channel convolution kernel, a normalization layer, a ReLU activation layer, and a 1*1 size, 2-channel convolution kernel. It then outputs three feature maps of the same length and width, representing the heat map, the coordinate prediction offset feature map, and the target size feature map. The value of each point on the heat map represents the probability that the center point of the target falls on that point. The coordinate prediction offset feature map represents the coordinate offset prediction value of each point on the corresponding heat map, and the target size feature map represents the target size prediction value of each point on the corresponding heat map. When doing inference, the network output will also pass through a score threshold filter, that is, delete the key points with a small probability of heat map prediction, thereby giving the final key point prediction value, that is, after one network inference, a rough estimate of the position of all user aircraft and scatterer aircraft is obtained.

[0063] Step 3, first moment channel estimation phase: based on the aircraft position information obtained in step 2 Design an orthogonal matching pursuit optimization of a small range and fine-grained codebook near each path to obtain accurate location information Use the least squares algorithm to get the path gain Implement channel estimation.

[0064] Step 4: Channel tracking at subsequent moments: based on the estimated position information at the previous moment Determine whether there is a path missed detection based on the received signal. If the residual signal satisfies the following formula, repeat step 3 to perform channel estimation at this moment:

[0065] ∥(a(x,z)) H y res (t)∥ 2 >=σ 2 (ln(M)-ln(-ln(1-P fa )))

[0066]

[0067] Among them, P fa is the false alarm probability. Otherwise, channel tracking can be completed by using orthogonal matching pursuit to perform channel extrapolation within the aircraft position detected at the last moment and the movable range of the aircraft.

[0068] Step 5, aircraft detection position association and trajectory matching stage: Based on the aircraft position information estimated at all times, the Hungarian algorithm is used to match and associate the detected positions at different times. All the positions estimated at the previous and next moments are regarded as two bipartite graphs, and the Hungarian algorithm is used to achieve the maximum matching of the bipartite graphs with a cost of 1-IOU, where IOU is the intersection-union ratio of the detection box. In this way, the optimal matching of the aircraft position estimation at different times is obtained, and the moving trajectory is formed.

[0069] Step 6, aircraft monitoring stage: The base station compares the estimated positions of user aircraft and scatterer aircraft with the authenticated aircraft path set on the base station. For authenticated aircraft, if its position is not detected at a certain moment, it is judged that the signal of the aircraft is interrupted at that moment, and the actual position of the aircraft is determined by the estimated trajectory information, and compared with the preset trajectory to realize trajectory error detection; for unauthenticated aircraft, if the position information detected at certain consecutive moments deviates greatly from all preset trajectories of authenticated aircraft of the base station, the aircraft is judged to be an intruder aircraft.

[0070] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aforementioned low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO are implemented.

[0071] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the aforementioned low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO are implemented.

[0072] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0073] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0074] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0075] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0076] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO, wherein the ultra-large-scale MIMO system in a preset observation area includes a base station, an authenticated user aircraft, and S scatterer aircraft, wherein the S scatterer aircraft include authenticated aircraft and unauthenticated aircraft; the base station is configured with M uniform linear array antennas and adopts a hybrid beamforming architecture; the base station antenna is divided into Q sub-arrays, each sub-array antenna number is M / Q, and the number of pilots is M / 2Q; It is characterized in that The trajectory detection and channel tracking method comprises the following steps: Step 1: The user aircraft sends a pilot signal to the base station. The base station designs a beamforming matrix under the hybrid beamforming architecture to recover the channel image from the base station received signal. Step 2: Use the key point detection network to perform network inference on the channel image recovered from the base station received signal to obtain the rough estimated position of the user aircraft and all scatterer aircraft; Step 3: For the channel estimation at the first moment, the rough estimated position obtained in step 2 is optimized by using orthogonal matching pursuit optimization to obtain an estimated position; The least squares algorithm is used to obtain the path gain of the base station received signal to achieve channel estimation; Step 4: For the channel tracking at the subsequent time, determine whether there is a path missed detection at the current time according to the received signal. If yes, perform channel estimation at the current time according to the channel estimation method at the first time in step 3. Otherwise, perform channel extrapolation using orthogonal matching pursuit optimization based on the aircraft position estimated at the previous time and the movable range of the aircraft to achieve channel tracking. Step 5: Based on the estimated positions of each aircraft at all times, the Hungarian algorithm is used to match and associate the positions at different times to obtain the movement trajectory of each aircraft; Step 6: The base station compares the movement trajectory of each aircraft with the paths of the authenticated user aircraft and the authenticated aircraft to correct the trajectory of each aircraft.

2. The method for low-altitude aircraft trajectory detection and channel tracking based on ultra-large-scale MIMO according to claim 1, characterized in that: The specific process of step 1 is as follows: The user aircraft sends a pilot signal to the base station. Under the hybrid beamforming architecture, when the number of subarrays Q is equal to 2, the simulated beamforming matrix W is designed. A ,Right now Among them, W A is a diagonal matrix, They are the first M / 4 rows and the last M / 4 rows of the M / 2-point discrete Fourier transform matrix; The base station received signal y(t) is expressed as: Among them, P r is the received power at the base station, n(t) is Gaussian white noise, h1(t) and h2(t) are the first M / 2 elements and the last M / 2 elements of the uplink channel h, respectively; h(t) is the uplink channel between the base station and the user aircraft, expressed as: Where L(t) represents the number of paths at time t, that is, the number of aircraft in the preset observation area, g l (t) represents the path complex gain of the lth path at time t, (z l (t),x l (t)) represents the position of the lth aircraft at time t in the Cartesian coordinate system, a(z l (t),x l (t)) and d(z l (t),x l (t)) represent the steering vector and Doppler frequency shift vector respectively; The mth element in the steering vector [a(z l (t),x l (t))] m Modeled as: The mth element in the Doppler frequency shift vector [d(z l (t),x l (t))] m Modeled as: Among them, D m (z l (t),x l (t)) represents the distance between the user aircraft and the mth antenna, d is the antenna spacing, i.e. half wavelength; k c is the wave number, v(t) is the moving speed, t sf is the subframe duration, N sf is the number of subframes, φ m is the angle between the moving direction and the x-axis; When the number of subarrays Q is greater than 2, the digital beamforming matrix F is designed B and the simulated beamforming matrix W B , so that a channel image with the same quality as when the number of subarrays Q is equal to 2 can be recovered from the base station received signal; F B It is expressed as: in, for dimensional all-zero matrix, W B It is expressed as: The analog beamforming matrix of the qth subarray is expressed as: The base station received signal y(t) is expressed as: in, h1(t), h2(t) and h Q (t) are the channels of the first, second and Qth subarrays respectively; The channel image is generated by: Among them, y′ img is the channel image, is the near-field transformed domain codebook under the hybrid beamforming architecture, It is the near-field conversion domain codebook under the all-digital architecture. is the hybrid beamforming matrix W Hybrid The pseudo-inverse of W Hybrid =F B ⊙W B .

3. The method for low-altitude aircraft trajectory detection and channel tracking based on ultra-large-scale MIMO according to claim 1, characterized in that: In the step 2, the key point detection network includes a skeleton network ResNet-50, a decoder and three detection heads with the same structure. The detection head includes a first convolutional layer, a normalization layer, a ReLU activation layer and a second convolutional layer connected in sequence. The convolution kernel size of the first convolutional layer is 3*3 and the channel is 64. The convolution kernel size of the second convolutional layer is 1*1 and the channel is 2. The skeleton network ResNet-50 is used to extract features of the channel image recovered from the signal received from the base station. The decoder decodes the extracted features and sends them to three detection heads. The three detection heads output feature maps of the same length and width, namely the heat map, the coordinate prediction offset feature map and the target size feature map. The value of each pixel on the heat map represents the predicted probability that the center point of the target falls on this point. The coordinate prediction offset feature map represents the coordinate offset prediction value of each point on the corresponding heat map. The target size feature map represents the target size prediction value of each point on the corresponding heat map. The score threshold filter is used to delete the key points in the heat map whose prediction probability is less than the preset threshold to obtain the final key point prediction value, which is the rough estimated position of the user aircraft and all scatterer aircraft obtained by network reasoning.

4. The method for low-altitude aircraft trajectory detection and channel tracking based on ultra-large-scale MIMO according to claim 2, characterized in that: In step 4, judging whether there is a path missed detection at the current moment according to the received signal is specifically as follows: If the residual signal satisfies the following equation, there is a path miss detection: ∥(a(z l (t),x l (t))) H y res (t)∥ 2 >=σ 2 (ln(M)-ln(-ln(1-P fa ))) Among them, P fa is the false alarm probability, is the estimated path gain, represents the estimated steering vector.

5. The method for low-altitude aircraft trajectory detection and channel tracking based on ultra-large-scale MIMO according to claim 1, characterized in that: The specific process of step 5 is as follows: Based on the estimated aircraft position at all times, the Hungarian algorithm is used to match and associate the positions at different times. That is, all the positions estimated at the previous and next moments are regarded as two bipartite graphs. The Hungarian algorithm is used to achieve the maximum matching of the bipartite graphs with a cost of 1-IOU. IOU is the intersection-union ratio of the detection box. In this way, the optimal matching of the aircraft position estimation at different times is obtained, and the movement trajectory is formed.

6. The method for low-altitude aircraft trajectory detection and channel tracking based on ultra-large-scale MIMO according to claim 1, characterized in that: The specific process of step 6 is as follows: The base station compares the estimated movement trajectories of user aircraft and scatterer aircraft with the preset paths of authenticated user aircraft and authenticated aircraft. If the position of an authenticated user aircraft or an authenticated aircraft is not detected at a certain moment, it is determined that the signal of the aircraft was interrupted at the moment when the position was not detected, and the estimated position at the moment of signal interruption is taken as the actual position of the aircraft, which is compared with the preset path to realize trajectory error detection. For an unauthenticated aircraft, if the movement trajectory formed by the positions detected at consecutive preset moments is inconsistent with the preset paths of the authenticated user aircraft and the authenticated aircraft of the base station, it is determined to be an intruding aircraft.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO are implemented as described in any one of claims 1 to 6.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the low-altitude aircraft trajectory detection and channel tracking method based on ultra-large-scale MIMO are implemented as described in any one of claims 1 to 6.

Images