A method for positioning a UAV based on beam scanning of a communication base station

Abstract

The application discloses a method for realizing unmanned aerial vehicle target positioning based on a communication base station beam scanning mode. The application aims at the safety problems and challenges brought by unmanned aerial vehicle "black flight" to public safety, privacy protection, aviation and key infrastructure, deploys a system on a communication base station, adopts a beam scanning method, and realizes positioning of the unmanned aerial vehicle. In the first stage, the base station periodically performs beam scanning on an observation area, and realizes coarse positioning of the unmanned aerial vehicle through echo data processing; in the second stage, based on the coarse positioning result in the first stage and in combination with a motion model, a complete motion track is acquired by using an optimization and filtering method, and two-dimensional positioning of the unmanned aerial vehicle is realized. The application has the advantages of communication and perception integration, low deployment cost of a hardware system and low algorithm complexity.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of electronic signal processing, and relates to a method for positioning unmanned aerial vehicles based on beam scanning of communication base stations. BACKGROUND

[0002] With the rapid development of information technology, unmanned aerial vehicle technology has expanded from the military field to the civilian field, and its application range is increasingly wide, covering aerial photography, agriculture, logistics, environmental protection and many other fields. Unmanned aerial vehicles have been widely used in military and civilian fields due to their easy operation, strong maneuverability and wide application range. However, with the continuous progress and wide application of unmanned aerial vehicle technology, the problem of unauthorized flight (i.e. "black flight") is becoming increasingly serious, bringing many challenges to society. "Black flight" refers to the illegal flight of unmanned aerial vehicles in no-fly zones, restricted flight zones or other sensitive areas without permission or in violation of regulations. Such behavior poses a serious threat to public safety, privacy protection, aviation safety and critical infrastructure. Therefore, how to effectively monitor unmanned aerial vehicles and prevent potential risks has become an important problem that needs to be solved.

[0003] Currently, the technologies for positioning unmanned aerial vehicles mainly include optical, radio, acoustic and radar technologies. These technologies have effects in specific scenarios, but also have obvious limitations: systems based on optical signals can directly obtain visual images of the target, but are greatly affected by the environment and weather, and are not suitable for long-distance application scenarios; systems based on radio have a high false alarm rate in complex communication environments; systems based on acoustics are greatly affected by environmental noise and have a short detection distance. Systems based on radar have a long detection distance and high precision, but the deployment cost is high.

[0004] The present application designs a method for positioning unmanned aerial vehicle targets using the beam scanning of communication base stations. The method is generally divided into two stages. In the first stage, the base station periodically performs beam scanning in the observation area to obtain the target scattering coefficient matrix, thereby realizing the coarse positioning of the base station. In the second stage, according to the target position obtained by coarse positioning, the motion model and position information are combined by the method of particle swarm optimization and extended Kalman filtering to optimize the coarse positioning point, obtain the complete point trajectory after optimization and prediction, and realize two-dimensional positioning of the unmanned aerial vehicle. SUMMARY

[0005] The present application aims to position unmanned aerial vehicle targets using the beam scanning of communication base stations. The method is generally divided into two stages. In the first stage, the base station scans to obtain the coarse positioning point of the target. In the second stage, according to the coarse positioning result, the motion model of the target is combined with the methods of particle swarm optimization and Kalman filtering to realize two-dimensional positioning of the unmanned aerial vehicle target. The method specifically includes the following steps:

[0006] The first stage is that the base station coarsely locates the unmanned aerial vehicle:

[0007] The first step is to set up the system and its working mode. The system is deployed on the base station in a side-mounted manner. The radar transceiver antenna is arranged at the same position in the system. The radar works in an electronic scanning mode. The transmitting platform is in a beam scanning working mode. It switches the beam in the elevation direction and the azimuth direction according to a fixed period on a preset monitoring plane. The receiving platform works in a fixed beam mode. The system working concept diagram is shown in Figure 1 .

[0008] The second step is that the transmitting platform transmits an OFDM signal in the millimeter wave band. The receiving platform receives the echo signal reflected by the target. After the carrier frequency is removed, the received echo signal can be expressed as

[0009]

[0010] In formula (1), t is the fast time, τ is the slow time, N is the number of subcarriers, f c is the carrier frequency, c is the speed of light, λ represents the wavelength, Δf is the subcarrier spacing, T is the pulse width, R(τ) represents the slant range between the target and the radar, represents the modulation function of the antenna pointing, θ P and represent the azimuth angle and the elevation angle of the target at the moment, respectively. Formula (1) can represent the convolution relationship of the target scattering and the point spread function. Using the matrix form, it can be expressed as

[0011] y=HsZ T (2)

[0012] Then the target scattering matrix is recovered. The solution of the target optimization function can be solved by the least square method as

[0013]

[0014] The third step is to find the position of the peak value of the amplitude of the target scattering coefficient matrix in formula (3). Then the two-dimensional coordinate position of the target is the two-dimensional coordinate position of the target scattering coefficient matrix corresponding to the peak value:

[0015]

[0016] Where x i and y i are the two-dimensional positioning coordinate points of the i-th frame target.

[0017] The second stage is the trajectory optimization of the unmanned aerial vehicle based on the coarse positioning parameters of the base station:

[0018] Fourth step: pre-filtering the rough positioning trajectory, smoothing the trajectory and eliminating unreasonable points. Apply average sliding filter to the entire trajectory to reduce the influence of unreasonable trajectory points on the trajectory and eliminate sharp changes. Then optimize the trajectory according to the speed threshold method, and use the average value obtained by the local average method to replace the unreasonable trajectory points.

[0019] Fifth step: generate point clouds around the two-dimensional coordinates of each positioning point obtained by rough positioning, and select points from each point cloud group to form a trajectory, as shown in Figure 2 The particle swarm optimization algorithm is used to optimize the trajectory of rough positioning according to the known target motion model, and the root mean square error (RMSE) of polynomial fitting is used as the cost function.

[0020] Sixth step: according to the optimized trajectory result, the position backward difference method is used to obtain the two-dimensional average motion speed of the moving target. The time sequence corresponding to the average speed and the average speed are used as the training data of the Gaussian process regression, and the complete time sequence is input to predict the speed of the first position. The speed between two positions is considered to be equal to the speed of the previous position.

[0021] Seventh step: extract distance information from the original echo signal. Equation (1) is used to perform FFT along the distance direction to compress the distance. The compressed signal is in the form of

[0022]

[0023] A1 is a constant coefficient term of signal amplitude. The histogram of the amplitude two-dimensional image of the compressed signal is counted, and the histogram bin size interval is determined by equation (6):

[0024]

[0025] σ is the standard deviation of the entire amplitude image, and n=NKL is the number of pixel points of the entire image. The quartile point of the histogram bin edge is determined as the threshold. The data points less than the threshold are set to zero, and the remaining data points are kept unchanged. Then the distance of the target is determined according to equation (7) along the peak value of each slow time τ:

[0026]

[0027] Finally, the obtained original distance information is subjected to an interpolation operation to ensure the smoothness of the distance while removing outliers.

[0028] Eighth step: using the extended Kalman filter algorithm, taking the two-dimensional coordinates of the target as the state variable, the two-dimensional velocity of the target as the control variable, and the distance measured by the base station as the observation value, iterative processing is carried out. By multiple prediction and update of the state variable and the observation value, the system error is gradually corrected, and finally the complete target motion trajectory after filtering and estimation is output.

[0029] The present application can achieve the following beneficial effects after adopting the above scheme: first, it can be directly deployed on the base station to realize communication and perception integration; second, the scheme of mounting the perception receiver on the base station side can reduce the requirements for hardware and reduce the deployment cost of the hardware system; third, the algorithm has low complexity. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 It is a geometric schematic diagram of the system working mode.

[0031] Figure 2 It is a particle generation schematic diagram.

[0032] Figure 3 It is a data processing flowchart. DETAILED DESCRIPTION

[0033] The first stage is the rough positioning of the base station to the unmanned aerial vehicle:

[0034] First step: the unmanned aerial vehicle positioning system is mainly divided into signal transmitting part, receiving part and information processing part. The system is deployed on the base station in a side-mounted manner, the radar transceiver antenna is arranged at the same position, and the radar working mode is electronic scanning mode. The transmitting platform opens the beam scanning working mode, switches the beam in the pitch direction and the azimuth direction in a fixed period on a pre-set monitoring plane, and the receiving platform adopts the fixed beam working mode, and its system working concept diagram is as shown in Figure 1 .

[0035] Second step: open the radar transmitting module beam scanning, receive the echo signal, and pretreat the echo signal. For a single target, the transmitting antenna transmits the OFDM signal in the millimeter wave band, the receiving platform receives the echo signal from the target, and after the carrier frequency is removed, the received echo signal can be expressed as

[0036]

[0037] In formula (8), t is fast time, tau is slow time, N is the number of subcarriers, f c is the carrier frequency, c is the speed of light, lambda represents the wavelength, delta f is the subcarrier spacing, T is the pulse width, R(tau) represents the slant range of the target and the radar, represents the modulation function of the antenna pointing, theta P and respectively, and the rectangular pulse function is

[0038]

[0039] The formula (8) can be expressed as a convolution relationship of target scattering and point spread function, which can be expressed in matrix form as

[0040] y=HsZ T (10)

[0041] Where y represents the echo matrix, s represents the target scattering coefficient matrix, H represents the azimuth convolution matrix, Z represents the elevation convolution matrix, and n represents the noise matrix. According to the periodicity of beam scanning being K pulses, the echo matrix is divided into two-dimensional signal sequences y1, y2, …, y L , L is the frame number, and the radar pattern matrix H and Z in a scanning period are known.

[0042] Solving the target scattering matrix is actually an optimization problem, and the regularization optimization problem is constructed as:

[0043]

[0044] Since two-dimensional linear inversion is difficult to implement, the matrix is vectorized into a one-dimensional inverse form. The signal is column vectorized as s i =vec(s i ), y i =vec(y i ), and the convolution matrix represents the Kronecker product of the matrix. For the two-dimensional signal matrix of each time point, the solution of the target optimization function can be solved by the least square method as

[0045]

[0046] Where I represents the unit matrix, λ is the regularization parameter, and the role is to ensure that the matrix is positive definite so that the inversion result is stable.

[0047] Third step: find the position of the peak value of the target scattering coefficient matrix amplitude in formula (12), then the two-dimensional coordinate position of the target is the two-dimensional coordinate position of the peak value corresponding to the target scattering coefficient matrix:

[0048]

[0049] Where x i , y i is the two-dimensional positioning coordinate point of the i-th frame target.

[0050] The second stage is the trajectory optimization of the unmanned aerial vehicle based on the coarse positioning parameters of the base station:

[0051] Fourth step: pre-filtering the rough positioning trajectory, smoothing the trajectory and eliminating unreasonable trajectory points. Apply average sliding filter to the whole trajectory to reduce the influence of unreasonable trajectory points on the trajectory and eliminate sharp changes. After eliminating obvious trajectory deviation values, there may still be some unreasonable positioning results in the trajectory. For trajectory optimization based on the speed threshold method, set the speed threshold values in the x direction and the y direction respectively, judge which points are unreasonable trajectory points according to the speed in each direction, and use the average value obtained by local averaging to replace the unreasonable trajectory points.

[0052] Fifth step: generate a point cloud around the two-dimensional coordinates of each positioning point obtained by rough positioning. The generated point cloud should be able to cover the error range of radar detection. Randomly select points from each point cloud group to form a trajectory As Figure 2 shown, select M trajectories as initial particle swarm, use particle swarm optimization algorithm, according to the known target motion model, use the root mean square error (RMSE) of polynomial fitting as the cost function, and optimize the trajectory of rough positioning.

[0053] First, initialize the particles, calculate the loss value of each particle through the loss function, and record the particle with the minimum loss value as the global optimal solution:

[0054]

[0055] Where E(·) represents the fitting error, and the P polynomial fitting method is adopted. The specific value of P depends on the motion model used. Within a certain range, the larger P is, the better the fitting effect of the nonlinear trajectory. Let any particle p i The x coordinate of the trajectory is the vector x i , and the y coordinate of the trajectory is the vector y i , which can be represented as

[0056]

[0057] Where n i is the fitting error, and it can be known that is the Vandermonde matrix of the vector x i . QR decomposition of the matrix W i can obtain an orthogonal matrix and an upper triangular matrix such that Without considering the fitting error n i , the coefficient vector is obtained by inversion

[0058]

[0059] The fitting error is

[0060] E(p i )=||n i ||2=||y i -W i b i ||2 (17)

[0061] Then the velocity of each particle is updated, and in order to prevent the particle from migrating to a far place due to too fast velocity, the velocity is limited in amplitude, and then the position of each particle is updated, and the local optimal solution and the global optimal solution are updated according to the new position of the particle. The velocity is repeatedly updated, and the local optimal solution and the global optimal solution are repeatedly updated, until the global optimal solution is no longer updated after a certain number of iterations, or the iteration count reaches the maximum iteration count, the algorithm is stopped.

[0062] Step 6: According to the trajectory result obtained by optimization, the two-dimensional average motion velocity of the moving target is obtained by using the position backward difference method The motion velocity of the first position can be obtained by Gaussian process regression: the time series corresponding to the velocity and the motion velocity of the two dimensions are used as the training data of the Gaussian probability model to train the model, and the complete time series is used as the input of the model, so that the fitted two-dimensional velocity sequence can be obtained. The velocity between two positions is considered to be equal to the velocity of the previous position.

[0063] Step 7: Extract the distance information from the original echo signal. The distance compression is performed by FFT along the distance direction of formula (8), and the compressed signal is in the form of

[0064]

[0065] A1 is a constant coefficient term of signal amplitude. A histogram of the amplitude two-dimensional image of the compressed signal is counted, and the histogram bin size interval is determined by formula (19):

[0066]

[0067] σ is the standard deviation of the entire amplitude image, and n=NKL is the number of pixel points of the entire image. The quartile point of the histogram bin edge is determined as the threshold value. The data points less than the threshold value are set to zero, and the remaining data points are kept unchanged. Then the distance of the target is determined according to the peak value of each slow time τ according to formula (20):

[0068]

[0069] Finally, the obtained original distance information is subjected to an interpolation operation to ensure the smoothness of the distance while removing outliers.

[0070] Step 8: using the extended Kalman filter algorithm, taking the two-dimensional coordinates of the target as the state variable X i , the two-dimensional velocity of the target as the control variable U i , and the distance r i measured by the base station as the observation value, the state transition equation is

[0071]

[0072] where Δt represents the velocity sampling interval. The observation equation is

[0073]

[0074] Repeat the prediction and update steps. By multiple predictions and updates of the state variable and the observation value, the system error is gradually corrected, and finally the complete target motion trajectory after filtering and estimation is output.

[0075] The entire data processing flow is shown in Figure 3 .

[0076] The above is only a specific implementation of the present application, any feature disclosed in the specification can be replaced by other equivalent or similar purpose alternative features unless specifically described; all features disclosed, or steps in all methods or processes, except for mutually exclusive features and / or steps, can be combined in any way.

Claims

1. A method for positioning unmanned aerial vehicle target based on communication base station beam scanning, comprising the following steps: First step: the system is deployed on the base station in a side-mounted manner, the radar transceiver antenna in the system is arranged at the same position, the radar works in an electronic scanning mode, the transmitting platform is started in a beam scanning mode, and the beam is switched in a fixed period on a preset monitoring plane in two dimensions of pitch and azimuth, and the receiving platform works in a fixed beam mode; Second step: after the transmitting platform transmits a signal, the receiving platform receives echo signals reflected by the target, and a least square method is used to solve a target scattering matrix; Third step: a position where a peak value of the target scattering coefficient matrix amplitude in the second step is located, and two-dimensional coordinate positions of the target corresponding to the peak value of the target scattering coefficient matrix are obtained; Fourth step: an average sliding filter is applied to perform a smoothing operation on the entire trajectory, a speed threshold method is used for trajectory optimization, and an average value obtained by using a local average method is used to replace unreasonable trajectory points; Fifth step: a point cloud is generated around each positioning point two-dimensional coordinate obtained by coarse positioning, a point is randomly selected from each point cloud group to form a trajectory, a plurality of trajectories are selected as initial particle groups, a particle swarm optimization algorithm is used, a known target motion model is used, a root mean square error of polynomial fitting is used as a cost function, and the trajectory of coarse positioning is optimized; Sixth step: according to the trajectory result obtained by optimization, a position backward difference method is used to obtain a two-dimensional average motion speed of the moving target, a time sequence corresponding to the average speed and the average speed are used as training data of Gaussian process regression, a first position speed is predicted by inputting a complete time sequence, and a speed between two positions is considered to be equal to a speed of a previous position; Seventh step: FFT is performed along the distance direction to perform distance compression, a histogram of a two-dimensional image of the compressed signal amplitude is counted, a quartile point of a bin edge of the histogram is determined as a threshold, data points less than the threshold are set to zero, remaining data points remain unchanged, a peak value is found along each slow time dimension of the compressed signal to determine a target distance, and finally an interpolation operation is performed on the obtained original distance information; Eighth step: an extended Kalman filter algorithm is used, two-dimensional coordinates of the target are used as state variables, two-dimensional speeds of the target are used as control variables, distances measured by the base station are used as observation values, iterative processing is performed, system errors are gradually corrected through multiple predictions and updates of the state variables and the observation values, and finally a complete target motion trajectory filtered and estimated is output.

Images