A perception-assisted beam tracking method for jittery drones

Abstract

The present invention discloses a perception-assisted jitter UAV beam tracking method, which comprises the following steps: establishing a jitter UAV beam tracking model; and combining an extended Kalman filter (EKF) and an enhanced particle swarm optimization (EPSO) for jitter UAV beam tracking. The perception-assisted jitter UAV beam tracking method solves the jitter UAV beam tracking problem by combining an extended Kalman filter (EKF) and an enhanced particle swarm optimization (EPSO), and utilizes a DFRC signal, thereby eliminating the need for a downlink pilot and improving the angle tracking accuracy, and can accurately achieve the beam tracking task while balancing the jitter effect of the UAV.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle beam tracking, and in particular to a perception-assisted jittered unmanned aerial vehicle beam tracking method. Background Art

[0002] The development of the sixth generation of wireless technology presents many key scenarios, among which the communication and perception integration (ISAC) technology based on unmanned aerial vehicles (UAVs) has attracted much attention. The future sixth generation (6G) network aims to provide users with ultra-reliable transmission and ultra-low latency, which may require the introduction of perception services. The integration of ISAC communication and perception functions in the same device in the same frequency band is a good solution to improve spectrum utilization and reduce hardware energy consumption. At present, due to the characteristics of high mobility, low cost, and easy deployment, UAVs are widely used in the fifth generation (5G) network to assist ground communications. At the same time, the use of multiple-input multiple-output (MIMO) technology enables UAVs to form sharp pencil-shaped spatial beams, which also puts higher requirements on the beam alignment between UAVs and integrated radar communication platforms (IRCPs). Therefore, real-time beam tracking is required to maintain a reliable link with less overhead. At present, beam tracking algorithms in ground millimeter wave MIMO system networks have been widely and deeply studied. For example, F. Liu et al. used the dual function radar and communication (DFRC) signal method for vehicle tracking, but this method is tailored for vehicle networks because the angle changes of vehicles are small and can be ignored. Y.Huangdengr et al. proposed a UAV beam tracking method with dynamically inserted pilots, which reduced the traditional pilot overhead but did not eliminate it, and did not consider the jitter effect of the UAV. W.Wang et al. used a learning-based method to solve the beam alignment problem between jittered UAVs and users. However, it assumes that the UAV is stationary and does not consider the beam tracking problem when the UAV is moving. In actual UAV communication scenarios, the inherent random gusts cause the UAV to have inevitable jitter, which may affect the beam tracking effect and even cause beam misalignment. At present, there is no relevant research to eliminate the negative impact of UAV jitter on beam tracking. At present, there is no research considering the beam tracking problem of ISAC UAV communication with jitter effect. Summary of the invention

[0003] In order to solve the above problems, the purpose of the present invention is to provide a perception-assisted jittered UAV beam tracking method to solve the negative impact of UAV jitter on beam tracking, and solve the beam tracking problem of jittered UAV by combining extended Kalman filter (EKF) and enhanced particle swarm optimization (EPSO). DFRC signals are used to eliminate the need for downlink pilots and improve the angle tracking accuracy, so that the beam tracking task can be accurately achieved while balancing the impact of UAV jitter.

[0004] In order to achieve the above object, the technical solution of the present invention is as follows:

[0005] The present invention provides a perception-assisted jitter UAV beam tracking method, comprising the following steps:

[0006] S1. Establish a jitter UAV beam tracking model;

[0007] S2. The extended Kalman filter (EKF) and enhanced particle swarm optimization (EPSO) are combined for beam tracking of jittered UAVs.

[0008] Furthermore, the S1 comprises the following steps:

[0009] S11. Establishing UAV jitter model:

[0010] S12. Establishing radar model;

[0011] S13. Establish a communication model;

[0012] S14. Establish a state transition model;

[0013] S15. Establish a measurement model.

[0014] Furthermore, the S11 comprises the following steps:

[0015] S111. Assume that uniform linear arrays (ULAs) are deployed on both the IRCP and the UAV, and the communication between the UAV and the IRCP is carried out through a line-of-sight LoS channel. The angle, distance, and speed of the UAV relative to the IRCP at the nth epoch are represented by θ n (t),d n (t) and v n (t), where all parameters are functions of t∈[0,T], where T is the maximum duration of interest. Without loss of generality, it is assumed that the UAV flies parallel to the antenna array of the IRCP;

[0016] S112. Use the following deterministic model to calculate the resulting pitch angle uncertainty;

[0017] φ n (t) = φ n (t-1)+Δφ n (t), (1)

[0018] Among them, φ n (t) and φ n (t-1) represents the elevation angle of two consecutive time slots t and (t-1) respectively represents the change of pitch angle at time t, The mean is 0 and the variance is The normal distribution of .

[0019] Furthermore, the S11 comprises the following steps:

[0020] S121. At time slot t, IRCP sends DFRC signal s to the drone n (t), the reflected echo on the IRCP side is expressed as

[0021]

[0022] in, p and τ represent the array gain factor, transmit power and delay respectively. is zero and has a variance of Gaussian complex noise, represents the Doppler frequency, the transmit beamforming matrix It is designed based on angle prediction, where represents the one-step angle prediction of IRCP at the nth epoch, represents the reflection coefficient, where d n and ∈ represent the distance between IRCP and UAV and the radar cross section RCS of UAV, respectively;

[0023] S122. Without loss of generality, the sending and receiving steering vectors of the IRCP and the UAV are as follows

[0024]

[0025]

[0026] Assuming that the antenna spacing of the ULA is half a wavelength, the transmit SNR is

[0027] Further, the S13 comprises the following steps:

[0028] S131. Express the arrival angle AOA from IRCP to the UAV as γ n =φ n +θ n , where θ n Represents the angle of the drone relative to the IRCP. The IRCP communicates with the drone through a line-of-sight channel. The channel matrix Expressed as

[0029]

[0030] in, is the channel path loss, β0 is the channel power gain at the reference distance d0 = 1m, d nrepresents the distance between the UAV and the IRCP, and the phase of the line-of-sight channel is Among them, f c and c represent the carrier frequency and the speed of light respectively. For convenience, we use represents the line-of-sight channel coefficient, the receiving steering vector b(γ n )for

[0031]

[0032] S132.s n (t) Using the transmit beamforming vector In N t The signal received and processed by the drone is expressed as

[0033]

[0034] in, and p represent the array gain coefficient and the transmission power respectively, represents the receive beamformer, and denote the two-step prediction and jitter effect balance angle of the (n-1)th epoch on the IRCP side, respectively. is zero and has a variance of Gaussian complex noise;

[0035] Assumptions n (t) is unit power, then the received signal-to-noise ratio at the nth moment is

[0036]

[0037] The corresponding reachable sum rates are as follows

[0038] R n =log2(1+SNR n ), (9)

[0039] S133. Assuming that the UAV flies at a constant speed at a certain altitude, the goal is to determine the changes in the UAV angle, distance, and reflection coefficient relative to the IRCP in two consecutive epochs.

[0040]

[0041] Furthermore, in S14, d n and n The relationship between the two states is established as follows:

[0042]

[0043] Among them, zθ , z d , z v and z β denote the corresponding systematic errors, which have zero mean and variances respectively. and Gaussian noise.

[0044] Further, the S15 comprises the following steps:

[0045] S151. By changing r in formula (2) n Apply a matched filter, using s with Doppler delay n (t) to estimate and τ, giving θ as n , d n , v n Measurement model

[0046]

[0047] Among them, δ θ , δ τ , The variances are The measurement noise, variance Inversely proportional to the received signal-to-noise ratio, assuming and where a i ,i=1,2,3, is a constant that depends on the system structure, G represents the gain of the matched filter;

[0048] S152. For convenience, formula (11) and formula (12) are expressed as

[0049] X n =g(X n-1 )+Z n , (13)

[0050] Y n =h(X n )+Δ n , (14)

[0051] Among them, X n =[θ n ,d n ,v n ,ζ n ] T represents the state variable, g(·) and h(·) are the functions in formula (11) and formula (12), Z n and Δ n is the covariance matrix and The noise vector is Represents a measured variable.

[0052] Further, the S2 comprises the following steps:

[0053] S21. Perform two-step state prediction according to the standard process of Kalman filtering;

[0054] S22. Targeted issues A one-dimensional optimal value search method based on EPSO and the problem The solution algorithm of

[0055] S23. EKF-EPSO-based jittered UAV beam tracking method.

[0056] Further, the S21 includes the following steps:

[0057] S211.

[0058]

[0059] in, and Represent one-step and two-step predictions respectively represents the Kalman filter value of the (n-1)th epoch;

[0060] S212. According to the state prediction value obtained by formula (15), IRCP uses Included in Perform transmit beamforming to include The drone with the information sends a DFRC signal, and the drone needs to find a The angular interval of the center is formulated as follows:

[0061]

[0062]

[0063]

[0064]

[0065] Among them, Θ is the n The maximum change in the constant.

[0066] Further, the S22 comprises the following steps:

[0067] S221. Assume that a population consists of N p Particle composition, first, give a random initial value as The solution is then followed by a maximum of K iterations. In each iteration, the maximum individual value is exchanged with the maximum overall value until the number of calculations reaches K or the change in the solution is less than the threshold κ. The particle swarm updates its speed and position in each iteration, and the updated formula is given.

[0068]

[0069] Among them, ω represents the inertia weight, which represents the confidence of the particle in its previous state of motion, and They represent the velocity and position of particle i at the kth and (k+1th) iterations respectively, c1 and c2 are two constant learning factors, r1 and r2 are two random numbers between 0 and 1, and They represent the individual optimal position and the overall optimal position of particle i in the past k iterations respectively;

[0070] S222. Use linearly decreasing inertia weight to enhance the particle swarm algorithm as follows:

[0071]

[0072] Among them, ω max and ω min Respectively represent the maximum and minimum values ​​of ω;

[0073] S223. Question The solution algorithm of .

[0074] By adopting the above technical solution, the present invention has the following advantages:

[0075] The present invention provides a perception-assisted jittered UAV beam tracking method for jittered UAV beam tracking. The UAV uses the angle information in the received DFRC signal as prior information to balance the jitter effect, thereby eliminating the need for a downlink pilot and improving the angle tracking accuracy. Compared with the traditional method that does not consider the jitter effect, the present invention has significantly improved achievable speed and rate, and ensures robust beam tracking performance in a dynamically jittered UAV scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 This is the jittered drone beam tracking scenario of the present invention. DETAILED DESCRIPTION

[0077] The technical solution of the present invention is described in detail below in conjunction with the drawings in the specification. It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0078] The present invention provides a perception-assisted jitter UAV beam tracking method, comprising the following steps:

[0079] S1. Establish a jitter UAV beam tracking model;

[0080] S1 includes the following specific steps:

[0081] S11. Establishing UAV jitter model:

[0082] S11 includes the following specific steps:

[0083] S111. The present invention considers a scenario where a single shaking drone performs beam tracking, such as Figure 1 The transmitter and receiver of the IRCP are equipped with N t and N r Antenna elements. The IRCP sends the DFRC signal to the UAV equipped with M receiving antennas. In addition, it is assumed that uniform linear arrays (ULA) are deployed on both the IRCP and the UAV. In addition, the communication between the UAV and the IRCP is carried out through a line-of-sight (LoS) channel. We denote the angle, distance, and velocity of the UAV relative to the IRCP at the nth epoch as θ n (t),d n (t) and v n (t). All parameters are functions of t∈[0,T], where T is the maximum duration of interest and T is discretized into N discrete ΔT for ease of representation. Without loss of generality, it is assumed that the UAV flies parallel to the antenna array of the IRCP.

[0084] S112. Generally speaking, the shaking of a drone is mainly caused by the pitch motion due to horizontal airflow. Therefore, the present invention mainly considers the influence of the pitch angle. In order to describe this effect, the present invention uses the following deterministic model to calculate the resulting pitch angle uncertainty.

[0085] φ n (t) = φ n (t-1)+Δφn (t), (1)

[0086] Among them, φ n (t) and φ n (t-1) represents the elevation angle of two consecutive time slots t and (t-1) respectively represents the change of pitch angle at time t, The mean is 0 and the variance is Normal distribution. Generally speaking, although the airflow is random, the wind direction will not change in the opposite direction, and the wind speed will not change significantly in a short time. Therefore, we further assume that σ φ is time-dependent, and each T σ The sampling period will change its value. In addition, σ φ The value of depends on the wind speed and is usually in the range [0.01,0.1] rad.

[0087] When the IRCP and the UAV are performing beam tracking, the initial connection between the two can be simply done through regular radar signal processing or beam training. Therefore, we mainly focus on the beam prediction and tracking stage on the IRCP side.

[0088] S12. Establishing radar model;

[0089] S12 includes the following steps:

[0090] S121. At time slot t, IRCP sends DFRC signal s to the drone n (t), for simplicity, (t) is omitted in the following. The reflected echo on the IRCP side can be expressed as

[0091]

[0092] in, p and τ represent the array gain factor, transmit power and delay respectively. is zero and has a variance of Gaussian complex noise. represents the Doppler frequency. Transmit beamforming matrix It is designed based on angle prediction, where represents the one-step angle prediction of the IRCP at the nth epoch. Finally, represents the reflection coefficient, where d n and ∈ represent the distance between the IRCP and the UAV and the radar cross section (RCS) of the UAV, respectively.

[0093] S122. Without loss of generality, the sending and receiving steering vectors of the IRCP and the UAV are as follows

[0094]

[0095]

[0096] Assuming that the antenna spacing of the ULA is half a wavelength, the transmit SNR is

[0097] S13 includes the following steps:

[0098] S131. Due to the jitter of the drone, the arrival angle AOA from IRCP to the drone is expressed as γ n =φ n +θ n , where θ n represents the angle of the UAV relative to the IRCP, and the IRCP communicates with the UAV through the line-of-sight channel. In particular, the channel matrix It can be expressed as

[0099]

[0100] in, is the channel path loss, β0 is the channel power gain at the reference distance d0 = 1m, d n represents the distance between the UAV and the IRCP. The phase of the line-of-sight channel is where f c and c represent the carrier frequency and the speed of light respectively. For convenience, we use To represent the line-of-sight channel coefficient, the receiving steering vector b(γ n )for

[0101]

[0102] S132. Similar to the radar signal expression in formula (2), s n (t) Using the transmit beamforming vector In N t The signal received and processed by the drone can be expressed as

[0103]

[0104] in, and p represent the array gain coefficient and transmit power respectively. represents the receive beamformer, and They represent the two-step prediction and jitter effect balance angle of the (n-1)th epoch on the IRCP side respectively. is zero and has a variance of Gaussian complex noise.

[0105] Assumptions n (t) is unit power, then the received signal-to-noise ratio at the nth moment is

[0106]

[0107] The corresponding reachable sum rates are as follows

[0108] R n =log2(1+SNR n ), (9)

[0109] S133. Assume that the UAV is flying at a constant speed at a certain altitude. In two consecutive epochs, the goal is to determine the changes in the UAV angle, distance, and reflection coefficient relative to the IRCP. Figure 1 The geometric relationship in

[0110]

[0111] S14. Establish a state transition model;

[0112] By approximating formula (10) and taking into account d n and n The relationship between the two states is established as follows:

[0113]

[0114] Among them, z θ , z d , z v and z β denote the corresponding systematic errors, which have zero mean and variances respectively. and Gaussian noise.

[0115] S15. Establish measurement model:

[0116] S15 includes the following steps:

[0117] S151. By changing r in formula (2) n Applying a matched filter, a Doppler-delayed n (t) to estimate and τ, giving θ as n , d n , v n Measurement model

[0118]

[0119] Among them, δ θ , δ τ , The variances are The measurement noise, according to existing literature, is Inversely proportional to the received signal-to-noise ratio. Assume and where a i , i = 1, 2, 3, is a constant that depends on the system structure, and G represents the gain of the matched filter.

[0120] S152. For convenience, formula (11) and formula (12) are expressed as

[0121] X n =g(X n-1 )+Z n , (13)

[0122] Y n =h(X n )+Δ n , (14)

[0123] Among them, X n =[θ n ,d n ,v n ,ζ n ] T represents the state variable, g(·) and h(·) are the functions in formula (11) and formula (12). n and Δ n is the covariance matrix and The noise vector. Represents a measured variable.

[0124] S2. The extended Kalman filter (EKF) and enhanced particle swarm optimization (EPSO) are combined for beam tracking of jittered UAVs.

[0125] S2 includes the following steps:

[0126] S21. Perform two-step state prediction according to the standard process of Kalman filtering;

[0127] According to the standard process of Kalman filtering, the two-step state prediction is first performed

[0128]

[0129] in, and Represent one-step and two-step predictions respectively Represents the Kalman filter value of the (n-1)th epoch.

[0130] According to the state prediction value obtained by formula (15), IRCP uses Included in Perform transmit beamforming to include The drone transmitting the DFRC signal is affected by the jitter effect. If the drone points the receiving beam directly at This is likely to cause beam deviation. To balance this effect, after obtaining prior information After that, the drone needs to find a The objective function is to maximize the received signal-to-noise ratio. In this way, the optimal receiving beamforming angle of the UAV can be obtained. Based on the above discussion, the following optimization problem is formulated.

[0131]

[0132]

[0133]

[0134]

[0135] Among them, Θ is the n The maximum change in the constant.

[0136] S22. Targeted issues A one-dimensional optimal value search method based on EPSO and the problem The solution algorithm of

[0137] Targeting the problem The present invention proposes a one-dimensional optimal value search method based on EPSO. Assume that a population consists of N p Particle composition, first, give a random initial value as The solution is then followed by a maximum of K iterations. Each iteration exchanges the maximum individual value with the maximum overall value until the number of calculations reaches K or the change in the solution is less than the threshold κ. The particle swarm updates its velocity and position in each iteration and gives the updated formula

[0138]

[0139] Among them, ω represents the inertia weight, which represents the confidence of the particle in its previous state of motion, and Represent the velocity and position of particle i at the kth and (k+1th) iterations respectively. c1 and c2 are two constant learning factors. r1 and r2 are two random numbers between 0 and 1. and They represent the individual optimal position and the overall optimal position of particle i in the past k iterations respectively.

[0140] The inertia weight ω has a great influence on the change of particle velocity. When the inertia weight is large, the global optimization ability of the algorithm becomes stronger and stronger, and it becomes smaller and smaller, which can promote the convergence of the algorithm. Therefore, a linearly decreasing inertia weight is used to enhance the particle swarm algorithm as follows

[0141]

[0142] Among them, ω max and ω min represent the maximum and minimum values ​​of ω respectively.

[0143] The problem The solution process is summarized as Algorithm 1, as shown below.

[0144]

[0145] When the drone receives When The direction with the highest signal-to-noise ratio can be obtained. This can balance the jitter effect of the drone and achieve beam alignment between the IRCP and the drone. Next, the IRCP will receive the reflected echo r from the drone. n , so that the measurement result Y can be obtained immediately n .

[0146] S23. The extended Kalman filter (EKF) and enhanced particle swarm optimization (EPSO) are combined to track the jittered UAV beam. The IRCP will continue the EKF process and optimize X n∣n-1 To get X n , then X n It will be used as the input for the next iteration to realize the whole process of jittering UAV beam prediction and tracking, and the process is summarized as Algorithm 2.

[0147]

[0148] The present invention proposes a method combining extended Kalman filter (EKF) and enhanced particle swarm optimization (EPSO) to solve the beam tracking problem of jittery UAVs. In particular, the present invention utilizes DFRC signals, thereby eliminating the need for downlink pilots and improving the angle tracking accuracy. Experimental results show that the method proposed in the present invention can accurately achieve the beam tracking task while balancing the impact of UAV jitter.

[0149] Finally, it should be pointed out that although the present invention has been described with reference to the current specific embodiments, ordinary technicians in this technical field should realize that the above embodiments are only used to illustrate the present invention and are not used as a limitation of the present invention. Various equivalent changes or substitutions can be made without departing from the concept of the present invention. Therefore, as long as the changes and modifications to the above embodiments are within the essential spirit of the present invention, they will fall within the scope of the claims of the present invention.

Claims

1. A perception-assisted jittered UAV beam tracking method, characterized in that: The following steps are involved: S1. Establish a jitter UAV beam tracking model; S2. The extended Kalman filter (EKF) and enhanced particle swarm optimization (EPSO) are combined for beam tracking of jittered UAVs. The S2 comprises the following steps: S21. Perform two-step state prediction according to the standard process of Kalman filtering; The S21 comprises the following steps: S211. in, and Represent one-step and two-step predictions respectively represents the Kalman filter value of the (n-1)th epoch; S212. According to the state prediction value obtained by the above formula, IRCP uses Included in Perform transmit beamforming to include The drone with the information sends a DFRC signal, and the drone needs to find a The angular interval of the center is formulated as follows: Among them, Θ is the n The constant of the maximum change of S22. Targeted issues A one-dimensional optimal value search method based on EPSO and the problem The solution algorithm of The S22 comprises the following steps: S221. Assume that a population consists of N p Particle composition, first, give a random initial value as The solution is then followed by a maximum of K iterations. In each iteration, the maximum individual value is exchanged with the maximum overall value until the number of calculations reaches K or the change in the solution is less than the threshold κ. The particle swarm updates its speed and position in each iteration, and the updated formula is given. Among them, ω represents the inertia weight, which represents the confidence of the particle in its previous state of motion, and They represent the velocity and position of particle i at the kth and (k+1th) iterations respectively, c1 and c2 are two constant learning factors, r1 and r2 are two random numbers between 0 and 1, and They represent the individual optimal position and the overall optimal position of particle i in the past k iterations respectively; S222. Use linearly decreasing inertia weight to enhance the particle swarm algorithm as follows: Among them, ω max and ω min Respectively represent the maximum and minimum values ​​of ω; S223. Question The solution algorithm of S23. EKF-EPSO-based jittered UAV beam tracking method.

2. The method for sensing-assisted jittered drone beam tracking according to claim 1, characterized in that: The S1 comprises the following steps: S11. Establishing UAV jitter model: S12. Establishing radar model; S13. Establish a communication model; S14. Establish a state transition model; S15. Establish a measurement model.

3. The perception-assisted jittered UAV beam tracking method according to claim 2, characterized in that: The S11 comprises the following steps: S111. Assume that uniform linear arrays (ULAs) are deployed on both the IRCP and the UAV, and the communication between the UAV and the IRCP is carried out through a line-of-sight LoS channel. The angle, distance, and speed of the UAV relative to the IRCP at the nth epoch are represented by θ n (t),d n (t) and v n (t), where all parameters are functions of t∈[0,T], where T is the maximum duration of interest. Without loss of generality, it is assumed that the UAV flies parallel to the antenna array of the IRCP; S112. Use the following deterministic model to calculate the resulting pitch angle uncertainty; f n (t)=φ n (t-1)+Δφ n (t), Among them, φ n (t) and φ n (t-1) represents the elevation angle of two consecutive time slots t and (t-1), represents the change of pitch angle at time t, The mean is 0 and the variance is The normal distribution of .

4. The perception-assisted jittered UAV beam tracking method according to claim 3, characterized in that: The S12 comprises the following steps: S121. At time slot t, IRCP sends DFRC signal s to the drone n (t), the reflected echo on the IRCP side is expressed as in, p and τ represent the array gain factor, transmit power and delay, respectively; is zero and has a variance of Gaussian complex noise, represents the Doppler frequency, the transmit beamforming matrix It is designed based on angle prediction, where represents the one-step angle prediction of IRCP at the nth epoch, represents the reflection coefficient, where d n and ∈ represent the distance between IRCP and UAV and the radar cross section RCS of UAV, respectively; S122. Without loss of generality, the sending and receiving steering vectors of the IRCP and the UAV are as follows Assuming that the antenna spacing of the ULA is half a wavelength, the transmit SNR is 5. The method for sensing-assisted jittered drone beam tracking according to claim 4, characterized in that: The S13 comprises the following steps: S131. Express the arrival angle AOA from IRCP to the UAV as γ n =φ n +θ n , where θ n Represents the angle of the drone relative to the IRCP. The IRCP communicates with the drone through a line-of-sight channel. The channel matrix Expressed as in, is the channel path loss, β0 is the channel power gain at the reference distance d0 = 1m, d n represents the distance between the UAV and the IRCP, and the phase of the line-of-sight channel is Among them, f c and c represent the carrier frequency and the speed of light respectively. For convenience, we use represents the line-of-sight channel coefficient, the receiving steering vector b(γ n )for S132.s n (t) Using the transmit beamforming vector In N t The signal received and processed by the drone is expressed as in, p and Represent the array gain coefficient, transmit power and Doppler frequency respectively, represents the receive beamformer, and denote the two-step prediction and jitter effect balance angle of the (n-1)th epoch on the IRCP side, is zero and has a variance of Gaussian complex noise; Assumptions n (t) is unit power, then the received signal-to-noise ratio at time n is The corresponding reachable sum rates are as follows R n =log2(1+SNR n ), S133. Assuming that the UAV flies at a constant speed at a certain altitude, the goal is to determine the changes in the UAV angle, distance, and reflection coefficient relative to the IRCP in two consecutive epochs.

6. The perception-assisted jittered UAV beam tracking method according to claim 5, characterized in that: S14 takes into account dx and ζ n The relationship between the two states is established as follows: Among them, z θ , z d , z v and z β denote the corresponding systematic errors, which have zero mean and variances respectively. and Gaussian noise.

7. The perception-assisted jittered UAV beam tracking method according to claim 6, characterized in that: The S15 comprises the following steps: S151. By calculating the r in the above IRCP side reflection echo formula n Apply a matched filter, using s with Doppler delay n (t) to estimate and τ, giving θ as n , d n , v n Measurement model Among them, δ θ , δ τ , The variances are The measurement noise, variance Inversely proportional to the received signal-to-noise ratio, assuming and where a i ,i=1,2,3, is a constant that depends on the system structure, G represents the gain of the matched filter; S152. For convenience of representation, the above state transfer equation and the above θ n , d n , v n The measurement model is expressed as X n =g(X n-1 )+Z n , Y n =h(X n )+Δ n , Among them, X n =[θ n ,d n ,v n ,ζ n ] T represents the state variable, g(·) and h(·) are the above state transition equation and the above θ n , d n , v n The function in the measurement model, Z n and Δ n is the noise vector, and its corresponding covariance matrix is and Represents a measured variable.

Images