Unmanned aerial vehicle resource optimization method and device based on outage probability spatial modeling

Abstract

The invention provides an unmanned aerial vehicle resource optimization method and device based on outage probability spatial modeling, and the method comprises the steps: building a sight distance probability model through employing a parameterized S-shaped curve function based on a pitch angle; constructing a large-scale fading model according to the sight distance probability model and the safety margin; establishing a small-scale fading model based on the Rice factor and the pitch angle; building an outage probability space according to the large-scale fading model and the small-scale fading model, and determining an unmanned aerial vehicle reachable area according to the outage probability space; and carrying out unmanned aerial vehicle resource optimization based on the outage probability space in an unmanned aerial vehicle reachable area. According to the method, the reliability and the resource utilization efficiency of the unmanned aerial vehicle relay communication system in a complex urban environment can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and in particular to a UAV resource optimization method and apparatus based on interruption probability space modeling. Background Technology

[0002] With the rapid development of drone technology, drones are increasingly being used in the field of communication relay. Drones, with their high flexibility and line-of-sight transmission advantages, play a vital role in emergency communications, cellular network supplementation, and IoT data collection. However, traditional drone relay communication systems face severe challenges in complex urban environments.

[0003] Patent CN120151995A proposes a power control method for UAV-assisted relay systems based on statistical knowledge. This method considers the probabilistic line-of-sight link between the UAV and ground nodes, as well as large-scale and small-scale fading effects. However, existing line-of-sight probabilistic models, large-scale fading models, and small-scale fading models have the following shortcomings: Existing line-of-sight probability models are highly complex. For example, the ITU-recommended models require detailed building parameters, depend on specific scenario configurations, and lack universality and adaptability, making it difficult to accurately predict channel conditions in different urban environments. Most studies employ simplified free-space path loss models or Rayleigh fading models, which fail to accurately reflect the actual characteristics of air-to-ground channels.

[0004] Traditional large-scale fading modeling does not fully consider the nonlinear relationship between pitch angle and shadow fading variance, and ignores the huge fluctuations in path loss in the low pitch angle region, resulting in inaccurate path loss estimation and affecting system performance evaluation.

[0005] Existing technologies typically employ Rayleigh channel models for small-scale fading modeling. However, UAV communication exhibits a significant line-of-sight component, and the Rayleigh model, which assumes no direct path, is severely inconsistent with actual channel characteristics. While some studies utilize Ricean channels, they lack accurate modeling of the relationship between the Ricean factor and spatial geometry.

[0006] The aforementioned drawbacks severely limit the reliability, coverage, and resource utilization efficiency of UAV relay communication systems in complex urban environments. Especially in high-density urban areas, the performance degradation of traditional methods is even more pronounced due to building obstruction and multipath effects. Summary of the Invention

[0007] This invention provides a method and apparatus for optimizing UAV resources based on interruption probability space modeling, which can effectively improve the reliability and resource utilization efficiency of UAV relay communication systems in complex urban environments.

[0008] A UAV resource optimization method based on interruption probability space modeling includes: Based on the pitch angle, a line-of-sight probability model is constructed using a parameterized S-curve function; A large-scale fading model is constructed based on the aforementioned line-of-sight probability model and safety margin. A small-scale fading model is established based on Rice factor and pitch angle; An interruption probability space is constructed based on the large-scale fading model and the small-scale fading model, and the reachable area of ​​the UAV is determined based on the interruption probability space. Within the reachable area of ​​the drone, drone resources are optimized based on the interruption probability space.

[0009] Furthermore, based on the pitch angle, a line-of-sight probability model is constructed using a parameterized S-curve function, including: Based on environmental factors, we obtained the offset parameter for controlling the horizontal position of the S-curve, the amplitude parameter for controlling the probability of the S-curve being in the horizontal position, and the slope parameter for controlling the steepness of the S-curve's ascent. An S-curve function is constructed based on the pitch angle, the offset parameter, the amplitude parameter, and the slope parameter to obtain the line-of-sight probability model.

[0010] Furthermore, a large-scale fading model is constructed based on the aforementioned line-of-sight probability model and safety margin, including: Calculate the free space path loss based on the communication distance between the UAV and the source node, the carrier frequency, and the speed of light; Establish a standard deviation function related to pitch angle, and calculate the first average additional loss under line-of-sight conditions and the second average additional loss under non-line-of-sight conditions based on the standard deviation function and the safety margin. Based on the line-of-sight probability model, the first average additional loss and the second average additional loss are summed by probability weighting to obtain the total average additional loss. The large-scale fading model is obtained by summing the free space path loss and the total average additional loss.

[0011] Furthermore, the standard deviation function is constructed using an exponential model based on pre-fitted relevant environmental parameters and the pitch angle.

[0012] Furthermore, a small-scale fading model is established based on the Rice factor and pitch angle, including: Construct a function relating the Rice factor and the pitch angle; Based on the relationship function, the scaling polynomial and shape adjustment polynomial in the Marcum-Q function are fitted using a polynomial approximation method to obtain the small-scale fading model. The Marcum-Q function is an exponential function of the scale adjustment polynomial and the shape adjustment polynomial.

[0013] Furthermore, an interruption probability space is constructed based on the large-scale fading model and the small-scale fading model, including: Calculate the total path loss based on the large-scale fading model. Substitute the total path loss and the Rice factor in the small-scale fading model into the Marcum-Q function to construct a monotonic interruption probability model; Based on the single-hop interruption probability model and interruption probability constraints, an interruption probability space is constructed.

[0014] Further, determining the reachable area of ​​the UAV based on the interruption probability space includes: For each point in the preset space, the interruption probability is calculated according to the single-hop interruption probability model; The region formed by the points whose interruption probability satisfies the interruption probability constraint is defined as the reachable region of the UAV.

[0015] Furthermore, UAV resource optimization is performed based on the interruption probability space, including: Establish a two-hop link signal-to-noise ratio model; Using time allocation factor, bandwidth allocation factor and power allocation factor as optimization variables, an objective function is established based on the single-hop interruption probability model. The range of values ​​for the time allocation factor, bandwidth allocation factor, and power allocation factor, as well as the range of values ​​for the UAV power constraint coefficient, are set as constraints. The objective function is solved based on the two-hop link signal-to-noise ratio model to obtain the resource optimization results of the UAV.

[0016] A UAV resource optimization device based on interruption probability space modeling, comprising: The line-of-sight modeling module uses a parameterized S-curve function to construct a line-of-sight probability model based on the pitch angle. A large-scale modeling module is used to construct a large-scale fading model based on the line-of-sight probability model and the safety margin. The small-scale modeling module establishes a small-scale fading model based on Rice factor and pitch angle. The interruption probability modeling module is used to construct an interruption probability space based on the large-scale fading model and the small-scale fading model, and to determine the reachable area of ​​the UAV based on the interruption probability space. An optimization module is used to optimize UAV resources based on the interruption probability space within the UAV's reachable area.

[0017] Furthermore, the line-of-sight modeling module constructs a line-of-sight probability model based on the pitch angle using a parameterized S-curve function, including: Based on environmental factors, we obtained the offset parameter for controlling the horizontal position of the S-curve, the amplitude parameter for controlling the probability of the S-curve being in the horizontal position, and the slope parameter for controlling the steepness of the S-curve's ascent. An S-curve function is constructed based on the pitch angle, the offset parameter, the amplitude parameter, and the slope parameter to obtain the line-of-sight probability model.

[0018] Furthermore, the large-scale modeling module constructs a large-scale fading model based on the line-of-sight probability model and the safety margin, including: Calculate the free space path loss based on the communication distance between the UAV and the source node, the carrier frequency, and the speed of light; Establish a standard deviation function related to pitch angle, and calculate the first average additional loss under line-of-sight conditions and the second average additional loss under non-line-of-sight conditions based on the standard deviation function and the safety margin. Based on the line-of-sight probability model, the first average additional loss and the second average additional loss are summed by probability weighting to obtain the total average additional loss. The large-scale fading model is obtained by summing the free space path loss and the total average additional loss.

[0019] Furthermore, the standard deviation function is constructed using an exponential model based on pre-fitted relevant environmental parameters and the pitch angle.

[0020] Furthermore, the small-scale modeling module establishes a small-scale fading model based on the Rice factor and pitch angle, including: Construct a function relating the Rice factor and the pitch angle; Based on the relationship function, the scaling polynomial and shape adjustment polynomial in the Marcum-Q function are fitted using a polynomial approximation method to obtain the small-scale fading model. The Marcum-Q function is an exponential function of the scale adjustment polynomial and the shape adjustment polynomial.

[0021] Furthermore, the interruption probability modeling module constructs an interruption probability space based on the large-scale fading model and the small-scale fading model, including: Calculate the total path loss based on the large-scale fading model. Substitute the Rice factor from the total path loss and small-scale fading model into the Marcum-Q function to construct a monotonic interruption probability model; Based on the single-hop interruption probability model and interruption probability constraints, an interruption probability space is constructed.

[0022] Furthermore, the interruption probability modeling module determines the reachable area of ​​the UAV based on the interruption probability space, including: For each point in the preset space, the interruption probability is calculated according to the single-hop interruption probability model; The region formed by the points whose interruption probability satisfies the interruption probability constraint is defined as the reachable region of the UAV.

[0023] Furthermore, the optimization module performs UAV resource optimization based on the interruption probability space, including: Establish a two-hop link signal-to-noise ratio model; Using time allocation factor, bandwidth allocation factor and power allocation factor as optimization variables, an objective function is established based on the single-hop interruption probability model. The range of values ​​for the time allocation factor, bandwidth allocation factor, and power allocation factor, as well as the range of values ​​for the UAV power constraint coefficient, are set as constraints. The objective function is solved based on the two-hop link signal-to-noise ratio model to obtain the resource optimization results of the UAV.

[0024] An electronic device includes a processor and a memory, the memory storing a plurality of instructions, the processor being configured to read the instructions and execute the method described above.

[0025] The UAV resource optimization method and apparatus based on interruption probability space modeling provided by this invention have at least the following beneficial effects: (1) A comprehensive model that better reflects the actual air-to-ground channel characteristics was established by using an environment-adaptive line-of-sight probability model, a safety margin path loss model, and a pitch-angle-dependent Ricean fading model. Simulation results show that the prediction error of the outage probability is reduced by about 40% in suburban environments and by about 35% in high-rise urban environments. (2) The proposed interruption probability space concept provides precise constraints for UAV trajectory planning, avoiding the idealized assumptions of traditional methods. In actual deployment, the system interruption probability can be reduced by about 30%, especially in complex urban environments; (3) Through the joint optimization allocation of time, bandwidth and power, the system throughput can be increased by more than 25% in the circular trajectory simulation. The resource allocation parameters can be adaptively adjusted according to environmental changes, which improves resource utilization while ensuring service quality. (4) By transforming the stochastic optimization problem into a deterministic optimization problem through the safety margin path loss model, and combining the polynomial approximation of the Marcum-Q function, the computational complexity is reduced by about 50%, thus realizing the feasibility of real-time resource scheduling. (5) The system can automatically adjust channel parameters and resource allocation strategies in different urban environments. In various environments from suburbs to high-rise cities, the system can maintain stable communication performance and significantly improve robustness.

[0026] (6) Strong engineering applicability: The present invention fully considers various constraints in actual deployment, including flight altitude restrictions, transmission power restrictions, etc., and the solution provided has good engineering implementation value. Attached Figure Description

[0027] Figure 1 This is a flowchart of one embodiment of the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0028] Figure 2 This is a schematic diagram of an embodiment of the line-of-sight probability variation with pitch angle in the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0029] Figure 3 This is a comparative schematic diagram of an embodiment of the large-scale fading model based on safety margin in the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0030] Figure 4 This is a schematic diagram of an embodiment of the UAV resource optimization method based on interruption probability space modeling provided by the present invention, in which the interruption probability varies with altitude.

[0031] Figure 5 This is a schematic diagram of an embodiment of the interruption probability space boundary in the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0032] Figure 6 This is a schematic diagram of an embodiment of the three-dimensional node interruption space in the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0033] Figure 7 This is a schematic diagram of path planning based on an embodiment of the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0034] Figure 8 This is a schematic diagram of one embodiment of the resource allocation parameter optimization in the UAV resource optimization method based on interruption probability space modeling provided by the present invention.

[0035] Figure 9 This is an optimization diagram illustrating an embodiment of the UAV resource optimization method based on interruption probability space modeling provided by the present invention, under different trajectory radii.

[0036] Figure 10 This is a schematic diagram of one embodiment of the UAV resource optimization device based on interruption probability space modeling provided by the present invention. Detailed Implementation

[0037] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0038] refer to Figure 1 In some embodiments, a UAV resource optimization method based on interruption probability space modeling is provided, including: S1. Based on the pitch angle, a line-of-sight probability model is constructed using a parameterized S-curve function. S2. Construct a large-scale fading model based on the line-of-sight probability model and safety margin; S3. Establish a small-scale fading model based on Rice factor and pitch angle; S4. Construct an interruption probability space based on the large-scale fading model and the small-scale fading model, and determine the reachable area of ​​the UAV based on the interruption probability space; S5. Within the reachable area of ​​the UAV, optimize UAV resources based on the interruption probability space.

[0039] Specifically, in step S1, based on the pitch angle, a line-of-sight probability model is constructed using a parameterized S-curve function, including: S11. Based on environmental factors, the offset parameter used to control the horizontal position of the S-curve, the amplitude parameter used to control the probability of the S-curve being in the horizontal position, and the slope parameter used to control the steepness of the S-curve's ascent are obtained through fitting. S12. Construct an S-curve function based on the pitch angle, the offset parameter, the amplitude parameter, and the slope parameter to obtain the line-of-sight probability model.

[0040] Specifically, the line-of-sight probability model is as follows: (1) Among them, P LOS θ represents the probability of a line-of-sight relationship between the UAV and the ground node, θ is the pitch angle between the UAV and the ground node, a1 is the magnitude parameter controlling the probability of the S-curve being in a horizontal position, b1 is the slope parameter controlling the steepness of the S-curve rise, and a2 is the offset parameter controlling the horizontal position of the S-curve.

[0041] a1, a2, and b1 are all parameters related to environmental factors. These three parameters together control the position, shape, and offset of the S-curve.

[0042] For the amplitude parameter a1 that controls the probability of the S-curve being in a horizontal position, the probability of θ=a2 is controlled. The larger a1 is, the greater the probability P of a line-of-sight between the UAV and the ground node. LOS The smaller the value, the greater the probability P that there is a line-of-sight between the UAV and the ground node, even if the pitch angle θ reaches a2. If a1 is large, it means that even if the pitch angle θ reaches a2, there is still a line-of-sight between the UAV and the ground node. LOSThe probability P is still very low (in densely populated cities). If a1 is close to 1, it means that when the pitch angle θ reaches a2, there is a line-of-sight between the UAV and the ground node. LOS Approximately 50%.

[0043] For the slope parameter b1 that controls the steepness of the S-shaped rise, when b1>0, the curve rises in an S-shape. The larger b1 is, the faster the transition from low probability to high probability (the occlusion effect changes drastically), and the smaller b1 is, the smoother the transition (the occlusion effect changes slowly).

[0044] The offset parameter a2, which controls the horizontal position of the S-curve, represents the horizontal offset of the probability curve in the direction of the pitch angle. In urban environments, a2 is larger (e.g., 20°~30°) because a higher angle is needed to avoid building obstruction; in open areas, a2 is smaller (e.g., 5°~15°).

[0045] Therefore, in densely populated cities: a1 is larger, a2 is larger, and b1 is smaller, requiring a higher pitch angle to obtain a higher probability of sight distance. In open rural areas, a1 is smaller, a2 is smaller, and b1 is larger, allowing for a higher probability of sight distance even with a smaller pitch angle.

[0046] In some embodiments, parameters a1, a2, and b1 can be fitted based on simulation, for example, in different environments (e.g., urban, suburban, rural), and a series of data points (θ) can be obtained from the simulation. i P LOS，i The fitting is based on nonlinear least squares or linear regression, and typical fitting values ​​are shown in Table 1: Table 1

[0047] refer to Figure 2 It provides curves showing the relationship between sight distance probability and elevation angle in different urban environments, displaying the relationship curves between sight distance probability and elevation angle in four environments: suburbs, cities, dense cities, and high-rise cities, illustrating the difference in the minimum elevation angle required to achieve the same sight distance probability in different environments.

[0048] Further, in step S2, a large-scale fading model is constructed based on the line-of-sight probability model and the safety margin, including: S21. Calculate the free space path loss based on the communication distance between the UAV and the source node, the carrier frequency, and the speed of light. S22. Establish a standard deviation function related to pitch angle, and calculate the first average additional loss under line-of-sight conditions and the second average additional loss under non-line-of-sight conditions based on the standard deviation function and the safety margin. S23. Based on the line-of-sight probability model, the first average additional loss and the second average additional loss are summed with probability weights to obtain the total average additional loss. S24. Sum the free space path loss and the total average additional loss to obtain the large-scale fading model.

[0049] Specifically, the large-scale fading model is as follows: (2) Where PL represents the total path loss, PL F P represents the free space path loss. LOS P represents the probability that there is a line-of-sight (LOS) relationship between the drone and the ground node, i.e., the LOS probability. NLOS Indicates the non-line-of-sight probability. This represents the first average additional loss under line-of-sight conditions. This represents the second average additional loss under non-line-of-sight conditions.

[0050] In some embodiments, the free space path loss is calculated using the following formula: (3) Where d represents the communication distance between the drone and the source node, f represents the carrier frequency, and c represents the speed of light.

[0051] The first average additional loss under line-of-sight conditions and the second average additional loss under non-line-of-sight conditions are calculated using the following formulas: (4) (5) Where, μ LOS This represents the fixed additional loss under line-of-sight conditions, where α is the safety margin. Let μ be the standard deviation function related to pitch angle, representing the fluctuation of loss with pitch angle under line-of-sight conditions. NLOS This represents the fixed additional loss in non-line-of-sight situations. This is the standard deviation function related to the pitch angle, representing the fluctuation of loss with pitch angle in non-line-of-sight situations.

[0052] In some embodiments, the safety margin ranges from 0.5 to 2.0, with a preferred value of 1.28.

[0053] In some embodiments, the pitch angle-related standard deviation function Based on pre-fitted relevant environmental parameters and the pitch angle, an exponential model is constructed: (6) Where k1 and k2 represent the pre-fitted relevant environmental parameters.

[0054] Standard deviation function related to pitch angle This describes how changes in the propagation environment, as the pitch angle θ changes, lead to varying degrees of loss fluctuation.

[0055] Fixed additional loss μ at line-of-sight LOS This represents the average additional attenuation caused by atmospheric absorption, slight diffraction, etc., during pure line-of-sight propagation, typically ranging from 1 to 3 dB. It can be obtained through field measurements or ray tracing.

[0056] Fixed additional loss μ in non-line-of-sight situations NLOS This represents the average additional attenuation caused by building penetration, strong reflection, diffraction, etc., during non-line-of-sight propagation; a typical value is 20-30 dB. It can be obtained through field measurements or ray tracing.

[0057] First average additional loss under line-of-sight probability model Second average additional loss in non-line-of-sight situations Perform probability-weighted summation: We obtain the total average additional loss, and then combine it with the free space path loss to obtain the total path loss, thus obtaining the large-scale fading model described above.

[0058] refer to Figure 3 The paper provides a comparison chart of large-scale fading models based on safety margin, comparing the path loss of the traditional mean model and the large-scale fading model based on safety margin proposed in this embodiment at different elevation angles, highlighting the advantages of the safety margin model in the low elevation angle region.

[0059] Further, in step S3, a small-scale fading model is established based on the Rice factor and pitch angle, including: S31. Construct the relationship function between Rice factor and pitch angle; S32. Based on the relationship function, the scaling polynomial and shape adjustment polynomial in the Marcum-Q function are fitted using a polynomial approximation method to obtain the small-scale fading model. The Marcum-Q function is an exponential function of the scale adjustment polynomial and the shape adjustment polynomial.

[0060] Specifically, in step S31, the relationship function between the Rice factor and the pitch angle is as follows: (7) Where K(θ) represents the Rice factor, and p and q represent fitting parameters related to frequency and environment.

[0061] Specifically, the Rice factor K is the ratio of direct path power to scattered power, p is the scale parameter, when θ=0, K(0)=p, which is the reference Rice factor in the horizontal direction, and q is the growth rate parameter, which controls the rate at which the Rice factor K increases with the pitch angle θ. If q>0, it means that K increases exponentially with θ.

[0062] Furthermore, in step S32, the Marcum-Q function is as follows: (8) in, Let represent the Marcum-Q function, where 'a' represents the non-centrality parameter, reflecting the relative strength of the direct path, and 'b' represents the detection threshold, reflecting the system's minimum requirement for signal strength. A larger 'b' indicates a more stringent requirement and a higher probability of interruption.

[0063] (9) (10) γ represents the average signal-to-noise ratio. th This represents the threshold signal-to-noise ratio.

[0064] Specifically, v(a) represents the scaling polynomial in the Marcum-Q function, which adjusts the rate of exponential decay and controls e. v(a) This scaling factor, as a increases, causes v(a) to generally decrease, thus slowing down the decay and reducing the probability of interruption. It captures the nonlinear effect of the direct component on the decay rate.

[0065] q(a) represents the shape adjustment polynomial in the Marcum-Q function, which controls the shape of the attenuation curve and reflects the sensitivity of the channel fading depth to the threshold.

[0066] v(a) and q(a) are usually low-order polynomials: v(a) = c0 + c1a + c2a 2 (11) q(a) = d0 + d1a + d1a 2 (12) Where c0, c1, c2, d0, d1, and d1 are the parameters to be fitted.

[0067] Further, in step S4, the interruption probability space is constructed based on the large-scale fading model and the small-scale fading model, including: S41. Calculate the total path loss based on the large-scale fading model. S42. Substitute the Rice factor in the total path loss and small-scale fading model into the Marcum-Q function to construct a single-hop interruption probability model; S43. Construct the interruption probability space based on the single-hop interruption probability model and interruption probability constraints.

[0068] Specifically, in step S41, the total path loss is calculated according to the large-scale fading model. The specific calculation formula is shown in formula (2), which will not be repeated here.

[0069] Furthermore, in step S42, the single-hop interruption probability model is as follows: (13) Among them, P out Let Q1 represent the Marcum-Q function, PL represent the total path loss, and γ represent the interruption probability. U Indicates the reference signal-to-noise ratio. This represents the interrupt threshold signal-to-noise ratio.

[0070] Specifically, the Marcum-Q function takes the form of: The derivation process of the single-hop interruption probability model is as follows: For Ricean fading channels, the received signal power gain |h| 2 The complementary cumulative distribution function (CCDF) is: (14) Interrupt event: SNR < ξ; The instantaneous SNR is: (15) therefore: (16) Substitute into Rice distribution CCDF: (17) The pitch angle θ affects the interruption probability through two paths: the Rice factor K(θ) and the total path loss PL.

[0071] Furthermore, in step S43, the interruption probability constraint is: (18) in, This represents the preset interruption probability threshold, where x, y, and z are the coordinates of a point in space.

[0072] Therefore, the interruption probability space can be represented as: (19) Among them, R 3 Representing three-dimensional space, This represents the interruption probability space.

[0073] Further, in step S4, determining the reachable area of ​​the UAV based on the interruption probability space includes: For each point in the preset space, the interruption probability is calculated according to the single-hop interruption probability model; The region formed by the points whose interruption probability satisfies the interruption probability constraint is defined as the reachable region of the UAV.

[0074] That is, the area formed by the points in the preset space that satisfy equation (19) is the area accessible to the UAV.

[0075] refer to Figure 4 The paper presents curves showing the change of interruption probability with altitude under different environments, demonstrating the trend of interruption probability with UAV altitude in four urban environments, and identifying the optimal flight altitude point in each environment.

[0076] Figure 5 The interruption probability space boundary curve diagram shows the boundary of the UAV reachable area that satisfies the interruption probability constraint on the horizontal distance-height plane, demonstrating the differences in boundary curve morphology in different environments.

[0077] Figure 6 This is a schematic diagram of the three-dimensional node interruption space, showing the three-dimensional distribution of the interruption probability space centered on the communication node. The reachable area is intuitively displayed using a teardrop-shaped three-dimensional graphic.

[0078] Figure 7 This is a schematic diagram of path planning based on the interruption probability space, showing the path planning of UAVs considering the interruption probability constraint in an environment with multiple communication nodes, and demonstrating the identification and processing strategies for overlapping and isolated areas.

[0079] Further, in step S5, resource optimization of the operational drone is performed based on the interruption probability space, including: S51. Establish a two-hop link signal-to-noise ratio model; S52. Using the time allocation factor, bandwidth allocation factor, and power allocation factor as optimization variables, establish an objective function based on the single-hop interruption probability model. S53. Set the value range of the time allocation factor, bandwidth allocation factor and power allocation factor, as well as the value range of the UAV power constraint coefficient, as constraint conditions. S54. Solve the objective function based on the two-hop link signal-to-noise ratio model to obtain the resource optimization results of the operation drone.

[0080] Specifically, in step S51, the two-hop link signal-to-noise ratio model is as follows: (20) ;(twenty one) Where, λsu λ represents the signal-to-noise ratio of the first-hop link. ud P represents the signal-to-noise ratio of the second-hop link. S P represents the transmit power of the source node. U D represents the transmit power of the drone node. su D represents the transmission distance of the first hop. ud G represents the transmission distance of the second hop. su G represents the large-scale fading gain of the first hop. ud This represents the large-scale fading gain of the second hop. This represents the small-scale fading gain of the first hop. Let B represent the small-scale fading gain of the second hop, B represent the total bandwidth, β1 represent the bandwidth allocation factor of the first hop, β2 represent the bandwidth allocation factor of the second hop, and N0 represent the noise power spectral density.

[0081] In step S52, the objective function is as follows: ;(twenty two) in, β represents the time allocation factor, and β represents the bandwidth allocation factor. P represents the power allocation factor. total This represents the total probability of system interruption.

[0082] In step S53, the constraints are as follows: ;(twenty three) ;(twenty four) in, is the power constraint coefficient of the source node. This represents the power constraint coefficient for the drone node.

[0083] In step S54, the optimization problem is solved using the gradient-based interior-point method, and a nonlinear optimizer combined with a sequential quadratic programming algorithm is used for numerical solution.

[0084] Specifically, the two-hop link signal-to-noise ratio model is used as the input to the optimization problem, and the optimization variables are mapped to the two-hop link signal-to-noise ratio model.

[0085] 1. Power Distribution: The total power is P = Ps + Pu, which is determined by the power allocation factor. Distribute: (25) (26) 2. Bandwidth allocation: β1=β, β2=(1-β).

[0086] Therefore: (20) ;(twenty one) 3. Time allocation: Although time allocation does not appear in the two-hop link signal-to-noise ratio model, it affects the effective transmitted data within time T and the minimum required signal-to-noise ratio.

[0087] refer to Figure 8 , Figure 8 The results of the resource allocation parameter optimization are shown in the figure. The variation of the time allocation factor δ and the power allocation factor η with the position and angle of the UAV under different environments is shown, and the performance difference between fixed parameters and optimized parameters is compared.

[0088] Figure 9 The comparison chart of optimization results under different trajectory radii shows the impact of trajectory radius changes on resource allocation parameters, illustrating the system's sensitivity to trajectory parameters. The method provided in this embodiment will be further explained through specific application scenarios below.

[0089] Taking a suburban environment as an example, the system parameters are configured as follows: Environmental parameters: (Sight distance probability model) Rice factor: (2.4GHz band) Communication frequency: 2 GHz UAV transmit power: 30 dBm (1W) Noise power: -81 dBm Reference signal-to-noise ratio: 110 dB Interrupt threshold: 5 dB (corresponding to normal data transmission) Safety margin factor:

[0090] Channel modeling: First, a line-of-sight probability model was established to calculate the line-of-sight probability at different elevation angles. In a suburban environment, the line-of-sight probability exceeds 0.8 when the elevation angle reaches 30°, and approaches 1.0 when the elevation angle reaches 60°.

[0091] Furthermore, a large-scale fading model was established, employing a safety margin path loss. In the low elevation angle region (<30°), the path loss was 3-5 dB higher than that of the traditional mean model, which better reflects actual measurement data.

[0092] Furthermore, a small-scale fading model was established to calculate the relationship between the Rice factor and the pitch angle. Within the pitch angle range corresponding to typical operating altitudes (100-500m), the Rice factor linearly increases from 2.5 to 8.5.

[0093] Interruption probability analysis: Based on the aforementioned channel model, the variation of single-hop outage probability with UAV altitude was calculated. The results show that the outage probability is lowest at an altitude of approximately 300m, which aligns with the legally restricted altitude for UAVs and demonstrates good engineering applicability.

[0094] An interruption probability space was constructed, and the optimal coverage angle was calculated to be 43.2°, constrained by an interruption probability threshold ε=0.1. This angle provides important guidance for UAV deployment.

[0095] Resource allocation optimization: Assume the drone flies along a circular trajectory with a radius of 100m at an altitude of 500m. The source node and the destination node are located on opposite sides of the circular trajectory, 500m from the center of the trajectory.

[0096] An optimization problem was established, incorporating a time allocation factor δ, a bandwidth allocation factor β, and a power allocation factor η. The problem was solved using MATLAB's fmincon solver, with a maximum iteration count of 1000 and a function tolerance of 1e⁻⁶.

[0097] The optimization results show that in a suburban environment, the resource allocation parameters change gradually. The time allocation factor δ varies from 0.45 to 0.55, and the power allocation factor η varies from 0.48 to 0.52, indicating that the system is not sensitive to angle changes and the resource allocation is relatively stable.

[0098] Performance evaluation: Compared with traditional methods, the solution provided in this embodiment can reduce the probability of interruption by about 35%, increase system throughput by about 28%, and reduce computation time by about 45% in suburban environments.

[0099] Taking a high-rise urban environment as an example, the system parameters are configured as follows: Environmental parameters: (Sight distance probability model) Other parameters are described in the above embodiments.

[0100] Implementation steps: During the channel modeling phase, due to the high density of tall buildings in urban areas, the line-of-sight probability increases slowly with the pitch angle. When the pitch angle is 30°, the line-of-sight probability is only 0.4; when the pitch angle reaches 60°, the line-of-sight probability is still less than 0.8.

[0101] Interruption probability analysis shows that although the environment is more complex, the optimal flight altitude still needs to be maintained at around 300m, but the maximum coverage area is significantly reduced. The optimal coverage angle is 57.5°, indicating that a higher elevation angle is needed to overcome the effects of non-line-of-sight.

[0102] The resource allocation optimization results show that parameters δ and η change more drastically with angle. The time allocation factor δ varies in the range of 0.35-0.65, and the power allocation factor η varies in the range of 0.40-0.60, indicating that the system requires more refined parameter adjustments to maintain link quality. Near 90°, the optimal parameters deviate by 0.5, shifting towards the destination node, reflecting the impact of environmental asymmetry.

[0103] Performance Evaluation: In high-rise urban environments, the solution provided in this embodiment reduces the probability of outages by approximately 40%, increases system throughput by approximately 22%, and reduces performance fluctuations at different locations and angles by approximately 30% compared to traditional methods. Multi-node resource allocation under dynamic trajectory: Considering the flight trajectory of drones in mixed environments, the implementation of the present invention in multi-node scenarios is further explained.

[0104] System Configuration: Six ground communication nodes were deployed, randomly distributed within a 500m×500m area; the node environment types included suburbs, cities, and densely populated cities; the UAV flew at an altitude of 300m, a speed of 10m / s, and a total communication time of 100s.

[0105] Step 1: Environment Identification and Parameter Configuration By using prior environmental awareness or map information, the environment type of each node can be identified, and the corresponding channel parameters can be configured.

[0106] Step 2: Constructing the Interruption Probability Space An independent outage probability space is constructed for each node. Due to different environment types, the outage probability space of each node has different shapes: suburban nodes are flat and teardrop-shaped with a large coverage area; dense urban nodes are long and thin teardrop-shaped with a small coverage area but extending vertically.

[0107] Step 3: Trajectory Planning and Resource Allocation Based on the outage probability space of each node, the drone's flight trajectory is planned. Regions with overlapping outage probability spaces are prioritized for access, allowing multiple nodes to be served simultaneously within these regions, thus improving resource utilization efficiency.

[0108] In isolated node areas, adjust the UAV flight strategy, adopting hovering or low-speed flight methods to ensure communication quality. Calculate the channel status of each node in real time and dynamically adjust resource allocation parameters.

[0109] Performance evaluation: In dynamic multi-node scenarios, the solution provided in this embodiment has an average interruption probability of less than 0.1, a total system throughput that is 35% higher than that of the fixed trajectory solution, and a resource utilization rate of over 85%, meeting the minimum service quality requirements of all nodes.

[0110] Furthermore, the method provided in this embodiment can also be modified in the following ways: Safety margin factor adjustment: The safety margin factor α can be adjusted within the range of 0.5-2.0 according to the actual reliability requirements. For scenarios with extremely high reliability requirements (such as emergency communication), a larger α value (1.5-2.0) can be selected; for cost-sensitive scenarios, a smaller α value (0.5-1.0) can be selected.

[0111] Relay protocol extension: In addition to decode-forward relay, amplified forwarding or coded cooperative relay can also be used. Only the interruption probability model needs to be adjusted accordingly; the core idea of ​​this invention remains applicable.

[0112] Optimization algorithm selection: In addition to the fmincon optimizer, intelligent optimization methods such as genetic algorithms and particle swarm optimization can also be used. Especially in high-dimensional optimization problems, intelligent algorithms may have better global search capabilities.

[0113] Multi-UAV Collaboration: Scalable to multi-UAV collaborative relay scenarios, improving system performance through distributed optimization. Each UAV shares channel state information and collaborates on trajectory planning and resource allocation.

[0114] Hybrid environment handling: For nodes located at the boundary of an environment, parameter interpolation methods can be used to smoothly transition model parameters between different environments and avoid sudden performance changes.

[0115] Real-time adaptation: By combining machine learning methods, the channel parameters are learned and adaptively adjusted online, further improving the system's performance in unknown environments.

[0116] refer to Figure 10 In some embodiments, a UAV resource optimization device based on interruption probability space modeling is proposed, comprising: The line-of-sight modeling module 201 constructs a line-of-sight probability model based on the pitch angle and using a parameterized S-curve function. Large-scale modeling module 202 is used to construct a large-scale fading model based on the line-of-sight probability model and the safety margin. Small-scale modeling module 203 establishes a small-scale fading model based on Rice factor and pitch angle; The interruption probability modeling module 204 is used to construct an interruption probability space based on the large-scale fading model and the small-scale fading model, and to determine the reachable area of ​​the UAV based on the interruption probability space. The optimization module 205 is used to optimize UAV resources based on the interruption probability space within the reachable area of ​​the UAV.

[0117] Furthermore, the line-of-sight modeling module constructs a line-of-sight probability model based on the pitch angle using a parameterized S-curve function, including: Based on environmental factors, we obtained the offset parameter for controlling the horizontal position of the S-curve, the amplitude parameter for controlling the probability of the S-curve being in the horizontal position, and the slope parameter for controlling the steepness of the S-curve's ascent. An S-curve function is constructed based on the pitch angle, the offset parameter, the amplitude parameter, and the slope parameter to obtain the line-of-sight probability model.

[0118] Furthermore, the large-scale modeling module constructs a large-scale fading model based on the line-of-sight probability model and the safety margin, including: Calculate the free space path loss based on the communication distance between the UAV and the source node, the carrier frequency, and the speed of light; Establish a standard deviation function related to pitch angle, and calculate the first average additional loss under line-of-sight conditions and the second average additional loss under non-line-of-sight conditions based on the standard deviation function and the safety margin. Based on the line-of-sight probability model, the first average additional loss and the second average additional loss are summed by probability weighting to obtain the total average additional loss. The large-scale fading model is obtained by summing the free space path loss and the total average additional loss.

[0119] Furthermore, the standard deviation function is constructed using an exponential model based on pre-fitted relevant environmental parameters and the pitch angle.

[0120] Furthermore, the small-scale modeling module establishes a small-scale fading model based on the Rice factor and pitch angle, including: Construct a function relating the Rice factor and the pitch angle; Based on the relationship function, the scaling polynomial and shape adjustment polynomial in the Marcum-Q function are fitted using a polynomial approximation method to obtain the small-scale fading model. The Marcum-Q function is an exponential function of the scale adjustment polynomial and the shape adjustment polynomial.

[0121] Furthermore, the interruption probability modeling module constructs an interruption probability space based on the large-scale fading model and the small-scale fading model, including: Calculate the total path loss based on the large-scale fading model. Substitute the total path loss and the Rice factor from the small-scale fading model into the Marcum-Q function to construct a monotonic interruption probability model. Based on the single-hop interruption probability model and interruption probability constraints, an interruption probability space is constructed.

[0122] Furthermore, the interruption probability modeling module determines the reachable area of ​​the UAV based on the interruption probability space, including: For each point in the preset space, the interruption probability is calculated according to the single-hop interruption probability model; The region formed by the points whose interruption probability satisfies the interruption probability constraint is defined as the reachable region of the UAV.

[0123] Furthermore, the optimization module performs UAV resource optimization based on the interruption probability space, including: Establish a two-hop link signal-to-noise ratio model; Using time allocation factor, bandwidth allocation factor and power allocation factor as optimization variables, an objective function is established based on the single-hop interruption probability model. The range of values ​​for the time allocation factor, bandwidth allocation factor, and power allocation factor, as well as the range of values ​​for the UAV power constraint coefficient, are set as constraints. The objective function is solved based on the two-hop link signal-to-noise ratio model to obtain the resource optimization results of the operational UAV.

[0124] In some embodiments, an electronic device is provided, including a processor and a memory, the memory storing a plurality of instructions, the processor being configured to read the instructions and execute the method described above.

[0125] The UAV resource optimization method and apparatus based on interruption probability space modeling provided in the above embodiments have at least the following beneficial effects: (1) A comprehensive model that better reflects the actual air-to-ground channel characteristics was established by using an environment-adaptive line-of-sight probability model, a safety margin path loss model, and a pitch-angle-dependent Ricean fading model. Simulation results show that the prediction error of the outage probability is reduced by about 40% in suburban environments and by about 35% in high-rise urban environments. (2) The proposed interruption probability space concept provides precise constraints for UAV trajectory planning, avoiding the idealized assumptions of traditional methods. In actual deployment, the system interruption probability can be reduced by about 30%, especially in complex urban environments; (3) Through the joint optimization allocation of time, bandwidth and power, the system throughput can be increased by more than 25% in the circular trajectory simulation. The resource allocation parameters can be adaptively adjusted according to environmental changes, which improves resource utilization while ensuring service quality. (4) By transforming the stochastic optimization problem into a deterministic optimization problem through the safety margin path loss model, and combining the polynomial approximation of the Marcum-Q function, the computational complexity is reduced by about 50%, thus realizing the feasibility of real-time resource scheduling. (5) The system can automatically adjust channel parameters and resource allocation strategies in different urban environments. In various environments from suburbs to high-rise cities, the system can maintain stable communication performance and significantly improve robustness.

[0126] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for optimizing UAV resources based on interruption probability space modeling, characterized in that, include: Based on the pitch angle, a line-of-sight probability model is constructed using a parameterized S-curve function; A large-scale fading model is constructed based on the aforementioned line-of-sight probability model and safety margin. A small-scale fading model is established based on Rice factor and pitch angle; An interruption probability space is constructed based on the large-scale fading model and the small-scale fading model, and the reachable area of ​​the UAV is determined based on the interruption probability space. Within the reachable area of ​​the drone, drone resources are optimized based on the interruption probability space.

2. The method according to claim 1, characterized in that, Based on the pitch angle, a line-of-sight probability model is constructed using a parameterized S-curve function, including: Based on environmental factors, we obtained the offset parameter for controlling the horizontal position of the S-curve, the amplitude parameter for controlling the probability of the S-curve being in the horizontal position, and the slope parameter for controlling the steepness of the S-curve's ascent. An S-curve function is constructed based on the pitch angle, the offset parameter, the amplitude parameter, and the slope parameter to obtain the line-of-sight probability model.

3. The method according to claim 1, characterized in that, Based on the aforementioned line-of-sight probability model and safety margin, a large-scale fading model is constructed, including: Calculate the free space path loss based on the communication distance between the UAV and the source node, the carrier frequency, and the speed of light; Establish a standard deviation function related to pitch angle, and calculate the first average additional loss under line-of-sight conditions and the second average additional loss under non-line-of-sight conditions based on the standard deviation function and the safety margin. Based on the line-of-sight probability model, the first average additional loss and the second average additional loss are summed by probability weighting to obtain the total average additional loss. The large-scale fading model is obtained by summing the free space path loss and the total average additional loss.

4. The method according to claim 3, characterized in that, The standard deviation function is constructed using an exponential model based on pre-fitted relevant environmental parameters and the pitch angle.

5. The method according to claim 1, characterized in that, A small-scale fading model is established based on Rice factor and pitch angle, including: Construct a function relating the Rice factor and the pitch angle; Based on the relationship function, the scaling polynomial and shape adjustment polynomial in the Marcum-Q function are fitted using a polynomial approximation method to obtain the small-scale fading model. The Marcum-Q function is an exponential function of the scale adjustment polynomial and the shape adjustment polynomial.

6. The method according to claim 5, characterized in that, Based on the large-scale fading model and the small-scale fading model, an interruption probability space is constructed, including: Calculate the total path loss based on the large-scale fading model. Substitute the Rice factor from the total path loss and small-scale fading model into the Marcum-Q function to construct a monotonic interruption probability model; Based on the single-hop interruption probability model and interruption probability constraints, an interruption probability space is constructed.

7. The method according to claim 6, characterized in that, The reachable area of ​​the UAV is determined based on the interruption probability space, including: For each point in the preset space, the interruption probability is calculated according to the single-hop interruption probability model; The region formed by the points whose interruption probability satisfies the interruption probability constraint is defined as the reachable region of the UAV.

8. The method according to claim 6, characterized in that, Optimizing UAV resources based on the interruption probability space includes: Establish a two-hop link signal-to-noise ratio model; Using time allocation factor, bandwidth allocation factor and power allocation factor as optimization variables, an objective function is established based on the single-hop interruption probability model. The range of values ​​for the time allocation factor, bandwidth allocation factor, and power allocation factor, as well as the range of values ​​for the UAV power constraint coefficient, are set as constraints. The objective function is solved based on the two-hop link signal-to-noise ratio model to obtain the resource optimization results of the UAV.

9. A UAV resource optimization device based on interruption probability space modeling, characterized in that, include: The line-of-sight modeling module uses a parameterized S-curve function to construct a line-of-sight probability model based on the pitch angle. A large-scale modeling module is used to construct a large-scale fading model based on the line-of-sight probability model and the safety margin. The small-scale modeling module establishes a small-scale fading model based on Rice factor and pitch angle. The interruption probability modeling module is used to construct an interruption probability space based on the large-scale fading model and the small-scale fading model, and to determine the reachable area of ​​the UAV based on the interruption probability space. An optimization module is used to optimize UAV resources based on the interruption probability space within the UAV's reachable area.

10. An electronic device, characterized in that, It includes a processor and a memory, the memory storing multiple instructions, and the processor being used to read the instructions and execute the method as described in any one of claims 1-8.

Images