Task offloading performance evaluation method of unmanned aerial vehicle assisted MEC

Abstract

The invention discloses a task unloading performance evaluation method for an unmanned aerial vehicle assisted MEC. The task unloading performance evaluation method comprises the following steps: constructing an unmanned aerial vehicle dynamic collaborative unloading model based on a Delou triangulation network structure; establishing a collaborative switching mechanism based on a Delou inner triangulation network structure through the unmanned aerial vehicle dynamic collaborative unloading model; and according to the unmanned aerial vehicle dynamic collaborative unloading model, calculating a successful calculation probability of an unloading task of the user equipment, and obtaining a successful edge calculation probability of typical user equipment based on a successful uplink communication probability and the successful calculation probability under the switching of the unmanned aerial vehicle. The technical problem of how to establish an unmanned aerial vehicle cooperative switching mechanism to switch the service cooperative nodes of the unmanned aerial vehicle in real time and perform switching performance evaluation based on the movement characteristics of the unmanned aerial vehicle is solved.

Description

Technical Field

[0001] The present invention relates to the field of mobile communication technologies, and in particular to a method for evaluating the performance of unmanned aerial vehicle (UAV)-assisted MEC task offloading. Background Art

[0002] Drones play an intuitive and important role in mobile edge computing (MEC) in air-to-ground integrated networks, bringing task processing closer to ground users. However, faced with the challenges of link stability and real-time task processing brought by aerial node mobility, it is necessary to re-evaluate and optimize the current drone-assisted MEC architecture to meet the stringent requirements of ultra-reliable and low-latency communications in the 5G era and beyond. In addition, it is very difficult to achieve edge computing capabilities and expand coverage by simply improving the performance of a single drone or expanding drone-assisted MEC. This approach is difficult to ensure the stability and reliability of the system, which seriously hinders the full utilization of the potential of large-scale drone MEC architecture. To this end, it is possible to consider improving the reliability of network services and effectively optimize complex interference problems by utilizing collaborative services between multiple drone nodes. However, in this collaborative offloading mode, the challenges of mobility management increase. Compared with traditional single-node networks, the switching mode of multi-node collaborative networks is more complex, mainly due to the irregular shape and complexity of the service area boundaries formed by multi-node collaboration, which makes accurate description challenging. When a user device crosses the boundary of the collaborative area, its service collaboration node needs to be switched in real time. Considering the random movement of drones, the collaborative switching of drones becomes more unpredictable. Therefore, it is urgent to propose a UAV-assisted MEC task offloading performance evaluation method to solve the technical problem of how to establish a UAV collaborative switching mechanism based on the mobility characteristics of the UAV to switch the service collaboration nodes of the UAV in real time and conduct switching performance evaluation. Summary of the invention

[0003] The main purpose of this invention is to propose a task offloading performance evaluation method for UAV-assisted MEC, which aims to solve the technical problem of how to establish a UAV collaborative switching mechanism based on the mobility characteristics of the UAV to switch the service collaboration nodes of the UAV in real time and perform switching performance evaluation.

[0004] To achieve the above object, the present invention provides a method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading, wherein the method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading comprises the following steps:

[0005] S1. Construct a UAV dynamic collaborative unloading model based on Delaunay triangulation network structure;

[0006] S2. Establishing a collaborative switching mechanism based on the Delaunay triangulation network structure through the UAV dynamic collaborative unloading model;

[0007] S3. According to the UAV dynamic collaborative offloading model, the successful calculation probability of the user equipment offloading task is calculated, and the successful edge computing probability of the typical user equipment is obtained based on the successful uplink communication probability and the successful computing probability under the UAV switching.

[0008] One of the preferred solutions is that step S1 constructs a UAV dynamic collaborative unloading model based on the Delaunay triangulation network structure, specifically:

[0009] All drones are subject to the density λ U Homogeneous Poisson point process, and operating at the same flight altitude h, each UAV is equipped with M antennas, which are connected to the central server through backhaul links;

[0010] All user devices are distributed according to the density λ UE A homogeneous Poisson point process is shown in FIG. 1 , where each user equipment is equipped with a single antenna, and the user equipment at the origin is a typical user equipment; the data of the typical user equipment is simultaneously unloaded to three drones in a specified unloading CoMP set, which is represented by U0={A0, A2, A3}, where U0 is the unloading CoMP set, A0, A1, and A2 are drones in the unloading CoMP set, the drones in the unloading CoMP set are serving drones, and the remaining drones are interfering drones.

[0011] One of the preferred solutions is that the step S2 establishes a collaborative switching mechanism based on the Delaunay triangulation network structure, specifically:

[0012] Defining a switching event, where the switching event is whether an average received power of a typical user equipment changes;

[0013] If the average received power of typical user equipment changes, that is, switching occurs from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set, the equivalent model of the ground-to-air wireless network is used to treat the initial cooperative switching area as a Voronoi structure with a nearest neighbor association criterion, and the Voronoi structure is quantitatively characterized by random geometry theory.

[0014] In one of the preferred solutions, the switching from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set is specifically:

[0015]

[0016] Among them, E i (t) is the average received power of the i-th offloaded CoMP set, U i(t) = {A i ,B i ,C i} is the i-th unloading CoMP set, P(t) is the transmit power of the UAV in the i-th unloading CoMP set, α is the path loss coefficient, u k (t) is the distance between the kth UAV and the projected typical user device, k∈{A i ,B i ,C i}, E j (s) is the jth offload CoMP set, U j (t) = {A j ,B j ,C j} is the jth offloaded CoMP set, is the transmit power of the jth unloaded CoMP centralized UAV, u n (s) is the distance between the nth UAV and the projected typical user device, n∈{A j ,B j ,C j}.

[0017] In one of the preferred solutions, the Voronoi structure with the nearest neighborhood association criterion is specifically:

[0018] According to the nearest neighbor association criterion, the exclusion zone is introduced. O is the coordinate origin, is the horizontal distance between the UAV and the typical user, where the interfering UAV follows a non-homogeneous point process; the density of the UAV is divided into two parts, namely, the UAV in the exclusion zone and the interfering UAV. The density of the UAV in the exclusion zone is w i is the horizontal distance between the ith UAV and the typical user; the interference UAV density is Among them, the total density of drones is λ U ,but

[0019] In one of the preferred solutions, the success calculation probability of the user equipment offloading task is:

[0020] P scp (t') = pP[T c ≤t']+(1-p)P[T m ≤t']

[0021] Among them, P scp (t') is the probability of successful calculation, p is the probability that the task is calculated by the central processor, T c , T mare the computing delays of the central processor and MEC respectively, t' is the time threshold, P[T c ≤t'] is the probability of successful calculation in the CPU for a given target calculation delay, P[T m ≤t'] is the probability of successful calculation of the given target calculation delay at the MEC end.

[0022] In one of the preferred solutions, the probability of successful calculation of the given target calculation delay in the central processor is:

[0023] P[T c ≤t']=1-exp(-u c t'+r c t')

[0024] Among them, u c is the CPU service rate, r c is the reachability rate of tasks on the central server when the proportion p of total tasks is assigned to the central server.

[0025] In one of the preferred solutions, the reachability of the task on the central server is:

[0026] r c =pλ UE |E|P sup h (t)

[0027] Among them, |E| is the entire network area, λ UE Distribution density of ground user equipment, P sup h (t) is the probability of successful uplink communication under UAV switching.

[0028] In one of the preferred solutions, the probability of successful calculation of the given target calculation delay at the MEC end is:

[0029]

[0030] Among them, u m is the service rate of the MEC server, ρ m It is the ratio of the achievable rate of MEC server tasks to the service rate.

[0031] In one of the preferred solutions, the probability of successful edge computing of the typical user equipment is:

[0032] P secp ≤P sup h (t)P scp (t')

[0033] Among them, P secp is the probability of successful edge calculation, P sup h (t) is the probability of successful uplink communication under UAV switching.

[0034] In the above technical solution of the present invention, the task offloading performance evaluation method of the drone-assisted MEC includes the following steps: constructing a drone dynamic collaborative offloading model based on the Delaunay triangulation network structure; establishing a collaborative switching mechanism based on the Delaunay triangulation network structure through the drone dynamic collaborative offloading model; calculating the probability of successful calculation of the user equipment offloading task according to the drone dynamic collaborative offloading model, and obtaining the successful edge computing probability of the typical user equipment based on the successful uplink communication probability and the successful computing probability under the drone switching. According to the mobility characteristics of the drone in the 3GPP protocol, the present invention constructs a drone dynamic collaborative offloading model based on the Delaunay triangulation network structure and an efficient collaborative switching mechanism of the Delaunay triangulation structure, and optimizes the successful edge computing probability of the typical user equipment by combining the successful uplink communication probability with the established drone dynamic collaborative offloading model, thereby solving the technical problem of how to establish a drone collaborative switching mechanism based on the mobility characteristics of the drone to switch the service collaboration node of the drone in real time and perform switching performance evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0036] Figure 1 A schematic diagram of a method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading according to an embodiment of the present invention;

[0037] Figure 2 A model of the movement of the drone according to an embodiment of the present invention;

[0038] Figure 3 Embodiment P of the present invention sup h =0.8, the probability of successful edge calculation as a function of (p, t');

[0039] Figure 4 In the embodiment of the present invention, when t'=2ms, the probability of successful edge calculation is (p, P sup h ) function’s changing trend graph.

[0040] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with the implementation methods and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] It should be noted that all directional indications (such as up, down, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0043] In addition, in the present invention, the descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" or "second" may explicitly or implicitly include at least one of the features.

[0044] Furthermore, the technical solutions between the various embodiments of the present invention may be combined with each other, but this must be based on the fact that they can be implemented by ordinary technicians in the field. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0045] See also Figure 1-Figure 4 According to one aspect of the present invention, the present invention provides a method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading, wherein the method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading comprises the following steps:

[0046] S1. Construct a UAV dynamic collaborative unloading model based on Delaunay triangulation network structure;

[0047] S2. Establishing a collaborative switching mechanism based on the Delaunay triangulation network structure through the UAV dynamic collaborative unloading model;

[0048] S3. According to the UAV dynamic collaborative offloading model, the successful calculation probability of the user equipment offloading task is calculated, and the successful edge computing probability of the typical user equipment is obtained based on the successful uplink communication probability and the successful computing probability under the UAV switching.

[0049] Specifically, in this embodiment, the step S1 constructs a UAV dynamic collaborative unloading model based on the Delaunay triangulation network structure, specifically:

[0050] The present invention supports a MEC-assisted UAV communication network, in which all UAVs are subject to a density of λU Homogeneous Poisson point process Φ U , and operate at the same flight altitude h. Each UAV is equipped with M antennas and connected to the central server through backhaul links;

[0051] All user devices are distributed according to the density λ UE Homogeneous Poisson point process Φ UE , each user equipment is equipped with a single antenna, where the user equipment at the origin is a typical user equipment; the data of the typical user equipment is simultaneously unloaded to three drones in the specified unloading CoMP set, represented as U0 = {A0, A2, A3}, U0 is the unloading CoMP set, A0, A1, A2 are drones in the unloading CoMP set, the drones in the unloading CoMP set are service drones, and the remaining drones are interference drones; each unloading CoMP set contains a circular area with a radius of r0, and the Euclidean distance from the i-th drone to the typical user equipment at O ​​is u i , where i∈{1,2,...,∞}, w i is the distance between the i-th drone and the projected typical user (0,0,h), then The offloading CoMP set is a typical user among ground users, selecting the nearest drones A0 and A1 as two related drone base stations, establishing the edge of the cooperative triangle area, and defining a quadrilateral with vertices {A0, A1, A2, A3} in two adjacent triangle areas that share the same edge; the typical user selects the two drones with the shortest distance among the four drones located at the vertices of the quadrilateral, and the drone located between the other two drones and closest to the typical user; in the drone dynamic cooperative offloading model, each drone navigates along a straight line at a random speed and altitude, and its direction obeys a uniform and independent distribution.

[0052] Specifically, in this embodiment, the step S2 establishes a collaborative switching mechanism based on the Delaunay triangulation network structure, specifically:

[0053] A switching event is defined, where the switching event is whether the average received power of a typical user equipment changes; wherein the average received power of the typical user equipment is relative to channel fading;

[0054] If the average received power of typical user equipment changes, that is, switching occurs from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set, the equivalent model of the ground-to-air wireless network is used to treat the initial cooperative switching area as a Voronoi structure with a nearest neighbor association criterion, and the Voronoi structure is quantitatively characterized by random geometry theory.

[0055] Specifically, in this embodiment, when the average received power of the typical user equipment changes, that is, switching occurs from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set, due to the movement of the drone, the drone cooperation set may not be able to form a triangular Delaunay unit at the next moment, that is, the position of the drone and the triangle list of the typical UE are both time-varying; therefore, it is possible to consider using an equivalent model of the ground-to-air wireless network, in which case the coverage area of ​​the triangular area depends only on the density of the corresponding PPP, and the challenging initial CoMP switching area can be equivalent to a classic Voronoi structure with a nearest neighbor association criterion for processing, and the resulting Voronoi cell exhibits convexity, which can be quantitatively characterized using random geometry theory; the switching from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set is specifically:

[0056]

[0057] Among them, E i (t) is the average received power of the i-th offloaded CoMP set, U i (t) = {A i ,B i ,C i} is the i-th unloading CoMP set, P(t) is the transmit power of the UAV in the i-th unloading CoMP set, α is the path loss coefficient, u k (t) is the distance between the kth UAV and the projected typical user device, k∈{A i ,B i ,C i}, E j (s) is the jth offload CoMP set, U j (t) = {A j ,B j ,C j} is the jth offloaded CoMP set, is the transmit power of the jth unloaded CoMP centralized UAV, u n (s) is the distance between the nth UAV and the projected typical user device, n∈{A j ,B j ,C j}.

[0058] Specifically, in this embodiment, the Voronoi structure with the nearest neighbor association criterion is specifically: an exclusion zone is introduced according to the nearest neighbor association criterion, and the exclusion zone is O is the coordinate origin, is the horizontal distance between the UAV and the typical user, where the interfering UAV follows a non-homogeneous point process; the density of the UAV is divided into two parts, namely, the UAV in the exclusion zone and the interfering UAV. The density of the UAV in the exclusion zone is w i is the horizontal distance between the ith UAV and the typical user; the interference UAV density is Among them, the total density of drones is λ U ,but

[0059] Specifically, in this embodiment, according to the coordinated switching mechanism, the probability of successful uplink communication under the UAV switching is calculated, and the probability of successful uplink communication under the UAV switching is:

[0060]

[0061] Among them, P sup h (t) is the probability of successful uplink communication under UAV switching, is the probability of switching failure, which reflects the system's tolerance to switching events. P[Hoff(t)] is the switching probability of the UAV cooperation area. c,ul (t) is the probability of successful uplink communication before switching.

[0062] Specifically, in this embodiment, the probability of successful uplink communication before switching is:

[0063]

[0064] Among them, w k is the horizontal distance from the kth UAV to the typical user, M is the number of antennas equipped by the UAV, α is the path loss coefficient, and u io is the link distance between the ith UAV and the typical user, γ is the signal-to-interference ratio threshold, L I (s) is the Laplace transform of the uplink interference power, To apply Palm distribution theory, we get the horizontal distance w i The joint probability density function of .

[0065] Specifically, in this embodiment, the horizontal distance w i The joint probability density function is:

[0066]

[0067] Among them, λ U The total density of drones is .

[0068] Specifically, in this embodiment, the Laplace transform of the uplink interference power is:

[0069]

[0070] Among them, λ UE is the distribution density of ground user equipment, u o is the Euclidean distance between the UAV closest to the typical user and the typical user at the origin, is the hypergeometric function, and s is the parameter variable.

[0071] Specifically, in this embodiment, the switching probability of the UAV cooperation area is the lower bound of the switching probability, and the required switching probability can be obtained by taking the equal sign; the switching probability of the UAV cooperation area is:

[0072]

[0073] in, is the horizontal distance between the UAV and the typical user, λ U The total density of drones, is the probability density function of v, In order to convert the initial collaboration area into the equivalent service drone movement speed corresponding to the Voronoi structure with the nearest neighbor association criterion, is the distance between the drone and the typical user at time t, w is the closest horizontal distance between the drone and the typical user, To disrupt the density of drones, The moving angle of the equivalent service drone corresponding to the initial collaboration area is converted into a Voronoi structure with the nearest neighbor association criterion.

[0074] Specifically, in this embodiment, the distance between the drone and the typical user at time t is:

[0075]

[0076] Specifically, in this embodiment, the density of the interfering drones is:

[0077]

[0078] Among them, λ U The total density of drones, is the cumulative distribution function of the UAV flight speed v, w i is the distance between the i-th drone and the projected typical user, is the probability density function of v, v is the flight speed of the drone, is the horizontal distance between the ith UAV and the typical user.

[0079] Specifically, in this embodiment, the moving speed of the equivalent service drone corresponding to the initial cooperation area being equivalent to the Voronoi structure with the nearest neighbor association criterion is:

[0080]

[0081] Among them, v k To offload the movement speed of each UAV in the CoMP cluster, θ k The movement angle of each UAV is centralized for offloading CoMP.

[0082] Specifically, in this embodiment, the movement angle of the equivalent service drone corresponding to the initial cooperation area being equivalent to the Voronoi structure with the nearest neighbor association criterion is:

[0083]

[0084] Specifically, in this embodiment, the probability density function is:

[0085]

[0086] Among them, Γ(a) is the gamma function of parameter a, a is a variable parameter, is the mean of the square of the moving speed of each UAV in the unloaded CoMP set, b is a variable parameter,

[0087] Specifically, in this embodiment, the success calculation probability of the user equipment uninstallation task is:

[0088] P scp (t') = pP[T c ≤t']+(1-p)P[T m ≤t']

[0089] Among them, P scp (t') is the probability of successful calculation, p is the probability that the task is calculated by the central processor, T c , T m are the computing delays of the central processor and MEC respectively, t' is the time threshold, P[T c ≤t'] is the probability of successful calculation in the CPU for a given target calculation delay, P[T m ≤t'] is the probability of successful calculation of the given target calculation delay at the MEC end;

[0090] The probability of successful calculation of the given target calculation delay in the CPU is:

[0091] P[T c ≤t']=1-exp(-u c t'+rc t')

[0092] Among them, u c is the CPU service rate, r c When the proportion p of the total tasks is assigned to the central server, the reachability rate of the tasks on the central server is:

[0093] r c =pλ UE |E|P sup h (t)

[0094] Among them, |E| is the entire network area, P sup h (t) is the probability of successful uplink communication under UAV switching;

[0095] The rate at which tasks are offloaded to the MEC server is:

[0096]

[0097] Among them, r0 is the rate at which tasks are offloaded to the MEC server;

[0098]

[0099] Among them, ρ m is the ratio of the achievable rate of the MEC server task to the service rate, r m is the task reachability rate of the MEC server, u m It is the service rate of MEC server;

[0100] The task reachability of the MEC server is:

[0101]

[0102] The probability of successful calculation of the given target calculation delay on the MEC side is:

[0103]

[0104] Specifically, in this embodiment, the probability of successful edge computing of the typical user equipment is:

[0105] P secp ≤P sup h (t)P scp (t')

[0106] Among them, P secp is the probability of successful edge calculation, P sup h (t) is the probability of successful uplink communication under UAV switching.

[0107] Specifically, in this embodiment, the task offloading performance evaluation method of the drone-assisted MEC proposed in the present invention uses the Monte Carlo method to perform simulation analysis in Matlab to obtain the simulation value of the successful edge computing probability of the model, and analyzes P sup h When and t' take different values, the changing trend of the probability of successful edge computing as a function of (p, t') is analyzed and characterized to promote the optimal design of the task offloading mechanism.

[0108] Specifically, in this embodiment, see Figure 2 , is the model of UAV movement under the UAV-assisted MEC task offloading performance evaluation method proposed in the present invention, and Q0 represents the distance from the typical user equipment projection point of the equivalent service UAV. Assume that the UAV reaches a new position Q1 at time t, The speed along Using the law of cosines, we can get the distance between Q2 and O′. Therefore, if another equivalent drone, for example, the drone at Q2, Closer to O′, switching will occur.

[0109] Specifically, in this embodiment, see Figure 3 and Figure 4 , the communication and computing performance indicators are comprehensively considered to evaluate the overall performance of the probability of successful edge computing. Based on this, the probability of successful edge computing is described as a function of the parameter p and the computing time threshold t'; Figure 3 In the paper, it is found through analysis that when the parameter p remains unchanged, the probability of successful edge computing tends to increase with the increase of the computing time threshold t', that is, under given conditions, the probability of successful edge computing is mainly constrained by the probability of successful computing, and the probability of successful computing increases with the increase of the computing time threshold; on the contrary, when the computing time t' is a constant, the probability of successful edge computing reaches a maximum value of a specific p in the time interval from 0 to 1, and the probability of successful edge computing changes with p, which is the same as the change of the probability of successful computing with respect to p; this alignment is mainly because the probability of successful edge computing is fundamentally affected by the probability of successful computing under the condition of the probability of successful uplink communication under consistent switching; at the same time, at lower p values, the probability of successful computing increases with the increase of p, while at higher p values, the probability of successful computing decreases with the increase of p;

[0110] Figure 4 The probability of successful edge computation at t' = 2ms is plotted as (p,P sup h) function, when the p value is relatively low, it is observed that the probability of successful edge computing increases with the increase of the probability of successful uplink communication. This phenomenon can be attributed to the fact that at this stage, the task processing mainly relies on the MEC server; specifically, as p increases, the arrival rate of tasks on the MEC server (denoted as rm) decreases, while the successful edge computing increases accordingly; when p is higher, a similar trend is also observed; although the successful edge computing decreases with the increase of p, the continuous growth of the probability of successful uplink communication will eventually lead to an increase in the probability of successful edge computing.

[0111] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. All equivalent structural changes made using the contents of the present invention's specification and drawings, or directly / indirectly applied in other related technical fields, are included in the patent protection scope of the present invention.

Claims

1. A method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading, characterized in that: The following steps are involved: S1. Construct a UAV dynamic collaborative unloading model based on Delaunay triangulation network structure; S2. Establishing a collaborative switching mechanism based on the Delaunay triangulation network structure through the UAV dynamic collaborative unloading model; S3. According to the UAV dynamic collaborative offloading model, the successful calculation probability of the user equipment offloading task is calculated, and the successful edge computing probability of the typical user equipment is obtained based on the successful uplink communication probability and the successful computing probability under the UAV switching.

2. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 1 is characterized in that: The step S1 constructs a UAV dynamic collaborative unloading model based on the Delaunay triangulation network structure, specifically: All drones are subject to the density λ U Homogeneous Poisson point process, and operating at the same flight altitude h, each UAV is equipped with M antennas, which are connected to the central server through backhaul links; All user devices are distributed according to the density λ UE A homogeneous Poisson point process is shown in FIG. 1 , where each user equipment is equipped with a single antenna, and the user equipment at the origin is a typical user equipment; the data of the typical user equipment is simultaneously unloaded to three drones in a specified unloading CoMP set, which is represented by U0={A0, A2, A3}, where U0 is the unloading CoMP set, A0, A1, and A2 are drones in the unloading CoMP set, the drones in the unloading CoMP set are serving drones, and the remaining drones are interfering drones.

3. A method for evaluating the performance of unmanned aerial vehicle-assisted MEC task offloading according to any one of claims 1 to 2, characterized in that: The step S2 establishes a collaborative switching mechanism based on the Delaunay triangulation network structure, specifically: Defining a switching event, where the switching event is whether the average received power of a typical user equipment changes; If the average received power of typical user equipment changes, that is, switching occurs from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set, the equivalent model of the ground-to-air wireless network is used to treat the initial cooperative switching area as a Voronoi structure with a nearest neighbor association criterion, and the Voronoi structure is quantitatively characterized by random geometry theory.

4. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 3 is characterized in that: The switching from the i-th offloading CoMP set to the adjacent j-th offloading CoMP set is specifically: Among them, E i (t) is the average received power of the i-th offloaded CoMP set, U i (t) = {A i ,B i ,C i } is the i-th unloading CoMP set, P(t) is the transmit power of the UAV in the i-th unloading CoMP set, α is the path loss coefficient, u k (t) is the distance between the kth UAV and the projected typical user device, k∈{A i ,B i ,C i }, E j (s) is the jth offload CoMP set, U j (t) = {A j ,B j ,C j } is the jth offloaded CoMP set, is the transmit power of the jth unloaded CoMP centralized UAV, u n (s) is the distance between the nth UAV and the projected typical user device, n∈{A j ,B j ,C j }.

5. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 3 is characterized in that: The Voronoi structure with the nearest neighborhood association criterion is specifically: According to the nearest neighbor association criterion, the exclusion zone is introduced. O is the coordinate origin, is the horizontal distance between the UAV and the typical user, where the interfering UAV follows a non-homogeneous point process; the density of the UAV is divided into two parts, namely, the UAV in the exclusion zone and the interfering UAV. The density of the UAV in the exclusion zone is w i is the horizontal distance between the ith UAV and the typical user; the interference UAV density is Among them, the total density of drones is λ U ,but 6. A method for evaluating the task offloading performance of UAV-assisted MEC according to any one of claims 1 to 2, characterized in that: The success calculation probability of the user equipment unloading task is: P scp (t')=pP[T c ≤t']+(1-p)P[T m ≤t'] Among them, P scp (t') is the probability of successful calculation, p is the probability that the task is executed by the central processor, T c , T m are the computing delays of the central processor and MEC respectively, t' is the time threshold, P[T c ≤t'] is the probability of successful calculation in the CPU for a given target calculation delay, P[T m ≤t'] is the probability of successful calculation of the given target calculation delay at the MEC end.

7. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 6 is characterized in that: The probability of successful calculation of the given target calculation delay in the CPU is: P[T c ≤t']=1-exp(-u c t'+r c t') Among them, u c is the CPU service rate, r c is the reachability rate of tasks on the central server when the proportion p of total tasks is assigned to the central server.

8. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 7 is characterized in that: The reachability of the task on the central server is: r c =pλ UE EP suph (t) Among them, |E| is the entire network area, λ UE Distribution density of ground user equipment, P suph (t) is the probability of successful uplink communication under UAV switching.

9. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 6 is characterized in that: The probability of successful calculation of the given target calculation delay on the MEC side is: Among them, u m is the service rate of the MEC server, ρ m It is the ratio of the achievable rate of MEC server tasks to the service rate.

10. The method for evaluating the task offloading performance of UAV-assisted MEC according to claim 6 is characterized in that: The probability of successful edge computing for the typical user device is: P secp ≤P suph (t)P scp (t') Among them, P secp is the probability of successful edge calculation, P suph (t) is the probability of successful uplink communication under UAV switching.

Images