Method for optimizing three-dimensional path based on block coordinate descent method

Abstract

The invention discloses a method for optimizing a three-dimensional path based on a block coordinate descent method, and belongs to the field of unmanned aerial vehicle path planning and time resource allocation. The implementation method comprises the following steps: constructing an unmanned aerial vehicle path planning and time resource allocation problem model, and converting an original continuous optimization problem into a discrete optimization problem by adopting a path discretization mode; equality constraints are changed into inequality constraints by adopting a relaxation method; an unmanned aerial vehicle energy consumption expression is simplified by introducing auxiliary variables, and a non-convex communication rate expression between the unmanned aerial vehicle and the ground is simplified; a block coordinate descent (BCD) method is utilized to solve the problem of simultaneous optimization of multiple types of variables, the optimization problem is divided into three sub-problems, and a continuous convex approximation (SCA) method is utilized to perform optimization iteration solution on flight time and task unloading time resource management of the unmanned aerial vehicle, a horizontal flight path of the unmanned aerial vehicle and a vertical flight path of the unmanned aerial vehicle. And realizing unmanned aerial vehicle assisted three-dimensional path optimization based on a block coordinate descent method until a convergence condition is reached.

Description

Technical Field

[0001] The invention relates to a method for optimizing a three-dimensional path based on a block coordinate descent method, and belongs to the field of unmanned aerial vehicle path planning and time resource allocation. Background Art

[0002] As mobile edge computing nodes, drones can make up for the shortcomings of static mobile edge computing networks on the ground, expand the service scope of mobile edge computing, improve the edge computing service quality of ground users, and provide more flexible edge computing services for smart terminals. In recent years, drone-assisted mobile edge computing has become a research hotspot worthy of attention. In particular, compared with static mobile edge computing networks on the ground, drone-assisted mobile edge computing has a higher degree of freedom, and the addition of drones makes resource management more complicated.

[0003] At present, most of the research on UAV flight trajectory and task offloading time resource management only designs the two-dimensional trajectory of the UAV, and models the channel between the UAV and the ground intelligent terminal as an ideal line-of-sight communication channel. However, in actual application scenarios, it is rare to encounter a completely unobstructed line-of-sight communication environment. Even in the suburbs, there are obstructions such as trees, and the communication environment in the city is even more complicated. At this time, the two-dimensional trajectory design that only considers line-of-sight transmission cannot effectively improve the task offloading capability of the UAV, and it is urgent to consider the three-dimensional UAV trajectory design under a more accurate channel model.

[0004] In addition, when managing the flight trajectory and task offloading time resources of drones, it is necessary to consider that the limited onboard energy of drones will limit their operating trajectory and task offloading capabilities. Existing studies mostly use simple linear or quadratic models to simplify the modeling of drone energy consumption, which will cause problems for the practical application of drone trajectory design. In addition, existing studies do not consider the total flight time of drones when planning drone trajectories. They usually assume that the flight time is known or fixed artificially. This assumption limits the freedom of drone trajectory design and further limits the drone's task offloading capabilities. Summary of the invention

[0005] The purpose of the present invention is to provide a method for optimizing a three-dimensional path based on a block coordinate descent method, construct a model of a UAV path planning and time resource allocation problem, and convert the original continuous optimization problem into a discrete optimization problem by using a path discretization method, so that the converted UAV path planning and time resource allocation problem model can be solved; for the non-convex nonlinear equality constraints existing in the UAV path planning and time resource allocation problem model, a relaxation method is used to convert the equality constraints into inequality constraints, and a continuous convex approximation (SCA) method is used to approximate the non-convex expression into a convex expression; the UAV energy consumption expression is simplified by introducing an auxiliary variable method, and the non-convex communication rate expression between the UAV and the ground is simplified; the block coordinate descent method (BCD) is used to solve the problem of simultaneous optimization of multiple types of variables, that is, the optimization problem is divided into three sub-problems, and the flight time and task offloading time resource management of the UAV, the horizontal flight trajectory of the UAV, and the vertical flight trajectory of the UAV are optimized respectively, and the three sub-problems are iteratively solved until the convergence condition is reached, and the UAV-assisted three-dimensional path optimization is realized based on the block coordinate descent method.

[0006] The objective of the present invention is achieved through the following technical solutions:

[0007] The present invention discloses a method for optimizing a three-dimensional path based on a block coordinate descent method, comprising the following steps:

[0008] Step 1: Construct the UAV path planning and time resource allocation problem model as shown in formula (1):

[0009]

[0010] in, and They represent the maximum horizontal speed and maximum vertical speed of the UAV, η represents the maximum energy consumption of the UAV, and Respectively represent the starting position and final position of the UAV, z min and z maxRespectively represent the minimum and maximum flight altitudes of the drone. Constraint C1.1 ensures that the energy consumption of the drone does not exceed its maximum value. Constraint C1.2 can be derived from the definition of the elevation angle. Constraint C1.3 ensures that the horizontal and vertical flight speeds of the drone do not exceed the maximum horizontal and vertical speeds, respectively. Constraints C1.4 and C1.5 ensure that the drone takes off from the specified location and lands at the specified location. Constraint C1.6 ensures that the flight altitude of the drone is within a reasonable range, avoiding collisions with buildings or towers and complying with government regulations. Constraints C1.7 and C1.8 ensure that the smart terminal can offload data to the drone in TDMA mode. Represents a collection of smart terminals.

[0011] Step 2: Convert the UAV path planning and time resource allocation problem model (1) constructed in step 1 into the continuous UAV path planning and time resource allocation optimization problem as shown in formula (2):

[0012]

[0013] Step 3: Use path discretization to convert the continuous UAV path planning and time resource allocation optimization problem into a discrete optimization problem of UAV path planning and time resource allocation as shown in formula (3):

[0014]

[0015]

[0016] Where T = {T[j]} j∈{0,…,J} , Q = {q[j]} j∈{0,…,J} , Z={z[j]} j∈{0,…,J} , In problem (3), constraints C1.13 and C1.14 not only constrain the maximum flight speed of the drone, but also constrain the maximum values ​​of the discrete segments used for path discretization in the horizontal and vertical directions; constraint C1.17 ensures that the total data transmission time of all smart terminals in each path segment will not exceed the flight time of the drone. This constraint corresponds to constraint C1.8 in the continuous time problem (2), that is, data transmission and task offloading are carried out in a TDMA manner. Since constraints C1.10, C1.11 and C1.12 are all non-convex, problem (3) is not a convex optimization problem and cannot be solved using standard convex optimization methods.

[0017] In path discretization, the 3D path of the UAV is divided into J+1 line segments, which are composed of J+2 points. characterization, and is affected by the starting position and end position If the distance between adjacent sites is limited to a certain range:

[0018]

[0019] in, and When the values ​​are appropriate, it can be assumed that the UAV flight speed and the distance between the UAV and each ground device remain constant within each line segment.

[0020] As a preferred embodiment, the total energy consumption E(T,Q,Z) of the drone in C1.11 is expressed as:

[0021]

[0022] Step 4: Use the relaxation method to transform the equation constraints in the UAV path planning and time resource allocation discrete optimization problem (3) into inequality constraints, and use the continuous convex approximation (SCA) method to approximate the non-convex expression into a convex expression.

[0023] Since the average task offloading rate function The expressions of the energy consumption function E(T,Q,Z) are extremely complex, constraint C1.12 is a nonlinear equality constraint, and problem (3) is difficult to solve.

[0024] In order to solve the discrete optimization problem of UAV path planning and time resource allocation (3), Find a lower bound:

[0025]

[0026] Since the task offloading rate under non-line-of-sight conditions is much lower than that under line-of-sight conditions, To approximate Based on this rate lower bound, the UAV path planning and time resource allocation discrete optimization problem (3) is transformed into the following approximate form:

[0027]

[0028] Problem (5) maximizes the lower bound of the minimum average task offloading data volume among all smart terminals. By solving problem (5), the task offloading performance of all smart terminals is guaranteed.

[0029] Introducing variables make

[0030]

[0031] This formula can be equivalently written as:

[0032]

[0033] By introducing the auxiliary variable Ξ={ξ[j]} j∈{0,…,J} Simplify the UAV energy consumption expression and simplify the non-convex communication rate expression between the UAV and the ground, and transform problem (5) into problem (6)

[0034]

[0035] By relaxing the nonlinear constraints C1.12 and C1.21, problem (6) is equivalently transformed into problem (7):

[0036]

[0037] Step 4: Solve the problem (7) obtained in step 3 based on the block coordinate descent method BCD (Block Coordinate Descent) and the continuous convex approximation SCA (Successive Convex Approximation) method. Use the block coordinate descent method BCD to solve the problem of simultaneous optimization of multiple types of variables in problem (7), that is, split the optimization problem into three sub-problems, optimize the flight time and task unloading time optimization problem of the drone, the horizontal flight trajectory optimization problem of the drone, and the vertical flight trajectory optimization problem of the drone, respectively, and use the continuous convex approximation SCA method to iteratively solve the three sub-problems until the convergence condition is reached. Based on the block coordinate descent method, the drone-assisted three-dimensional path optimization is realized to obtain the drone three-dimensional path optimization result. The drone three-dimensional path optimization result includes the optimized drone flight time and task unloading time, the drone horizontal flight trajectory, and the drone vertical flight trajectory.

[0038] Step 4.1: Use the block coordinate descent method (BCD) to solve the problem of optimizing multiple types of variables simultaneously in problem (7), that is, split the optimization problem into three sub-problems, namely, the optimization problem of the UAV's flight time and task offloading time, the optimization problem of the UAV's horizontal flight trajectory, and the optimization problem of the UAV's vertical flight trajectory.

[0039] Step 4.2: Use the continuous convex approximation SCA method to optimize the flight time and task offloading time optimization problems.

[0040] Given any feasible 3D UAV trajectory {Q, Z}, problem (7) degenerates into the subproblem of optimizing UAV flight and task offloading time:

[0041]

[0042] In problem (8), except for constraint C1.24, all other constraints and objective functions are about T[j], λ kConsidering that the left side of Equation C1.24 is a joint convex function of T[j] and ξ[j], the first-order Taylor expansion of the right side is used for SCA approximation, and problem (8) will be approximated as the following convex problem:

[0043]

[0044] Among them, ξ (l) [j] is the value of variable ξ[j] at the lth iteration. For problem (9), the standard convex optimization algorithm is used to optimize the flight time and task offloading time of the UAV, and the optimized flight time and task offloading time of the UAV are obtained.

[0045] Step 4.3: Use the continuous convex approximation SCA method to optimize the UAV horizontal flight trajectory optimization problem.

[0046] Given any feasible UAV flight and task offloading time {T,Λ}, and the UAV vertical flight trajectory Z, problem (7) degenerates into the sub-problem (10) of optimizing the UAV horizontal flight trajectory:

[0047]

[0048] The problem still has three non-convex constraints: C1.19, C1.23 and C1.24.

[0049] The approximate form of subproblem (10) is transformed into subproblem (11):

[0050]

[0051] The standard convex optimization algorithm is used to optimize and solve subproblem (11), thereby optimizing the horizontal flight trajectory of the UAV and obtaining the optimized horizontal flight trajectory of the UAV.

[0052] Step 4.4: Use the continuous convex approximation SCA method to optimize the vertical flight trajectory of the UAV.

[0053] Given the feasible UAV flight and task offloading time {T,Λ} and the UAV horizontal flight trajectory Q, problem (7) degenerates into subproblem (12):

[0054]

[0055] Subproblem (12) is transformed into subproblem (13):

[0056]

[0057] in

[0058]

[0059] Where: z (l) [j] is the value of z[j] at the lth iteration. In addition, the pitch angle constraint C1.23 is a convex function with respect to z[j] because its right side is concave with respect to z[j]. The standard convex optimization algorithm is used to optimize the vertical flight trajectory of the UAV and obtain the optimized vertical flight trajectory of the UAV.

[0060] Step 5: Based on the UAV three-dimensional path optimization results obtained in step 4, UAV-assisted three-dimensional path optimization is realized under the premise of minimum average task offloading data volume, thereby reducing the UAV flight drive energy consumption and communication energy consumption.

[0061] Beneficial effects:

[0062] 1. The present invention discloses a method for optimizing a three-dimensional path based on a block coordinate descent method, constructs a UAV path planning and time resource allocation problem model, and converts the original continuous optimization problem into a discrete optimization problem by using a path discretization method, so that the converted UAV path planning and time resource allocation problem model can be solved; for the non-convex nonlinear equality constraints existing in the UAV path planning and time resource allocation problem model, a relaxation method is used to convert the equality constraints into inequality constraints, and a continuous convex approximation SCA method is used to approximate the non-convex expression into a convex expression; the UAV energy consumption expression is simplified by introducing an auxiliary variable method, and the non-convex communication rate expression between the UAV and the ground is simplified; the block coordinate descent method BCD is used to solve the problem of simultaneous optimization of multiple types of variables, that is, the optimization problem is divided into three sub-problems, and the flight time and task offloading time resource management of the UAV, the horizontal flight trajectory of the UAV, and the vertical flight trajectory of the UAV are optimized respectively, and the three sub-problems are iteratively solved until the convergence condition is reached, and the UAV-assisted three-dimensional path optimization is realized based on the block coordinate descent method.

[0063] 2. The present invention discloses a method for optimizing three-dimensional paths based on block coordinate descent method, which optimizes the three-dimensional trajectory, flight time and task offloading time of a UAV with limited onboard energy in a probabilistic line-of-sight channel. According to the three-dimensional path optimization result of the UAV of the present invention, it is possible to realize UAV-assisted three-dimensional path optimization under the premise of minimum average task offloading data volume, thereby reducing the flight drive energy consumption and communication energy consumption of the UAV. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 Provide a model for the UAV path planning and time resource allocation system;

[0065] Figure 2 The relationship between the minimum average task offloading data volume and the number of iterations for each terminal;

[0066] Figure 3The relationship between the UAV trajectory and the onboard energy. DETAILED DESCRIPTION

[0067] In order to better illustrate the purpose and advantages of the present invention, the invention is further described below with reference to the accompanying drawings and examples.

[0068] Embodiment 1:

[0069] Table 1 Parameter setting table

[0070]

[0071] This embodiment discloses a method for optimizing a three-dimensional path based on a block coordinate descent method, the parameter settings are shown in Table 1, and the specific implementation steps are as follows:

[0072] Step 1: Construct the UAV path planning and time resource allocation problem model as shown in formula (1):

[0073]

[0074] in, and They represent the maximum horizontal speed and maximum vertical speed of the UAV, η represents the maximum energy consumption of the UAV, and Respectively represent the starting position and final position of the UAV, z min and z max Respectively represent the minimum and maximum flight altitudes of the drone. Constraint C1.1 ensures that the energy consumption of the drone does not exceed its maximum value. Constraint C1.2 can be derived from the definition of the elevation angle. Constraint C1.3 ensures that the horizontal and vertical flight speeds of the drone do not exceed the maximum horizontal and vertical speeds, respectively. Constraints C1.4 and C1.5 ensure that the drone takes off from the specified location and lands at the specified location. Constraint C1.6 ensures that the flight altitude of the drone is within a reasonable range, avoiding collisions with buildings or towers and complying with government regulations. Constraints C1.7 and C1.8 ensure that the smart terminal can offload data to the drone in TDMA mode. A collection of smart terminals.

[0075] Step 2: Convert the UAV path planning and time resource allocation problem model (1) constructed in step 1 into the continuous UAV path planning and time resource allocation optimization problem as shown in formula (2):

[0076]

[0077] Step 3: Use path discretization to convert the continuous UAV path planning and time resource allocation optimization problem into a discrete optimization problem of UAV path planning and time resource allocation as shown in formula (3):

[0078]

[0079]

[0080] Where T = {T[j]} j∈{0,…,J} , Q = {q[j]} j∈{0,…,J} , Z={z[j]} j∈{0,…,J} , In problem (3), constraints C1.13 and C1.14 not only constrain the maximum flight speed of the drone, but also constrain the maximum values ​​of the discrete segments used for path discretization in the horizontal and vertical directions; constraint C1.17 ensures that the total data transmission time of all smart terminals in each path segment will not exceed the flight time of the drone. This constraint corresponds to constraint C1.8 in the continuous time problem (2), that is, data transmission and task offloading are carried out in a TDMA manner. Since constraints C1.10, C1.11 and C1.12 are all non-convex, problem (3) is not a convex optimization problem and cannot be solved using standard convex optimization methods.

[0081] As a preferred embodiment, the total energy consumption E(T,Q,Z) of the drone in C1.11 is expressed as:

[0082]

[0083] Step 4: Use the relaxation method to transform the equation constraints in the UAV path planning and time resource allocation discrete optimization problem (3) into inequality constraints, and use the continuous convex approximation (SCA) method to approximate the non-convex expression into a convex expression.

[0084] Since the average task offloading rate function The expressions of the energy consumption function E(T,Q,Z) are extremely complex, constraint C1.12 is a nonlinear equality constraint, and problem (3) is difficult to solve.

[0085] In order to solve the discrete optimization problem of UAV path planning and time resource allocation (3), Find a lower bound:

[0086]

[0087] Since the task offloading rate under non-line-of-sight conditions is much lower than that under line-of-sight conditions, To approximate Based on this rate lower bound, the UAV path planning and time resource allocation discrete optimization problem (3) is transformed into the following approximate form:

[0088]

[0089] Problem (5) maximizes the lower bound of the minimum average task offloading data volume among all smart terminals. By solving problem (5), the task offloading performance of all smart terminals is guaranteed.

[0090] By introducing the auxiliary variable Ξ={ξ[j]} j∈{0,…,J} Simplify the UAV energy consumption expression and simplify the non-convex communication rate expression between the UAV and the ground, and transform problem (5) into problem (6)

[0091]

[0092] By relaxing the nonlinear constraints C4.12 and C4.21, problem (6) is equivalently transformed into problem (7):

[0093]

[0094] Step 4: Solve the problem (7) obtained in step 3 based on the block coordinate descent method BCD (Block Coordinate Descent) and the continuous convex approximation SCA (Successive Convex Approximation) method. Use the block coordinate descent method BCD to solve the problem of simultaneous optimization of multiple types of variables in problem (7), that is, split the optimization problem into three sub-problems, optimize the flight time and task unloading time optimization problem of the drone, the horizontal flight trajectory optimization problem of the drone, and the vertical flight trajectory optimization problem of the drone, respectively, and use the continuous convex approximation SCA method to iteratively solve the three sub-problems until the convergence condition is reached. Based on the block coordinate descent method, the drone-assisted three-dimensional path optimization is realized to obtain the drone three-dimensional path optimization result. The drone three-dimensional path optimization result includes the optimized drone flight time and task unloading time, the drone horizontal flight trajectory, and the drone vertical flight trajectory.

[0095] Step 4.1: Use the block coordinate descent method (BCD) to solve the problem of optimizing multiple types of variables simultaneously in problem (7), that is, split the optimization problem into three sub-problems, namely, the optimization problem of the UAV's flight time and task offloading time, the optimization problem of the UAV's horizontal flight trajectory, and the optimization problem of the UAV's vertical flight trajectory.

[0096] Step 4.2: Use the continuous convex approximation SCA method to optimize the flight time and task offloading time optimization problems.

[0097] Given any feasible 3D UAV trajectory {Q, Z}, problem (7) degenerates into the subproblem of optimizing UAV flight and task offloading time:

[0098]

[0099] In problem (8), except for constraint C1.24, all other constraints and objective functions are about T[j], λ k Considering that the left side of Equation C1.24 is a joint convex function of T[j] and ξ[j], the first-order Taylor expansion of the right side is used for SCA approximation, and problem (8) will be approximated as the following convex problem:

[0100]

[0101] Among them, ξ (l) [j] is the value of variable ξ[j] at the lth iteration. For problem (9), the standard convex optimization algorithm is used to optimize the flight time and task offloading time of the UAV, and the optimized flight time and task offloading time of the UAV are obtained.

[0102] Step 4.3: Use the continuous convex approximation SCA method to optimize the UAV horizontal flight trajectory optimization problem.

[0103] Given any feasible UAV flight and task offloading time {T,Λ}, and the UAV vertical flight trajectory Z, problem (7) degenerates into the sub-problem (10) of optimizing the UAV horizontal flight trajectory:

[0104]

[0105] The problem still has three non-convex constraints: C1.19, C1.23 and C1.24.

[0106] The approximate form of subproblem (10) is transformed into subproblem (11):

[0107]

[0108] The standard convex optimization algorithm is used to optimize and solve subproblem (11), thereby optimizing the horizontal flight trajectory of the UAV and obtaining the optimized horizontal flight trajectory of the UAV.

[0109] Step 4.4: Use the continuous convex approximation SCA method to optimize the vertical flight trajectory of the UAV.

[0110] Given the feasible UAV flight and task offloading time {T,Λ} and the UAV horizontal flight trajectory Q, problem (7) degenerates into subproblem (12):

[0111]

[0112] Subproblem (12) is transformed into subproblem (13):

[0113]

[0114] in

[0115]

[0116] Where: z (l) [j] is the value of z[j] at the lth iteration. In addition, the pitch angle constraint C4.23 is a convex function with respect to z[j] because its right side is concave with respect to z[j]. The standard convex optimization algorithm is used to optimize the vertical flight trajectory of the UAV and obtain the optimized vertical flight trajectory of the UAV.

[0117] Step 5: Based on the UAV three-dimensional path optimization results obtained in step 4, UAV-assisted three-dimensional path optimization is realized under the premise of minimum average task offloading data volume, thereby reducing the UAV flight drive energy consumption and communication energy consumption.

[0118] The specific description above further illustrates the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the above is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for optimizing a three-dimensional path based on a block coordinate descent method, characterized in that: The following steps are included: Step 1: Construct the UAV path planning and time resource allocation problem model as shown in formula (1): in, and They represent the maximum horizontal speed and maximum vertical speed of the UAV, η represents the maximum energy consumption of the UAV, and Respectively represent the starting position and final position of the UAV, z min and z max Respectively represent the minimum flight altitude and maximum flight altitude of the drone; constraint C1.1 ensures that the energy consumption of the drone does not exceed its maximum value, constraint C1.2 can be derived from the definition of the elevation angle, constraint C1.3 ensures that the horizontal flight speed and vertical flight speed of the drone do not exceed the maximum horizontal speed and maximum vertical speed respectively, constraints C1.4 and C1.5 ensure that the drone takes off from the specified location and lands at the specified location, constraint C1.6 ensures that the flight altitude of the drone is within a reasonable range, avoids collisions with buildings or towers and complies with government regulations, constraints C1.7 and C1.8 ensure that the smart terminal can unload data to the drone in TDMA mode, Represents a collection of smart terminals. Step 2: Convert the UAV path planning and time resource allocation problem model (1) constructed in step 1 into the continuous UAV path planning and time resource allocation optimization problem as shown in formula (2): stC4.1~C4.8, Step 3: Use path discretization to convert the continuous UAV path planning and time resource allocation optimization problem into a discrete optimization problem of UAV path planning and time resource allocation as shown in formula (3): C1.11: E(T,Q,Z)≤η, Where T = {T[j]} j∈{0,...,J} , Q={q[j]} j∈{0,...,J} , Z={z[j]} j∈{0,...,J} , In problem (3), constraints C1.13 and C1.14 not only constrain the maximum flight speed of the UAV, but also constrain the maximum values ​​of the discrete line segments used for path discretization in the horizontal and vertical directions; constraint C1.17 ensures that the total data transmission time of all smart terminals in each path segment will not exceed the flight time of the UAV. This constraint corresponds to constraint C1.8 in the continuous time problem (2), that is, data transmission and task offloading are carried out in a TDMA manner; since constraints C1.10, C1.11 and C1.12 are all non-convex, problem (3) is not a convex optimization problem; Step 4: Use the relaxation method to transform the equation constraints in the UAV path planning and time resource allocation discrete optimization problem (3) into inequality constraints, and use the continuous convex approximation SCA method to approximate the non-convex expression into a convex expression; In order to solve the discrete optimization problem of UAV path planning and time resource allocation (3), Find a lower bound: Since the task offloading rate under non-line-of-sight conditions is much lower than that under line-of-sight conditions, To approximate Based on this rate lower bound, the UAV path planning and time resource allocation discrete optimization problem (3) is transformed into the following approximate form: Problem (5) maximizes the lower bound of the minimum average task offloading data volume among all smart terminals. By solving problem (5), the task offloading performance of all smart terminals is guaranteed. By introducing the auxiliary variable Ξ={ξ[j]} j∈{0,...,J} Simplify the UAV energy consumption expression and simplify the non-convex communication rate expression between the UAV and the ground, and transform problem (5) into problem (6) By relaxing the nonlinear constraints C4.12 and C4.21, problem (6) is equivalently transformed into problem (7): Step 4: Solve the problem (7) obtained in step 3 based on the block coordinate descent method BCD (Block Coordinate Descent) and the continuous convex approximation SCA (Successive Convex Approximation) method. Use the block coordinate descent method BCD to solve the problem of simultaneous optimization of multiple types of variables in problem (7), that is, split the optimization problem into three sub-problems, and optimize the flight time and task offloading time optimization problem of the UAV, the horizontal flight trajectory optimization problem of the UAV, and the vertical flight trajectory optimization problem of the UAV respectively. Use the continuous convex approximation SCA method to iteratively solve the three sub-problems until the convergence condition is reached. Based on the block coordinate descent method, the UAV assisted three-dimensional path optimization is realized to obtain the UAV three-dimensional path optimization result; the UAV three-dimensional path optimization result includes the optimized UAV flight time and task offloading time, the UAV horizontal flight trajectory, and the UAV vertical flight trajectory.

2. A method for optimizing a three-dimensional path based on block coordinate descent method as claimed in claim 1, characterized in that: The total energy consumption E(T,Q,Z) of the drone in C1.11 is expressed as:

3. The method for optimizing a three-dimensional path based on block coordinate descent method according to claim 2, characterized in that: The implementation method of step 4 is: Step 4.1: Use the block coordinate descent method (BCD) to solve the problem of optimizing multiple types of variables simultaneously in problem (7), that is, split the optimization problem into three sub-problems, namely, the optimization problem of the UAV's flight time and task offloading time, the optimization problem of the UAV's horizontal flight trajectory, and the optimization problem of the UAV's vertical flight trajectory; Step 4.2: Use the continuous convex approximation SCA method to optimize the flight time and task offloading time optimization problem; Given any feasible 3D UAV trajectory {Q, Z}, problem (7) degenerates into the subproblem of optimizing UAV flight and task offloading time: In problem (8), except for constraint C1.24, all other constraints and objective functions are about T[j], λ k [j] and ξ[j]; Considering that the left side of Equation C1.24 is a joint convex function of T[j] and ξ[j], the first-order Taylor expansion of the right side is used for SCA approximation, and problem (8) will be approximated as the following convex problem: Among them, ξ (l) [j] is the value of variable ξ[j] at the lth iteration; for problem (9), the standard convex optimization algorithm is used to optimize the flight time and task offloading time of the UAV, and the optimized flight time and task offloading time of the UAV are obtained; Step 4.3: Use the continuous convex approximation SCA method to optimize the UAV horizontal flight trajectory optimization problem; Given any feasible UAV flight and task offloading time {T,Λ}, and the UAV vertical flight trajectory Z, problem (7) degenerates into the sub-problem (10) of optimizing the UAV horizontal flight trajectory: The problem still has three non-convex constraints, C1.19, C1.23, and C1.24; The approximate form of subproblem (10) is transformed into subproblem (11): The subproblem (11) is optimized by using a standard convex optimization algorithm to optimize the horizontal flight trajectory of the UAV and obtain the optimized horizontal flight trajectory of the UAV; Step 4.4: Use the continuous convex approximation SCA method to optimize the vertical flight trajectory of the UAV; Given the feasible UAV flight and task offloading time {T,Λ} and the UAV horizontal flight trajectory Q, problem (7) degenerates into subproblem (12): Subproblem (12) is transformed into subproblem (13): in Where: z (l) [j] is the value of z[j] at the lth iteration; in addition, the pitch angle constraint C1.23 is a convex function with respect to z[j] because its right end is concave with respect to z[j]; the standard convex optimization algorithm is used to optimize the vertical flight trajectory of the UAV to obtain the optimized vertical flight trajectory of the UAV.

4. A method for optimizing a three-dimensional path based on block coordinate descent method as claimed in claim 1, 2 or 3, characterized in that: The method also includes step five, which implements UAV-assisted three-dimensional path optimization under the premise of minimum average task offloading data volume based on the UAV three-dimensional path optimization result obtained in step four, thereby reducing the UAV flight drive energy consumption and communication energy consumption.

Images