A method and device for designing a UAV route network guided by traffic flow demand

Abstract

The present invention provides a method and device for designing a UAV route network under the guidance of traffic flow demand. The method comprises: selecting the site of the nodes of the UAV route network and initializing the topology of the UAV route network; constructing an optimization model of the UAV route network based on the initialized UAV route network topology with the goal of minimizing the risk of traffic flow operation; solving the optimization model of the UAV route network by genetic algorithm to obtain a design scheme of the UAV route network that minimizes the risk of traffic flow operation. The method and device of the present invention construct a reasonable and intensive UAV route network topology structure by organically matching the distribution of UAV traffic flow demand with the route structure layout configuration, so as to maximize the safety level of large-scale operation of UAVs and the efficiency of airspace resource utilization.

Description

Technical Field

[0001] The present invention relates to the field of air traffic management and unmanned aerial vehicle (UAV) technology, and in particular to a method and device for designing an unmanned aerial vehicle (UAV) route network under the guidance of traffic flow demand. Background Art

[0002] As an important physical space for UAVs to operate in the air, the rationality of route design will restrict the safety and efficiency of flight activities. Therefore, it is necessary to scientifically plan low-altitude UAV routes to meet the needs of large-scale commercial application development of UAVs in the future.

[0003] Traditional route planning focuses on the path planning level of a single UAV, with the purpose of searching for the optimal trajectory from the starting point to the target point in the airspace environment in combination with the UAV's target mission. Commonly used methods include graph search algorithms such as Dijkstra algorithm, Prim algorithm, Kruskal algorithm, and A* algorithm. However, route planning from the perspective of a single UAV is mostly task-oriented and one-time, and its airspace effectiveness will end after the UAV mission is completed. In addition, the planned route of a single UAV is only applicable to the safety control of that specific UAV. With the rapid expansion of the number and application scope of UAVs in low-altitude airspace, the route planning workload and coordination difficulty between a large number of UAVs operating in the same area will be greatly increased.

[0004] At present, the existing UAV route planning methods focus on the path planning level of a single UAV rather than the traffic flow level. In fact, the number and application scope of UAVs in low-altitude airspace will expand rapidly in the future, and the individual route planning work of a large number of UAVs will overlap.

[0005] Therefore, it is necessary to uniformly manage and dispatch drones in the airspace from the traffic flow level rather than the single flight level, and establish a public route network with strong globality and versatility, so as to improve the effectiveness of the use of low-altitude airspace resources and regulate the orderliness of drone flight activities. Summary of the invention

[0006] The embodiments of the present invention provide a method and device for designing a UAV route network under the guidance of traffic flow demand, so as to maximize the safety level of UAV traffic flow operation.

[0007] In order to achieve the above object, the present invention adopts the following technical scheme.

[0008] According to one aspect of the present invention, a method for designing a UAV route network under the guidance of traffic flow demand is provided, comprising:

[0009] Select the nodes of the UAV route network and initialize the topology of the UAV route network;

[0010] Based on the initialized UAV route network topology, an optimization model of the UAV route network is constructed with the goal of minimizing the risk of traffic flow operation;

[0011] The optimization model of the UAV route network is solved by a genetic algorithm to obtain a design scheme of the UAV route network that minimizes the risk of traffic flow operation.

[0012] Preferably, the site selection of nodes of the drone route network includes:

[0013] The critical ranking method is used to select route nodes suitable for UAV take-off, landing and transfer from multiple alternative route convergence points, and the original evaluation matrix X of m alternative route convergence points under n site selection factors is established. ij =(x ij ) m×n , x ij is the value of the jth alternative route convergence point under the location element i;

[0014] For the original evaluation matrix X ij Dimensionless processing is performed to obtain a standard evaluation matrix that eliminates the order of magnitude differences between different indicators.

[0015] Calculate the weight value w corresponding to the location factor i i ,in

[0016]

[0017] Where E i is the entropy value of location factor i;

[0018]

[0019] Determine the best solution for each location factor and the worst solution

[0020]

[0021] Calculate the criticality score R of each alternative route convergence point j , sort and get the first Φ route convergence points, and use the first Φ route convergence points as the nodes of the UAV route network;

[0022]

[0023] Preferably, the initialization of the drone route network topology includes:

[0024] The maximum range of the drone The adjacency matrix A of the route convergence point is established based on (a mn )Φ×Φ , if the length of the segment between the mth route convergence point and the nth route convergence point ω is the margin coefficient, then a mn =1, otherwise a mn = 0, the geographical environment in the airspace is segmented according to the preset grid granularity, and the variable-length A* path search algorithm is used to generate the optimal route between two connected route convergence points, including the following processing steps:

[0025] Step 1: Set a buffer zone with the straight line between the mth route convergence point and the nth convergence point as the center to form an initialized search space;

[0026] Step 2: Define the estimated cost function f(x) of grid point x, including the actual voyage cost g(x) and the estimated voyage cost h(x):

[0027]

[0028] g(x) is the actual distance from the starting convergence point m to the grid point x to be expanded, h(x) is the estimated distance from the grid point x to be expanded to the ending convergence point n, (X x ,Y x ) are grid point coordinates;

[0029] Step 3: Create feasible_set and finish_set lists, take the starting convergence point m as the parent node, calculate the estimated cost function value of each free grid point around the starting convergence point m, and add the estimated cost function value to the feasible_set list. Determine the search step length according to the proportion of obstacle grids in the local search space. When the proportion of obstacle grids is less than 30%, the step length uses 2 grid units, otherwise it uses 1 grid unit.

[0030] Step 4: Select the grid point P with the smallest f(x) in the feasible_set list as the next extended grid point, and add the grid point P to the finish_set list to determine whether the grid point P is the final convergence point n. If so, go to Step 7, otherwise go to Step 5.

[0031] Step 5: Take the grid point P as the new parent node and search for adjacent free grid points according to the search step length corresponding to the obstacle grid ratio around the grid point P;

[0032] Step 6: Recalculate the estimated cost function f(x) and return to Step 4.

[0033] Step 7: Store the optimal path obtained by the search algorithm, which is a free grid point sequence that minimizes the path cost between the mth route convergence point and the nth route convergence point.

[0034] Preferably, the optimization model of the UAV route network is constructed based on the initialized UAV route network topology with the goal of minimizing the traffic flow operation risk, including:

[0035] The traffic flow combination distribution K (K = Φ(Φ-1) / 2) of the starting and ending route convergence nodes with UAV flight demand is statistically analyzed. With the goal of minimizing the traffic flow operation risk, the optimization model of the UAV route network N(V, E) is established as follows:

[0036]

[0037] Where N is the route network topology to be optimized, including Φ route aggregation nodes, and E is the set of all routes in N. mn |m,n∈V and a mn =1}, f mn , l mn and r mn Route e mn The route flow, route mileage and route congestion level on the route are defined as:

[0038]

[0039] f k is the flow rate between the kth traffic flow OD pair, R k is the number of optional paths for traffic flow k, The probability of selecting path i for traffic flow k, is a 0-1 variable. When the selected path i of traffic flow k contains e mn hour Otherwise it is 0, c mn is the route capacity;

[0040] The optimization model satisfies the following constraints:

[0041]

[0042] f mn ≥f o (10)

[0043] Constraint (9) ensures the conservation of the UAV path flow and the OD pair flight flow, that is, the sum of the flow on the feasible section of the OD pair is equal to the traffic demand between the OD pair. Constraint (10) is the route utilization constraint, which requires that the route allocation flow must not be less than the route activation threshold.

[0044] Preferably, solving the optimization model of the drone route network by a genetic algorithm to obtain a design scheme of the drone route network that minimizes the risk of traffic flow operation includes:

[0045] The binary code of the route network topology is set as the upper layer code, and the real number code of the traffic flow distribution is set as the lower layer code. The optimization model of the UAV route network is solved by using a genetic algorithm with double-layer coding, including the following processing steps:

[0046] Step 1: Randomly generate an initial population p(0) = {X1, X2, ...X N}, N is the population size, X i is the i-th chromosome, which consists of two layers. Considering that the operation risk of traffic flow k on the route is dynamically updated with the allocated flow 1, 2, ... k-1, let the path selection probability The flight path i with the lowest risk in the current operation, i=1,2,...R k Take 1 at the position and 0 at the rest, that is:

[0047]

[0048] Step 2: Express the fitness function as the inverse of the objective function, that is, fitness = 1 / R(N). The smaller the risk value, the higher the fitness and the greater the probability of being selected. The chromosomes in the population are linearly sorted from small to large according to the fitness size, and the roulette wheel selection strategy is used to select proportionally according to the fitness value.

[0049] Step 3: Perform a crossover operation on the upper-layer route network topology code, that is, select two chromosomes corresponding to the crossover probability to exchange the network topology code after the intersection of the two chromosomes. If an illegal chromosome is generated after the crossover, the chromosome is repaired and the lower-layer traffic flow distribution code is adjusted accordingly.

[0050] Step 4: Perform mutation operation on the upper-layer route network topology coding, select a chromosome according to the mutation probability and randomly select a gene position to assign a new value. If an illegal chromosome is generated, the chromosome is repaired by using a repair method;

[0051] Step 5: Repeatedly generate new populations through the survival of the fittest until the maximum number of iterations is exceeded, and the optimal chromosome decoding with the lowest overall operation risk is taken as the final result to obtain the design plan of the UAV route network that minimizes the traffic flow operation risk.

[0052] According to another aspect of the present invention, there is provided a UAV route network design device guided by traffic flow demand, comprising: a waypoint site selection module, a network topology initialization module, a network optimization target modeling module and a route network iterative update module;

[0053] The waypoint site selection module is used to select the nodes of the UAV route network;

[0054] The network topology initialization module is used to initialize the network topology of the UAV route;

[0055] The network optimization target modeling module is used to construct an optimization model of the UAV route network based on the initialized UAV route network topology with the goal of minimizing the traffic flow operation risk;

[0056] The route network iterative update module is used to solve the optimization model of the UAV route network through a genetic algorithm to obtain a design plan for the UAV route network that minimizes the risk of traffic flow operation.

[0057] Preferably, the waypoint location selection module is used to select route nodes suitable for drone take-off, landing and transfer from multiple candidate route convergence points using a critical sorting method, and establish an original evaluation matrix X of m candidate route convergence points under n location selection factors ij =(x ij ) m×n , x ij is the value of the jth alternative route convergence point under the location element i;

[0058] For the original evaluation matrix X ij After dimensionless processing, we can get the standard evaluation matrix Y that eliminates the order of magnitude differences between different indicators. ij =(y ij ) m×n ,

[0059] Calculate the weight value w corresponding to the location factor i i ,in

[0060]

[0061] Where E i is the entropy value of location factor i;

[0062]

[0063] Determine the best solution for each location element and the worst solution

[0064]

[0065] Calculate the criticality score R of each alternative route convergence point j , sort and get the first Φ route convergence points, and use the first Φ route convergence points as the nodes of the UAV route network;

[0066]

[0067] Preferably, the network topology initialization module is used to initialize the network topology based on the maximum range of the drone. The adjacency matrix A of the route convergence point is established based on (a mn ) Φ×Φ , if the length of the segment between the mth route convergence point and the nth route convergence point ω is the margin coefficient, then a mn =1, otherwise a mn = 0, the geographical environment in the airspace is segmented according to the preset grid granularity, and the variable-length A* path search algorithm is used to generate the optimal route between two connected route convergence points, including the following processing steps:

[0068] Step 1: Set a buffer zone with the straight line between the mth route convergence point and the nth convergence point as the center to form an initialized search space;

[0069] Step 2: Define the estimated cost function f(x) of grid point x, including the actual voyage cost g(x) and the estimated voyage cost h(x):

[0070]

[0071] g(x) is the actual distance from the starting convergence point m to the grid point x to be expanded, h(x) is the estimated distance from the grid point x to be expanded to the ending convergence point n, (X x ,Y x ) are grid point coordinates;

[0072] Step 3: Create feasible_set and finish_set lists, take the starting convergence point m as the parent node, calculate the estimated cost function value of each free grid point around the starting convergence point m, and add the estimated cost function value to the feasible_set list. Determine the search step length according to the proportion of obstacle grids in the local search space. When the proportion of obstacle grids is less than 30%, the step length uses 2 grid units, otherwise it uses 1 grid unit.

[0073] Step 4: Select the grid point P with the smallest f(x) in the feasible_set list as the next extended grid point, and add the grid point P to the finish_set list to determine whether the grid point P is the final convergence point n. If so, go to Step 7, otherwise go to Step 5.

[0074] Step 5: Take the grid point P as the new parent node and search for adjacent free grid points according to the search step length corresponding to the obstacle grid ratio around the grid point P;

[0075] Step 6: Recalculate the estimated cost function f(x) and return to Step 4.

[0076] Step 7: Store the optimal path obtained by the search algorithm, which is a free grid point sequence that minimizes the path cost between the mth route convergence point and the nth route convergence point.

[0077] Preferably, the network optimization target modeling module is used to calculate the traffic flow combination distribution K (K = Φ (Φ-1) / 2) of the starting and ending route convergence nodes where there is a drone flight demand, and to minimize the traffic flow operation risk, and to establish the optimization model of the drone route network N (V, E) as follows:

[0078]

[0079] Where N is the route network topology to be optimized, including Φ route aggregation nodes, and E is the set of all routes in N. mn |m,n∈V and a mn =1}, f mn , l mn and r mn Route e mn The route flow, route mileage and route congestion level on the route are defined as:

[0080]

[0081] f k is the flow rate between the kth traffic flow OD pair, R k is the number of optional paths for traffic flow k, The probability of selecting path i for traffic flow k, is a 0-1 variable. When the selected path i of traffic flow k contains e mn hour Otherwise it is 0, c mn is the route capacity;

[0082] The optimization model satisfies the following constraints:

[0083]

[0084] f mn ≥f o (10)

[0085] Constraint (9) ensures the conservation of the UAV path flow and the OD pair flight flow, that is, the sum of the flow on the feasible section of the OD pair is equal to the traffic demand between the OD pair. Constraint (10) is the route utilization constraint, which requires that the route allocation flow must not be less than the route activation threshold.

[0086] Preferably, the route network iterative update module is used to set the binary code of the route network topology as the upper layer code, the real number code of the traffic flow distribution as the lower layer code, and adopt the genetic algorithm of the double-layer code to solve the optimization model of the UAV route network, including the following processing steps:

[0087] Step 1: Randomly generate an initial population p(0) = {X1, X2, ...X N}, N is the population size, X i is the i-th chromosome, which consists of two layers. Considering that the operation risk of traffic flow k on the route is dynamically updated with the allocated flow 1, 2, ... k-1, let the path selection probability The flight path i with the lowest risk in the current operation, i=1,2,...R k Take 1 at the position and 0 at the rest, that is:

[0088]

[0089] Step 2: Express the fitness function as the inverse of the objective function, that is, fitness = 1 / R(N). The smaller the risk value, the higher the fitness and the greater the probability of being selected. The chromosomes in the population are linearly sorted from small to large according to the fitness size, and the roulette wheel selection strategy is used to select proportionally according to the fitness value.

[0090] Step 3: Perform a crossover operation on the upper-layer route network topology code, that is, select two chromosomes corresponding to the crossover probability to exchange the network topology code after the intersection of the two chromosomes. If an illegal chromosome is generated after the crossover, the chromosome is repaired and the lower-layer traffic flow distribution code is adjusted accordingly.

[0091] Step 4: Perform mutation operation on the upper-layer route network topology coding, select a chromosome according to the mutation probability and randomly select a gene position to assign a new value. If an illegal chromosome is generated, the chromosome is repaired by using a repair method;

[0092] Step 5: Repeatedly generate new populations through the survival of the fittest until the maximum number of iterations is exceeded, and the optimal chromosome decoding with the lowest overall operation risk is taken as the final result to obtain the design plan of the UAV route network that minimizes the traffic flow operation risk.

[0093] It can be seen from the technical solutions provided by the above-mentioned embodiments of the present invention that the present invention proposes a method and device for designing a UAV route network under the guidance of traffic flow demand. By organically matching the distribution of UAV traffic flow demand and the route structure layout configuration, a reasonable and intensive UAV route network topology structure is constructed to maximize the safety level of UAV traffic flow operation.

[0094] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0096] Figure 1 A processing flow chart of a method for designing a UAV route network under the guidance of traffic flow demand provided by an embodiment of the present invention;

[0097] Figure 2 A processing diagram of a method for scientifically selecting a location for a candidate route convergence point provided by an embodiment of the present invention;

[0098] Figure 3 A route search flow chart provided by an embodiment of the present invention;

[0099] Figure 4 A solution flow chart of a genetic algorithm using double-layer encoding provided in an embodiment of the present invention.

[0100] Figure 5 A schematic diagram of a vector form of a first-layer candidate network topology adjacency matrix provided by an embodiment of the present invention;

[0101] Figure 6 A schematic diagram of the flow probability distribution results of k traffic flow OD pairs corresponding to a second-layer network topology provided by an embodiment of the present invention;

[0102] Figure 7 A structural diagram of a UAV route network design device guided by traffic flow demand provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0103] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0104] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items.

[0105] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0106] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings, and each embodiment does not constitute a limitation on the embodiments of the present invention.

[0107] Embodiment 1

[0108] This embodiment provides a method for designing a UAV route network under the guidance of traffic flow demand. First, the criticality sorting is used to scientifically select the location of the UAV route convergence point, and then the variable-length A* path search algorithm is used to generate the initial topology of the route network under the maximum range constraint of the UAV. Further, a UAV route network optimization model is established with the goal of minimizing the risk of traffic flow operation, and finally a double-layer coded genetic algorithm is used to iteratively update and solve the UAV route network. In this way, the organic matching of the distribution of UAV traffic flow demand and the route structure layout configuration can be achieved to the maximum extent, the effectiveness of the use of low-altitude airspace resources can be improved, and the safety level of UAV traffic flow operation can be guaranteed.

[0109] The UAV route network design under the guidance of traffic flow demand in the present invention refers to the design of the UAV route network that minimizes the risk of traffic flow operation by reasonably selecting the waypoints and waypoint connection relationships of the route network based on the future traffic flow demand distribution of the UAV, under the premise of meeting the constraints such as the UAV range performance limit, airspace obstacle limit and route utilization requirements. UAV route network planning is an important guarantee for improving the effectiveness of low-altitude airspace resource utilization and regulating the orderliness of UAV flight activities. Under the guidance of given traffic flow demand, the UAV route network aims to minimize the risk of traffic flow operation. By continuously adjusting the topological structure, it can achieve an organic match between the UAV traffic flow distribution and the route network structure, ensuring the effectiveness of low-altitude airspace resource utilization and the safety of UAV flight activities.

[0110] The processing flow of a method for designing a UAV route network under traffic flow demand guidance provided by an embodiment of the present invention is as follows: Figure 1 As shown, the specific steps include:

[0111] Step 101, selecting a location for a candidate route convergence point.

[0112] Figure 2 A processing diagram of a method for scientifically selecting a location for an alternative route convergence point provided by an embodiment of the present invention. In order to ensure that the drone route network nodes meet multiple influencing factors to the maximum extent, a criticality ranking method is used to select route nodes suitable for drone take-off, landing and transfer from multiple alternative route convergence points. The specific location selection principle steps are as follows:

[0113] (1) Establish the original evaluation matrix X of m alternative route convergence points under n location factors ij =(x ij ) m×n , where x ij is the value of the jth alternative route convergence point under the site selection factor i. According to the operation characteristics of the UAV, the site selection factors may include environmental restriction factors such as restricted areas, dangerous areas and restricted areas, and socio-economic factors such as population density and traffic convenience;

[0114] (2) The original evaluation matrix is ​​dimensionless processed to obtain the standard evaluation matrix Y that eliminates the order of magnitude differences between different indicators. ij =(y ij ) m×n ,in

[0115]

[0116] (3) Calculate the weight value w corresponding to the location factor i i ,in

[0117]

[0118] Where E i is the entropy value of location factor i.

[0119]

[0120] (4) Determine the optimal solution for each location factor and the worst solution

[0121]

[0122] (5) Calculate the criticality score R of each alternative route convergence point j , sort and get the first Φ (Φ is the preset value) route convergence points, and use the first Φ route convergence points as the nodes of the UAV route network.

[0123]

[0124] Step 102, initializing the drone route network topology.

[0125] The maximum range of the drone The adjacency matrix A of the route convergence point is established based on (a mn ) Φ×Φ , if the length of the segment between the mth route convergence point and the nth route convergence point (ω is the margin coefficient), then a mn =1, otherwise a mn =0.

[0126] The geographical environment in the airspace is segmented according to the preset grid granularity, and the obstacle grid and free grid are marked according to whether they contain severe weather prone areas, airport clearance protection areas, etc., and the variable length A* path search algorithm is used to generate the optimal route between two connected route convergence points. A route search process provided by an embodiment of the present invention is as follows Figure 3 As shown, the processing steps include the following:

[0127] Step 1: Initialize the search space. Set a buffer zone with the straight line between the mth route convergence point and the nth convergence point as the center to form a search space.

[0128] Step 2: Estimated cost function setting. Define the estimated cost function f(x) of grid point x, including the actual voyage cost g(x) and the estimated voyage cost h(x):

[0129]

[0130] Among them, g(x) is the actual distance from the starting convergence point m to the grid point x to be expanded, h(x) is the estimated distance from the grid point x to be expanded to the ending convergence point n, (X x ,Y x ) are the grid point coordinates.

[0131] Step 3: The feasible_set and finish_set lists are established. The starting convergence point m is taken as the parent node, and the estimated cost function value of each free grid point around the starting convergence point m is calculated and added to the feasible_set list. In order to improve the efficiency of path search, the search step length is determined according to the proportion of obstacle grids in the local search space, that is, when the proportion of obstacle grids is less than 30%, the step length is 2 grid units, otherwise it is 1 grid unit.

[0132] Step 4: Grid point expansion. Select the grid point P with the smallest f(x) in the feasible_set list as the next expanded grid point, and add the grid point P to the finish_set list to determine whether the grid point P is the final convergence point n. If so, go to Step 7, otherwise go to Step 5.

[0133] Step 5: Check the adjacent grid points of point P. Take grid point P as the new parent node and search for adjacent free grid points according to the search step length corresponding to the obstacle grid ratio around grid point P;

[0134] Step 6: Recalculate the estimated cost function f(x) and return to Step 4.

[0135] Step 7: Store the optimal path obtained by the search algorithm, which is a free grid point sequence that minimizes the path cost between the mth route convergence point and the nth route convergence point.

[0136] Step 103, with the goal of minimizing the risk of traffic flow operation, construct an optimization model of the UAV route network.

[0137] The traffic flow combination distribution K (K = Φ(Φ-1) / 2) of the origin and destination nodes where there are UAV flight demands is statistically analyzed. With the goal of minimizing the risk of traffic flow operation, the optimization model of the UAV route network N(V, E) is established as follows:

[0138]

[0139] Where N is the route network topology to be optimized, including Φ route aggregation nodes, and E is the set of all routes in N. mn |m,n∈V and a mn =1}. mn , lmn and r mn Route e mn The route flow, route mileage and route congestion level on the route are defined as:

[0140]

[0141] Specifically, k is the flow rate between the kth traffic flow OD pair, R k is the number of optional paths for traffic flow k, The probability of selecting path i for traffic flow k, is a 0-1 variable. When the selected path i of traffic flow k contains e mn hour Otherwise it is 0. mn is the route capacity.

[0142] Corresponding to the actual operation of the route network, the optimization model should also meet the following constraints:

[0143]

[0144] f mn ≥f o (10)

[0145] Among them, constraint (9) ensures the conservation of the UAV path flow and the OD pair flight flow, that is, the sum of the flow on the feasible section of the OD pair is equal to the traffic demand between the OD pair. Constraint (10) is the route utilization constraint, which requires that the route allocation flow must not be less than the route activation threshold.

[0146] Step 104, iteratively update and solve the optimization model of the UAV route network to obtain a design plan for the UAV route network that minimizes the risk of traffic flow operation.

[0147] The optimization model of the UAV route network is iteratively solved by using a genetic algorithm with double-layer coding to obtain the route network topology structure and the corresponding traffic flow distribution result that minimizes the risk of UAV traffic flow operation. The binary coding of the route network topology is the upper coding, and the real number coding of the traffic flow distribution is the lower coding. The solution process of a genetic algorithm with double-layer coding provided by the embodiment of the present invention is as follows: Figure 4 shown.

[0148] Step 1: Population initialization. Randomly generate the initial population p(0) = {X1, X2, ...X N}, N is the population size. i is the i-th chromosome, consisting of two layers, Figure 5 A schematic diagram of a vector form of a first-layer alternative network topology adjacency matrix provided in an embodiment of the present invention. Figure 6A schematic diagram of the flow probability distribution results of k traffic flow OD pairs corresponding to a second-layer network topology provided by an embodiment of the present invention, Figure 6 In the above example, considering that the operation risk of traffic flow k on the route is dynamically updated with the allocated flow 1, 2, ... k-1, let the path selection probability The flight path i with the lowest risk in the current operation, i=1,2,...R k 1 at the position and 0 at the rest, that is,

[0149]

[0150] Step 2: Selection operation. The fitness function is expressed as the inverse of the objective function, that is, fitness = 1 / R(N). The smaller the risk value, the higher the fitness, and the greater the probability of being selected. The chromosomes in the population are linearly sorted from small to large according to the fitness size. The roulette selection strategy is used to select proportionally according to the fitness value.

[0151] Step 3: Crossover operation. Since the lower-level traffic flow distribution code will change with the upper-level route network topology code, the crossover operation only operates on the upper-level route network topology code, that is, the network topology code after the intersection of two chromosomes is selected according to the crossover probability. If an illegal chromosome is generated after the crossover (such as lack of connectivity or exceeding the maximum range constraint of the drone), the chromosome needs to be repaired, and the gene bits corresponding to the missing or overlong edges need to be modified. At the same time, the lower-level traffic flow distribution code should also be adjusted accordingly.

[0152] Step 4: Mutation operation. The mutation operation is also only performed on the upper-level route network topology coding. A chromosome is selected according to the mutation probability and a gene position is randomly selected to assign a new value. If an illegal chromosome is generated, the chromosome is repaired using a repair method.

[0153] Step 5: Iteration operation. Repeatedly generate new populations through the survival of the fittest until the maximum number of iterations is exceeded, and the optimal chromosome with the lowest overall operation risk is decoded to obtain the final result.

[0154] Embodiment 2

[0155] The structure diagram of a UAV route network design device under traffic flow demand guidance provided by an embodiment of the present invention is as follows: Figure 7As shown, it can specifically include a waypoint site selection module 1, a network topology initialization module 2, a network optimization target modeling module 3 and a route network iterative update module 4. Among them, the waypoint site selection module 1 is used to realize the scientific site selection of route convergence points under multiple construction elements of drone take-off and landing points; the network topology initialization module 2 is used to search and generate the initial topology of the drone route network; the network optimization target modeling module 3 is used to establish a drone route network model that minimizes the risk of traffic flow operation; the route network iterative update module 4 is used to solve the optimal drone route network and the corresponding traffic flow distribution under double-layer coding.

[0156] The waypoint site selection module 1 may include an evaluation matrix processing module 11 and a waypoint sorting module 12. The evaluation matrix processing module 11 is used to generate an evaluation matrix of drone take-off and landing points under multiple site selection factors such as environmental restriction factors and socio-economic factors, and the waypoint sorting module 12 is used to select excellent candidate points as nodes of the drone route network.

[0157] The network topology initialization module 2 may include an adjacency matrix generation module 21 and a route generation module 22. The adjacency matrix generation module 21 is used to obtain the adjacency relationship of the route network nodes under the UAV range performance constraint, and the route generation module 22 is used to generate a free obstacle-free route between two connected nodes in a gridded airspace environment.

[0158] The network optimization target modeling module 3 may include a risk characterization module 31 and a constraint restriction module 32. The risk characterization module 31 is used to define the operational risk of the drone traffic flow, and the constraint restriction module 32 is used to define the necessary conditions that the drone route network should meet.

[0159] The route network iteration update module 4 may include a double-layer coding module 41 and an algorithm solution module 42. The double-layer coding module 41 is used to generate alternative solutions containing route structure information and traffic flow distribution information, and the algorithm solution module 42 is used to obtain the optimal drone route network structure and the corresponding drone traffic flow distribution.

[0160] The specific process of using the device of the embodiment of the present invention to design a UAV route network under the guidance of traffic flow demand is similar to the aforementioned method embodiment and will not be repeated here.

[0161] In summary, the embodiment of the present invention combines the route network topology design with the distribution of UAV traffic flow demand from the traffic flow level, and reasonably selects the route network waypoints and waypoint edge relationships under the premise of meeting the constraints such as UAV range performance limit, airspace obstacle limit, and route utilization requirements, and finally obtains the UAV route network design scheme that minimizes the traffic flow operation risk. The following steps are mainly used:

[0162] (1) Taking into account the environmental restrictions, social and economic factors involved in the construction of UAV take-off and landing points, scientifically select the location of alternative route convergence points;

[0163] (2) Establish an adjacency matrix of route convergence points and use the path search algorithm to generate the initial topology of the UAV route network in the gridded airspace;

[0164] (3) Based on the distribution of UAV traffic flow demand, the UAV route network target modeling is carried out to minimize the traffic flow operation risk;

[0165] (4) A double-layer coded genetic algorithm is used to solve the established network optimization target model, and the initial topology of the UAV route network is iteratively adjusted until convergence.

[0166] The present invention can realize the reasonable planning of UAV route network under the guidance of traffic flow demand, while ensuring the safety level of UAV traffic flow operation, realize the organic matching of traffic flow demand and airspace structure configuration, and ultimately provide technical support for promoting large-scale commercial application of UAV in the future.

[0167] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0168] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention or certain parts of the embodiments.

[0169] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0170] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for designing a UAV route network guided by traffic flow demand, characterized in that: include: Select the nodes of the UAV route network and initialize the topology of the UAV route network; Based on the initialized UAV route network topology, an optimization model of the UAV route network is constructed with the goal of minimizing the risk of traffic flow operation; Solving the optimization model of the UAV route network through a genetic algorithm to obtain a design scheme for the UAV route network that minimizes the risk of traffic flow operation; The site selection of nodes of the UAV route network includes: The critical ranking method is used to select route nodes suitable for UAV take-off, landing and transfer from multiple alternative route convergence points, and the original evaluation matrix X of m alternative route convergence points under n site selection factors is established. ij =(x ij ) m×n , x ij is the value of the jth alternative route convergence point under the location factor i; for the original evaluation matrix X ij After dimensionless processing, we can get the standard evaluation matrix Y that eliminates the order of magnitude differences between different indicators. ij =(y ij ) m×n , Calculate the weight value w corresponding to the location factor i i ,in Where E i is the entropy value of location factor i; Determine the best solution for each location factor and the worst solution Calculate the criticality score R of each alternative route convergence point j , sort and get the first Φ route convergence points, and use the first Φ route convergence points as the nodes of the UAV route network; The initialization of the UAV route network topology includes: The maximum range of the drone The adjacency matrix A of the route convergence point is established based on (a mn ) Φ×Φ , if the length of the segment between the mth route convergence point and the nth route convergence point ω is the margin coefficient, then a mn =1, otherwise a mn = 0, the geographical environment in the airspace is segmented according to the preset grid granularity, and the variable-length A* path search algorithm is used to generate the optimal route between two connected route convergence points, including the following processing steps: Step 1: Set a buffer zone with the straight line between the mth route convergence point and the nth route convergence point as the center to form an initialized search space; Step 2: Define the estimated cost function f(x) of grid point x, including the actual voyage cost g(x) and the estimated voyage cost h(x): g(x) is the actual voyage cost from the starting convergence point m to the grid point x to be expanded, h(x) is the estimated voyage cost from the grid point x to be expanded to the ending convergence point n, (X x ,Y x ) are grid point coordinates; Step 3: Create feasible_set and finish_set lists, take the starting convergence point m as the parent node, calculate the estimated cost function value of each free grid point around the starting convergence point m, and add the estimated cost function value to the feasible_set list. Determine the search step length according to the proportion of obstacle grids in the local search space. When the proportion of obstacle grids is less than 30%, the step length uses 2 grid units, otherwise it uses 1 grid unit. Step 4: Select the grid point P with the smallest f(x) in the feasible_set list as the next extended grid point, and add the grid point P to the finish_set list to determine whether the grid point P is the final convergence point n. If so, go to Step 7, otherwise go to Step 5; Step 5: Take the grid point P as the new parent node and search for adjacent free grid points according to the search step length corresponding to the obstacle grid ratio around the grid point P; Step 6: Recalculate the estimated cost function f(x) and return to Step 4; Step 7: Store the optimal path obtained by the search algorithm, which is a free grid point sequence that minimizes the path cost between the mth route convergence point and the nth route convergence point; The optimization model of the UAV route network is constructed based on the initialized UAV route network topology with the goal of minimizing the risk of traffic flow operation, including: The traffic flow combination distribution K of the starting and ending route convergence nodes with UAV flight demand is statistically analyzed, K = Φ(Φ-1) / 2, and the optimization model of the UAV route network N(V,E) is established with the goal of minimizing the traffic flow operation risk as follows: Where N is the route network topology to be optimized, including Φ route aggregation nodes, and E is the set of all routes in N. mn |m,n∈V and a mn =1}, f mn , l mn and r mn Route e mn The route flow, route mileage and route congestion level on the route are defined as: f k is the flow rate between the kth traffic flow OD pair, R k is the number of optional paths for traffic flow k, The probability of selecting path i for traffic flow k, is a 0-1 variable. When the selected path i of traffic flow k contains e mn hour Otherwise it is 0, c mn is the route capacity; The optimization model satisfies the following constraints: f mn ≥f ο (10) Constraint (9) ensures the conservation of the UAV path flow and the OD pair flight flow, that is, the sum of the flow on the feasible section of the OD pair is equal to the traffic demand between the OD pair. Constraint (10) is the route utilization constraint, which requires that the route allocation flow must not be less than the route activation threshold; The method of solving the optimization model of the drone route network by a genetic algorithm to obtain a design scheme of the drone route network that minimizes the risk of traffic flow operation includes: The binary code of the route network topology is set as the upper layer code, and the real number code of the traffic flow distribution is set as the lower layer code. The optimization model of the UAV route network is solved by using a genetic algorithm with double-layer coding, including the following processing steps: Step 1: Randomly generate an initial population p(0) = {x1, x2, ... x N′ }, N′ is the population size, x i′ is the i′th chromosome, which consists of two layers. Considering that the operation risk of traffic flow k on the route is dynamically updated with the allocated flow 1, 2, ... k-1, let the path selection probability The flight path i with the lowest risk in the current operation, i=1,2,...R k Take 1 at the position and 0 at the rest, that is: Step 2: Express the fitness function as the inverse of the objective function, that is, fitness = 1 / R(N′). The smaller the risk value, the higher the fitness, and the greater the probability of being selected. The chromosomes in the population are linearly sorted from small to large according to the fitness size, and the roulette wheel selection strategy is used to select proportionally according to the fitness value; Step 3: Perform a crossover operation on the upper-layer route network topology code, that is, select two chromosomes corresponding to the crossover probability to exchange the network topology code after the intersection of the two chromosomes. If an illegal chromosome is generated after the crossover, the chromosome is repaired and the lower-layer traffic flow distribution code is adjusted accordingly. Step 4: Perform mutation operation on the upper-layer route network topology coding, select a chromosome according to the mutation probability and randomly select a gene position to assign a new value. If an illegal chromosome is generated, the chromosome is repaired by using a repair method; Step 5: Repeatedly generate new populations through the survival of the fittest until the maximum number of iterations is exceeded, and the optimal chromosome decoding with the lowest overall operation risk is taken as the final result to obtain the design plan of the UAV route network that minimizes the traffic flow operation risk.

2. A UAV route network design device guided by traffic flow demand, characterized in that: include: Waypoint location selection module, network topology initialization module, network optimization target modeling module and route network iterative update module; The waypoint site selection module is used to select the nodes of the UAV route network; The network topology initialization module is used to initialize the network topology of the UAV route; The network optimization target modeling module is used to construct an optimization model of the UAV route network based on the initialized UAV route network topology with the goal of minimizing the traffic flow operation risk; The route network iteration update module is used to solve the optimization model of the UAV route network through a genetic algorithm to obtain a design scheme of the UAV route network that minimizes the risk of traffic flow operation; The waypoint location selection module is used to select route nodes suitable for UAV take-off, landing and transfer from multiple alternative route convergence points using a critical sorting method, and establish an original evaluation matrix X of m alternative route convergence points under n location selection factors. ij =(x ij ) m×n , x ij is the value of the jth alternative route convergence point under the location element i; For the original evaluation matrix X ij After dimensionless processing, we can get the standard evaluation matrix Y that eliminates the order of magnitude differences between different indicators. ij =(y ij ) m×n , Calculate the weight value w corresponding to the location factor i i ,in Where E i is the entropy value of location factor i; Determine the best solution for each location factor and the worst solution Calculate the criticality score R of each alternative route convergence point j , sort and get the first Φ route convergence points, and use the first Φ route convergence points as the nodes of the UAV route network; The network topology initialization module is used to initialize the UAV with the maximum range The adjacency matrix A of the route convergence point is established based on (a mn ) Φ×Φ , if the length of the segment between the mth route convergence point and the nth route convergence point ω is the margin coefficient, then a mn =1, otherwise a mn = 0, the geographical environment in the airspace is segmented according to the preset grid granularity, and the variable-length A* path search algorithm is used to generate the optimal route between two connected route convergence points, including the following processing steps: Step 1: Set a buffer zone with the straight line between the mth route convergence point and the nth route convergence point as the center to form an initialized search space; Step 2: Define the estimated cost function f(x) of grid point x, including the actual voyage cost g(x) and the estimated voyage cost h(x): g(x) is the actual voyage cost from the starting convergence point m to the grid point x to be expanded, h(x) is the estimated voyage cost from the grid point x to be expanded to the ending convergence point n, (X x ,Y x ) are grid point coordinates; Step 3: Create feasible_set and finish_set lists, take the starting convergence point m as the parent node, calculate the estimated cost function value of each free grid point around the starting convergence point m, and add the estimated cost function value to the feasible_set list. Determine the search step length according to the proportion of obstacle grids in the local search space. When the proportion of obstacle grids is less than 30%, the step length uses 2 grid units, otherwise it uses 1 grid unit. Step 4: Select the grid point P with the smallest f(x) in the feasible_set list as the next extended grid point, and add the grid point P to the finish_set list to determine whether the grid point P is the final convergence point n. If so, go to Step 7, otherwise go to Step 5; Step 5: Take the grid point P as the new parent node and search for adjacent free grid points according to the search step length corresponding to the obstacle grid ratio around the grid point P; Step 6: Recalculate the estimated cost function f(x) and return to Step 4; Step 7: Store the optimal path obtained by the search algorithm, which is a free grid point sequence that minimizes the path cost between the mth route convergence point and the nth route convergence point; The network optimization target modeling module is used to calculate the traffic flow combination distribution K of the start and end route convergence nodes with UAV flight demand, K = Φ(Φ-1) / 2, and to minimize the traffic flow operation risk. The optimization model of the UAV route network N(V, E) is established as follows: Where N is the route network topology to be optimized, including Φ route aggregation nodes, and E is the set of all routes in N. mn |m,n∈V and a mn =1}, f mn , l mn and r mn Route e mn The route flow, route mileage and route congestion level on the route are defined as: f k is the flow rate between the kth traffic flow OD pair, R k is the number of optional paths for traffic flow k, The probability of selecting path i for traffic flow k, is a 0-1 variable. When the selected path i of traffic flow k contains e mn hour Otherwise it is 0, c mn is the route capacity; The optimization model satisfies the following constraints: f mn ≥f ο (10) Constraint (9) ensures the conservation of the UAV path flow and the OD pair flight flow, that is, the sum of the flow on the feasible section of the OD pair is equal to the traffic demand between the OD pair. Constraint (10) is the route utilization constraint, which requires that the route allocation flow must not be less than the route activation threshold; The route network iterative update module is used to set the binary code of the route network topology as the upper layer code and the real number code of the traffic flow distribution as the lower layer code, and adopts the genetic algorithm of the double-layer code to solve the optimization model of the UAV route network, including the following processing steps: Step 1: Randomly generate an initial population p(0) = {x1, x2, ... x N' }, N' is the population size, x i' is the i′th chromosome, which consists of two layers. Considering that the operation risk of traffic flow k on the route is dynamically updated with the allocated flow 1, 2, ... k-1, let the path selection probability The flight path i with the lowest risk in the current operation, i=1,2,...R k Take 1 at the position and 0 at the rest, that is: Step 2: Express the fitness function as the inverse of the objective function, that is, fitness = 1 / R(N′). The smaller the risk value, the higher the fitness, and the greater the probability of being selected. The chromosomes in the population are linearly sorted from small to large according to the fitness size, and the roulette wheel selection strategy is used to select proportionally according to the fitness value; Step 3: Perform a crossover operation on the upper-layer route network topology code, that is, select two chromosomes corresponding to the crossover probability to exchange the network topology code after the intersection of the two chromosomes. If an illegal chromosome is generated after the crossover, the chromosome is repaired and the lower-layer traffic flow distribution code is adjusted accordingly. Step 4: Perform mutation operation on the upper-layer route network topology coding, select a chromosome according to the mutation probability and randomly select a gene position to assign a new value. If an illegal chromosome is generated, the chromosome is repaired by using a repair method; Step 5: Repeatedly generate new populations through the survival of the fittest until the maximum number of iterations is exceeded, and the optimal chromosome decoding with the lowest overall operation risk is taken as the final result to obtain the design plan of the UAV route network that minimizes the traffic flow operation risk.

Images