Urban low-altitude unmanned aerial vehicle route dynamic planning method

Abstract

The invention discloses a dynamic planning method for an air route of an urban low-altitude unmanned aerial vehicle. The method comprises the following steps: firstly, constructing an urban low-altitude environment model based on an airspace rasterization technology, fusing multi-dimensional constraint factors such as geographic data, meteorological conditions and airspace control information through multi-source environment information, and accurately calibrating the position of a take-off and landing point; secondly, a dynamic grid availability evaluation model is established in combination with environmental constraints and a real-time airspace state, and the navigation feasibility of each grid unit is quantitatively analyzed; then, an improved A * algorithm is combined with a grid availability evaluation result to generate a global optimal initial route; finally, a rolling time domain optimization strategy is introduced, and uncertain factors such as sudden obstacles and airspace dynamic limitation are responded in real time through periodic non-flight route re-planning after the unmanned aerial vehicle takes off. Through collaborative fusion of static environment modeling and a dynamic optimization mechanism, the problem of real-time planning of the air route of the unmanned aerial vehicle in the urban low-altitude complex environment is effectively solved, and the environmental adaptability and task reliability of the system are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle (UAV) airspace management, and in particular relates to a method for dynamic planning of urban low-altitude UAV routes. Background Art

[0002] In recent years, with the widespread application of drone technology in logistics, agricultural plant protection, and emergency rescue, the number of drones operating in low-altitude airspace has exploded. Efficient and safe route planning has become crucial for operational safety. Dynamic route planning for drones, particularly in urban areas with high-density low-altitude flight, directly impacts flight safety and efficiency. Existing methods are often based on static airspace models and fail to effectively integrate multiple sources of information, such as real-time meteorological data, dynamic obstacles, and communication and navigation signals. This results in insufficient adaptability of route planning in complex environments. Traditional algorithms, such as A* and Dijkstra, suffer from low computational efficiency and poor real-time performance. While they can generate globally optimal paths, they lack the ability to dynamically respond to real-time airspace conditions, which can easily lead to sudden course changes or node congestion, increasing collision risk and energy consumption. To address these issues, this paper proposes a method for dynamic route planning for low-altitude drones in urban areas. This method utilizes multi-source data fusion modeling, real-time environmental perception, and optimization algorithms to address the route planning problem in low-altitude urban environments. Furthermore, through periodic route replanning to respond to sudden obstacles and dynamic airspace constraints, it enables real-time dynamic route planning in complex environments, significantly improving flight safety and operational efficiency. Summary of the Invention

[0003] In response to the above problems, the present invention discloses a method for dynamic planning of urban low-altitude UAV routes. First, an urban low-altitude environmental model is constructed based on airspace rasterization technology, and multi-dimensional constraint factors such as geographic data, meteorological conditions, and airspace control information are integrated through multi-source environmental information, and the take-off and landing points are accurately calibrated; secondly, a dynamic grid availability evaluation model is established in combination with environmental constraints and real-time airspace status to quantitatively analyze the navigation feasibility of each grid unit; then, an improved A* algorithm is used in combination with the grid availability evaluation results to generate a global optimal initial route; finally, a rolling time domain optimization strategy is introduced to respond to sudden obstacles, airspace dynamic restrictions and other uncertainties in real time through periodic non-flying route re-planning after the UAV takes off. The present invention effectively solves the problem of real-time planning of UAV routes in complex urban low-altitude environments through the synergistic integration of static environmental modeling and dynamic optimization mechanisms, and significantly improves the environmental adaptability and mission reliability of the system.

[0004] The present invention aims to solve the following technical problems:

[0005] Static airspace models cannot dynamically adapt to real-time changes in information such as different airspace structures, dynamic weather, obstacles, and communication and navigation signals;

[0006] Traditional algorithms cannot meet the real-time response requirements of low-altitude, high-density urban flights;

[0007] Insufficient quantification of environmental risks in single-aircraft route planning leads to low route safety;

[0008] The average response time of existing methods to emergencies is too long, and the replanned route has poor smoothness, which does not meet the dynamic constraints of the UAV.

[0009] In order to achieve the above technical objectives and the above technical effects, the present invention is implemented through the following technical solutions:

[0010] A method for dynamic planning of urban low-altitude UAV routes, comprising:

[0011] S1. Build an urban low-altitude environmental model based on airspace rasterization technology. By integrating multi-source environmental information, we integrate geographic data, meteorological conditions, airspace control information, and other multi-dimensional constraints, and calibrate the locations of drone take-off and landing points:

[0012] Discretize the planning space into a 3D grid array:

[0013] G={g ijk |i=1,2,...,m;j==1,2,...,n;k=1,2,...,p}

[0014] Where g ijk represents the grid cell in row i, column j, and layer k, and m×n×p is the grid resolution.

[0015] Divide the target airspace into three-dimensional grid cells with a side length of 1000m×1000m, and layer them at intervals of 50m in the height direction;

[0016] In the grid environment, the take-off point S of the UAV is calibrated. s ,y s ,z s ) and landing point T=(x t ,y t ,z t ), and integrates multi-dimensional constraints such as geographic data, meteorological conditions, and airspace control information to build an environmental model.

[0017] S2. Build a dynamic grid availability assessment model by combining environmental constraints with real-time airspace status (airspace structure, real-time meteorological data, obstacles, and communication and navigation signals).

[0018] Based on three meteorological information, namely wind speed, precipitation intensity and visibility, obstacle location, airspace structure and communication and navigation signal parameters, a weighted fusion model is used to quantitatively evaluate the navigation feasibility of each grid cell. The formula is:

[0019]

[0020] Where K is the basic coefficient of airspace structure, v is the real-time wind speed, p is the precipitation intensity, and p max is the UAV rain resistance level threshold, d vis is the visibility distance, d obs is the distance from the grid center to the nearest obstacle, C is the communication quality index, N is the navigation signal strength index, is the weight coefficient and satisfies

[0021] Among them, when the airspace range is a no-fly airspace, K=0; when the airspace range is a controlled airspace, K=0.3; when the airspace range is a restricted airspace, K=0.6; when the airspace range is a suitable-fly airspace, K=1.

[0022] Wherein, the grid availability weight W g ∈[0,1], the higher the weight value, the better the grid availability. When the grid is completely occupied by obstacles, set W g =0 means the grid is unavailable. g ∈[0.8,1] indicates that the state of the grid area is excellent and it can be passed through the grid area first; when W g ∈[0.6,0.8) indicates that the grid area is in good condition and can be used normally; when W g ∈[0.4,0.6) indicates that the grid area is in a warning state and the speed limit is required to pass through the grid area; when W g <0.4 means that the grid area is disabled and the drone must automatically avoid the grid area during flight.

[0023] S3. Based on the grid availability evaluation results, an improved A* algorithm is used to generate an initial single UAV optimal route. The cost function of the improved A* algorithm is:

[0024] F(n)=G(n)+H(n)+λ·Σ(1-W g )

[0025] Where G(n) is the actual flight distance cost from the take-off point S to the current node n, calculated as the flight distance; H(n) is the heuristic function cost from the current node to the destination, calculated using the Manhattan distance from the current node n to the landing point T; ∑(1-W g ) is the safety cost of the route, which is the sum of the availability costs of all grid nodes that the route passes through; λ is the safety adjustment coefficient, which is used to adjust the safety cost ∑(1-W g ) units are equivalent to the units of G(n) and H(n) and are dynamically adjusted according to the safety level of the flight mission.

[0026] Based on this, the constraints for improving the A* algorithm to design routes include:

[0027]

[0028] Where W g (n) is the availability weight of the grid where node n is located, φ n is the heading angle of the UAV at node n, Δφ max is the maximum allowable turning angle, P k is the three-dimensional position coordinate of the UAV at time k (x k ,y k ,z k ), P obs is the real-time position coordinates (x obs ,y obs ,z obs ), d safe is the minimum safety distance, C(n) is the communication signal strength of node n, and C min To maintain the minimum signal threshold required for the control link, E l is the estimated energy consumption of flight segment l, E max The maximum available energy of the drone.

[0029] S4. Introducing a rolling horizon optimization strategy to periodically re-route the drone after takeoff to respond to unexpected obstacles and dynamic airspace restrictions in real time, continuously ensuring route safety and operational efficiency; including:

[0030] When the drone detects W after taking off g (k)≤0.4 and the grid is located at the current position of the drone P current When there is an unflying segment between the landing point T, re-planning is triggered immediately, and the rolling horizon optimization method is used to update the unflying subsequent segments with the current drone position as the new starting point. The specific method is as follows:

[0031]

[0032] Where U t:t+H is the control input from the current time k to the predicted time domain H, including the motion control instructions of the UAV, such as speed, heading angle change rate, etc.; P k is the predicted position of the UAV at time k; T is the coordinate of the landing point; W g (P k ) is the position P k The availability weight of the grid; μ, ρ, ε are weight coefficients, and μ+ρ+ε=1.

[0033] Among them, the objective function J represents the comprehensive optimization goal of the UAV route dynamic replanning, which is used to quantitatively evaluate the quality of the candidate routes. The smaller the J value, the better the route. k -T||2 represents the terminal cost, which indicates the position of the drone in the predicted time domain P k The Euclidean distance from the landing point T guides the drone to approach the target as quickly as possible to avoid detouring too far; 1-W g (P k ) represents the safety cost, which penalizes passing through low-safety areas and forces routes to avoid risks; ||U k || 2 Represents the control cost, avoiding sharp turns or sudden speed changes, ensuring flight smoothness and energy consumption optimization.

[0034] Based on this, when the rolling horizon optimization strategy is subsequently adopted to respond to sudden obstacles and dynamic airspace restrictions in real time through periodic route replanning, the constraints of the generated route should also meet the constraints of the route designed by the improved A* algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive work, among which:

[0036] Figure 1 This is a flow chart of a method for dynamic planning of urban low-altitude UAV routes according to an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of a three-dimensional grid model of an urban low-altitude environment according to an embodiment of the present invention;

[0038] Figure 3 A schematic diagram of three-dimensional grid availability distribution according to an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of a single route generated based on an improved A* algorithm according to an embodiment of the present invention;

[0040] Figure 5 A schematic diagram of dynamic replanning of a rolling time domain optimized route according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] To make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. It should be noted that the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0042] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating relative importance or implying the number of technical features. The terms "including" and "comprising" should be understood in a broad sense, that is, they cover the content without excluding elements that are not explicitly listed. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0043] The present invention provides a method for dynamic planning of urban low-altitude UAV routes. Figure 1 As shown, the following steps are included:

[0044] (a) Build an urban low-altitude environmental model based on airspace rasterization technology, integrate multi-source environmental information with geographic data, meteorological conditions, airspace control information, and other multi-dimensional constraints, and accurately calibrate the take-off and landing points;

[0045] (b) Combine environmental constraints with real-time airspace status to establish a dynamic grid availability assessment model to quantitatively analyze the air traffic feasibility of each grid cell;

[0046] (c) Using the improved A* algorithm combined with the grid availability assessment results to generate the global optimal initial route;

[0047] (d) A rolling horizon optimization strategy is introduced to respond to uncertainties such as sudden obstacles and dynamic airspace restrictions in real time by periodically replanning unflying routes after the UAV takes off.

[0048] In a specific embodiment of the present invention, Figure 2 As shown, a schematic diagram of an urban low-altitude environment model of an embodiment of the present invention is shown, wherein the method includes:

[0049] The target airspace of 5000m×5000m×500m is divided into two-dimensional grids with a side length of 1000m×1000m in the horizontal direction and layers of 50m in the vertical direction;

[0050] Real-time wind speed, precipitation intensity, and visibility data are acquired through meteorological sensors; obstacle information is acquired through obstacle detection radar; communication quality index and navigation signal strength are acquired in real time through airborne communication modules and satellite navigation receivers; and airspace structure types are dynamically updated through digital airspace maps published by the air traffic control system;

[0051] The take-off point S = (100, 200, 50) and landing point T = (4500, 4800, 50) of the UAV are calibrated in a grid environment, and multi-dimensional constraints such as geographic data, meteorological conditions, and airspace control information are integrated to construct an environmental model.

[0052] In a specific embodiment of the present invention, Figure 3 As shown in the three-dimensional grid availability distribution diagram, the method includes:

[0053] A dynamic grid availability assessment model is constructed by combining information such as airspace structure, real-time meteorological data, obstacles, and communication and navigation signals:

[0054] Based on three meteorological information, namely wind speed, precipitation intensity and visibility, obstacle location, airspace structure and communication and navigation signal parameters, a weighted fusion model is used to quantify the availability of grid cells. The formula is:

[0055]

[0056] Where K is the basic coefficient of airspace structure, v is the real-time wind speed, p is the precipitation intensity, and p max is the UAV rain resistance level threshold, d vis is the visibility distance, d obs is the distance from the grid center to the nearest obstacle, C is the communication quality index, N is the navigation signal strength index, is the weight coefficient and satisfies

[0057] Among them, when the airspace range is a no-fly airspace, K=0; when the airspace range is a controlled airspace, K=0.3; when the airspace range is a restricted airspace, K=0.6; when the airspace range is a suitable-fly airspace, K=1.

[0058] The coefficients of each indicator are:

[0059] α=0.25, β=0.2, γ=0.2, δ=0.15, η=0.1, p max =20mm / h, v max =15m / s;

[0060] According to the threshold classification grid state, when W g∈[0.8,1] indicates that the state of the grid area is excellent and it can be passed through the grid area first; when W g ∈[0.6,0.8) indicates that the grid area is in good condition and can be used normally; when W g ∈[0.4,0.6) indicates that the grid area is in a warning state and the speed limit is required to pass through the grid area; when W g <0.4 means that the grid area is disabled and the drone must automatically avoid the grid area during flight.

[0061] In a specific embodiment of the present invention, Figure 4 The improved single route generation method of the A* algorithm, the route planning process includes:

[0062] Initialize the data, input the starting point, end point coordinates and grid weight data;

[0063] Starting from the starting point, expand the nodes according to the eight-neighborhood and calculate the cost of each node. The formula is:

[0064] F(n)=G(n)+H(n)+λ·∑(1-W g )

[0065] Where G(n) is the actual flight distance cost from the starting point to the current node n, calculated as the flight distance; H(n) is the estimated cost from the current node to the end point, calculated using the improved Manhattan distance considering airspace restrictions; Σ(1-W g ) is the route path safety cost factor, which accumulates the availability costs of all grid nodes that the route passes through; λ is the safety weight coefficient, which is dynamically adjusted according to the safety level of the flight mission.

[0066] The values ​​of each constraint condition are:

[0067] Δφ max =30°, d safe =100m, C min =-90dBm, E max =20000J.

[0068] Optionally, the specific constraints of the single route generation method of the improved A* algorithm are:

[0069]

[0070] Where W g (n) is the availability weight of the grid where node n is located, φ n is the heading angle of the UAV at node n, P k is the three-dimensional position coordinate of the UAV at time k (x k ,y k ,zk ), P obs is the real-time position coordinates (x obs ,y obs ,z obs )C(n) is the communication signal strength of node n, E l is the estimated energy consumption of flight segment l.

[0071] The following principles must be observed during route generation:

[0072] Prioritize grids with an "Excellent" status. If the path must pass through a "Warning" area, the speed limit is 5 m / s.

[0073] The route needs to completely avoid the "disabled" grid.

[0074] Finally, the generated routes are subjected to constraint checks.

[0075] In a specific embodiment of the present invention, Figure 5 The steps of executing the rolling horizon optimization (RHO) are as follows:

[0076] First, the triggering conditions of the rolling horizon optimization dynamic replanning method are:

[0077] Perform periodic testing with a cycle of 30 seconds;

[0078] W g <0.4 and the flight segment is in an unflown segment.

[0079] When W is detected g (k)≤0.4 and the grid is located at the current position of the drone P current When there is an unflying segment between the landing point T, re-planning is triggered immediately, and the rolling horizon optimization method is used to update the unflying subsequent segments with the current drone position as the new starting point. The specific method is as follows:

[0080]

[0081] Where U t:t+H is the control input from the current time k to the predicted time domain H, including the motion control instructions of the UAV, such as speed, heading angle change rate, etc.; P k is the predicted position of the UAV at time k; T is the coordinate of the landing point; W g (P k ) is the position P k The availability weight of the grid; μ, ρ, ε are weight coefficients, and μ+ρ+ε=1.

[0082] The values ​​of each constraint condition are:

[0083] H = 10 time steps. If the UAV has a fast response speed (such as a multi-rotor UAV), the time step is usually shorter (such as 1 to 5 seconds). If it is a fixed-wing UAV, the time step can be appropriately extended (such as 5 to 10 seconds); μ = 0.5, ρ = 0.3, ε = 0.2; the control input U k Medium speed change rate Δv≤3m / s, heading angle change rate Δφ≤15m / s

[0084] The replanning process based on rolling horizon optimization is as follows:

[0085] Take the current position of the drone as the starting point;

[0086] Call the improved A* algorithm to generate subsequent new flight segments to meet all constraints.

Claims

1. A method for dynamic planning of urban low-altitude UAV routes, characterized in that: The following steps are involved: (a) Build an urban low-altitude environmental model based on airspace rasterization technology, integrate multi-source environmental information with geographic data, meteorological conditions, and multi-dimensional constraints of airspace control information, and accurately calibrate the locations of take-off and landing points; (b) Combine environmental constraints with real-time airspace status to establish a dynamic grid availability assessment model to quantitatively analyze the air traffic feasibility of each grid cell; (c) Using the improved A* algorithm combined with the grid availability assessment results to generate the global optimal initial route; (d) A rolling horizon optimization strategy is introduced to respond to the uncertainties of sudden obstacles and dynamic airspace restrictions in real time by periodically replanning the unflying routes after the UAV takes off.

2. The method according to claim 1, characterized in that In the step (a), the urban low-altitude environment model is constructed based on the airspace rasterization technology as follows: Discretize the planning space into a 3D grid array: G={g ijk |i=1,2,...,m;j==1,2,...,n;k=1,2,...,p} Where g ijk represents the grid cell in row i, column j, and layer k, and m×n×p is the grid resolution; And calibrate the take-off point S of the UAV in the grid environment s ,y s ,z s ) and landing point T=(x t ,y t ,z t ).

3. The method according to claim 1, characterized in that In the step (b), the dynamic grid availability evaluation model is: Where K is the basic coefficient of airspace structure, v is the real-time wind speed, p is the precipitation intensity, and p max is the UAV rain resistance level threshold, d vis is the visibility distance, d obs is the distance from the grid center to the nearest obstacle, C is the communication quality index, N is the navigation signal strength index, α, β, γ, δ, η, is the weight coefficient and satisfies Among them, when the airspace range is a no-fly airspace, K = 0; when the airspace range is a controlled airspace, K = 0.3; when the airspace range is a restricted airspace, K = 0.6; when the airspace range is a suitable airspace, K = 1; Among them, the grid availability weight W g ∈[0,1], the higher the weight value, the better the grid availability; when the grid is completely occupied by obstacles, set W g =0 means the grid is unavailable; when W g ∈[0.8,1] indicates that the state of the grid area is excellent and it can be passed through the grid area first; when W g ∈[0.6,0.8) indicates that the grid area is in good condition and can be used normally; when W g ∈[0.4,0.6) indicates that the grid area is in a warning state and the speed limit is required to pass through the grid area; when W g <0.4 means that the grid area is disabled and the drone must automatically avoid the grid area during flight.

4. The method according to claim 1, wherein In step (c), the cost function of the improved A* algorithm is: F(n)=G(n)+H(n)+λ·∑(1-W g ) Where G(n) is the actual flight distance cost from the take-off point S to the current node n, calculated as the flight distance; H(n) is the heuristic function cost from the current node to the destination, calculated using the Manhattan distance from the current node n to the landing point T; ∑(1-W g ) is the safety cost of the route, which is the cumulative availability cost of all grid nodes that the route passes through; λ is the safety adjustment coefficient, and the safety cost ∑(1-W g ) units are equivalent to those of G(n) and H(n), converting the "risk value" into an "equivalent flight distance" that is dynamically adjusted based on the safety level of the flight mission.

5. The method according to claim 4, characterized in that In step (d), the rolling time domain optimization strategy is: when W is detected after the UAV takes off g (k)≤0.4 and the grid is located at the current position of the drone P current When there is an unflown segment between the current UAV position and the landing point T, re-planning is triggered immediately. The rolling horizon optimization method is used to update the unflown subsequent segments with the current UAV position as the new starting point. The specific method is as follows: Where U t:t+H is the control input from the current time k to the predicted time domain H, including the motion control instructions of the UAV; P k is the predicted position of the UAV at time k; T is the coordinate of the landing point; W g (P k ) is the position P k The availability weight of the grid where it is located; μ, ρ, ε are weight coefficients, and μ+ρ+ε=1; Among them, the objective function J represents the comprehensive optimization goal of the UAV route dynamic replanning, which is used to quantitatively evaluate the quality of the candidate routes. The smaller the J value, the better the route. k -T||2 represents the terminal cost, which indicates the position of the drone in the predicted time domain P k The Euclidean distance from the landing point T guides the drone to approach the target as quickly as possible to avoid detouring too far; 1-W g (P k ) represents the safety cost, which penalizes passing through low-safety areas and forces routes to avoid risks; ||U k || 2 Represents the control cost, avoiding sharp turns or sudden speed changes, ensuring flight smoothness and energy consumption optimization.

Images