Unmanned aerial vehicle path planning method based on comprehensive risk cost

Abstract

The invention discloses an unmanned aerial vehicle path planning method based on comprehensive risk cost, which can generate a low-risk unmanned aerial vehicle flight path. The method comprises the steps that firstly, an unmanned aerial vehicle path planning model containing comprehensive risks and operation cost is constructed, the model comprises a target function and constraint conditions, and the target function is composed of airspace risk cost and unmanned aerial vehicle transportation cost; then, setting three-dimensional constraint conditions including a voyage constraint, a turning angle constraint and a maximum takeoff weight constraint, and constructing a feasible solution space conforming to physical characteristics of the unmanned aerial vehicle; then, performing initial path planning based on a heuristic search weight-A * algorithm with weight, and generating an initial feasible path meeting a constraint condition by dynamically adjusting a heuristic function weight balance path length and risk cost; and finally, based on a parallelogram rule, optimizing the turning point optimization of the initial planning path. According to the invention, a path planning solution with safety and economical efficiency is provided for unmanned aerial vehicle transportation in a complex low-altitude environment.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and specifically to a parallelogram turning point optimization path planning method that considers comprehensive risk costs. The method is particularly suitable for turning point optimization and flight safety improvement in low-altitude UAV route planning scenarios. Background Art

[0002] With the rapid development of drone technology and the ever-expanding scope of its applications, drones are increasingly being used in logistics, agricultural monitoring, emergency rescue, and other fields. However, low-altitude drone flight faces complex airspace environments and diverse risk factors. In particular, the placement of turning points in route planning directly impacts flight safety and efficiency. Traditional path planning methods focus on the shortest path or the lowest energy consumption, while ignoring the additional risks and costs associated with turning points.

[0003] The A* algorithm and its improved algorithms are widely used in existing drone path planning, but they have limitations when it comes to optimizing turning points. In low-altitude flight environments, factors such as the drone's steering angle, range limitations, and payload capacity place higher demands on path planning. Furthermore, traditional methods fail to fully consider the cumulative effect of turning points on flight risk, potentially leading to safety risks in actual operation of the planned path.

[0004] Therefore, there is an urgent need for a path planning method that comprehensively considers risk cost and turning point optimization. By constructing objective functions and constraints and combining parallelogram transformation technology, the turning point setting can be optimized, flight risks can be reduced, and the safety and efficiency of UAV flight can be improved. Summary of the Invention

[0005] Purpose of the invention: To provide a parallelogram turning point optimization path planning method considering comprehensive risk cost. By constructing a UAV path planning model, combining airspace risk cost and transportation cost, the turning point setting is optimized, the number of turning points is reduced, the flight risk is reduced, and the safety and efficiency of UAV flight are improved.

[0006] Technical solution: A method for two-dimensional grid generation and take-off and landing field grid configuration in urban low-altitude airspace, comprising the following steps:

[0007] S1: Construct a UAV path planning model that includes comprehensive risks and operating costs, including an objective function and constraints. The objective function is composed of the airspace risk cost and the UAV transportation cost.

[0008] S2: Set constraints, including range constraints, turning angle constraints, and maximum takeoff weight constraints, to construct a feasible solution space that conforms to the physical characteristics of the drone, ensuring that the drone meets the range, turning angle, and payload limits during flight.

[0009] S3: An initial path planning method based on the weighted heuristic search weight-A* algorithm balances path length and risk cost by dynamically adjusting the heuristic function weight, calculates the path planning cost function, and generates an initial planning path that meets the constraints;

[0010] S4: Based on the parallelogram turning point optimization method combined with risk calculation, the initial planned path is optimized to reduce the number of turning points and lower the flight risk;

[0011] Furthermore, the specific method for constructing the UAV path planning model described in S1, including the objective function and constraints, where the objective function is composed of the airspace risk cost and the UAV transportation cost, is:

[0012] Layer h q There is a directed graph G(N,E) in the grid, where N is h q Layer grid point set, E represents the edge set. When planning the low-altitude UAV route, the center coordinate of the starting grid S is (x s ,y s ,h q ), the center coordinate of the end grid E is (x e ,y e ,h q ), the center coordinate of the path grid t in route planning is (x t ,y t ,h q ), t represents the grid number of the route, and t=1,2,...,n-1,n, n is the total number of grids to which the route belongs. Its objective function mainly includes the airspace risk cost C risk_cost and drone transportation cost C trans_cost , which can be expressed as:

[0013]

[0014] Where: P t-1,t is the transportation operation cost from grid t-1 to adjacent grid t, and the formula is:

[0015] P t-1,t =l t-1,t ×f e ×τ(m cargo )

[0016] Among them: t-1,t is the length from grid i to the adjacent grid t; f e The cost paid for the unit energy consumption of the drone; τ(m cargo ) is the load m cargo The penalty coefficient when , which can be expressed as:

[0017]

[0018] Where: τ max is the maximum cargo loading penalty value, m max The maximum cargo weight.

[0019] In summary, the comprehensive objective function of the UAV path planning model can be expressed as:

[0020] minC cost =(C risk_cost +1)×C trans_cost

[0021] Furthermore, the specific method and steps for setting the constraints described in S2, including range constraints, turning angle constraints, and maximum take-off weight constraints, to ensure that the UAV meets the range, turning angle, and payload limits during flight are:

[0022] P1: The calculation formula for range constraint is:

[0023] d aircraft_min ≥len(r SE )≤d aircraft_max

[0024] Among them, the constraint is the flight range limit constraint of the UAV, and the planned route r SE The distance len(r SE ) cannot be less than the minimum range d of the UAV aircraft_min , nor can it be greater than its maximum range d aircraft_max ;

[0025] P2: The calculation formula for the turning angle constraint is:

[0026] 0≤δ j ≤δ max

[0027] Among them, this constraint is the maximum steering angle limit of the UAV, indicating that the UAV

[0028] Steering angle δ j The maximum steering angle δ that it can withstand must not be exceeded max ;

[0029] P3: Maximum Take-Off Weight The calculation formula for the maximum take-off weight constraint is:

[0030] m uav +m cargo <m max

[0031] Among them, this constraint is for the maximum take-off mass of the UAV, indicating that the UAV's own weight m uav and load m cargo The sum of the weights must be less than the maximum take-off weight of the drone m max .

[0032] Furthermore, the initial path planning method based on the weighted-A* algorithm described in S3 calculates the path planning cost function and generates the initial planned path in a specific manner:

[0033] The path planning cost function f(t) is:

[0034] f(t)=g(t)+h(t)

[0035] Where g(t) is the actual cost of moving from the starting grid s to the expanded grid t; h(t) is the estimated cost of moving from grid t to the end grid e. g(t) and h(t) are the actual cost and heuristic function of the Parallel-A* algorithm, respectively, and the formulas are:

[0036]

[0037] in, and x is the actual environmental operation risk cost and route planning risk cost generated from the starting point S to the extended grid corresponding to the extended route path; e and y e is the grid coordinate of grid e; x t and y t is the grid coordinate of grid t.

[0038] Furthermore, the specific method of optimizing the initial planned path based on the parallelogram turning point optimization method combined with risk calculation in S4, reducing the number of turning points and reducing flight risk is:

[0039] If segment ab / / cd, segment ab can be transformed into a parallelogram. When ab or bc satisfies:

[0040] ab / / cd and(len(ab)≤l max or len(bc)≤l max )

[0041] Among them, l max The maximum length of the segment translation needs to be met. len(ab) is the Euclidean distance between turning points a and b, and len(bc) is the Euclidean distance between turning points b and c. The translation operation can be performed on ab and bc.

[0042] The steps of the parallelogram turning point optimization method considering risk cost are as follows:

[0043] D1: Calculate the initial planned path segment set segment_path_ini = {s1, s2, ..., s n}, s n For the segments included in the planned path; create a new path segment set segment_path_new = {};

[0044] D2: Determine the number of segments in segment_path_ini. If the number of segments is greater than 3, proceed to step D3; otherwise, add the segment in segment_path_ini to segment_path_new and proceed to step D7.

[0045] D3: Find the first set of consecutive segments ab, bc, and cd from route_path. If ab / / cd is satisfied, proceed to step D4; if not, proceed to step D5.

[0046] D4: Get the path grid point set corresponding to the initial segments ab and bc ini , the path grid point set corresponding to the parallel transformation segments ab1 and b1c trans . For set ini and set trans To calculate the risk, the formula is:

[0047]

[0048] Where M is the number of grids in the path grid point set. If grid ρ is a prohibited flying grid, then Risk seg = +∞. Then, the risk of the initial segment and the changed segment can be calculated according to the formula seg_ini and Risk seg_trans , proceed to step D5;

[0049] D5: If Risk seg_ini ≥Risk seg_trans , then perform parallelogram transformation to add ab1 to segment_path_new; if not satisfied, go to step D6;

[0050] D6: Add ab to segment_path_new and remove it from segment_path_ini, and repeat step D2;

[0051] D7: Conversion ends. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 The figure is a flow chart of a UAV path planning method based on comprehensive risk cost according to the present invention.

[0053] Figure 2 It is the path translation map of the UAV, where (a) is the initial path map and (b) is the path map after translation. DETAILED DESCRIPTION

[0054] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0055] Figure 1 This is a flowchart of a drone path planning method based on comprehensive risk cost according to the present invention. The present invention constructs a drone path planning model, combines airspace risk cost and transportation cost, optimizes turning point settings, reduces the number of turning points, reduces flight risks, and improves the safety and efficiency of drone flight.

[0056] The main steps include:

[0057] S1: Construct a UAV path planning model, including the objective function and constraints. The objective function is composed of the airspace risk cost and the UAV transportation cost. The specific method is:

[0058] Layer h q There is a directed graph G(N,E) in the grid, where N is h q Layer grid point set, E represents the edge set. When planning the low-altitude UAV route, the center coordinate of the starting grid S is (x s ,y s ,h q ), the center coordinate of the end grid E is (x e ,y e ,h q ), the center coordinate of the path grid t in route planning is (x t ,y t ,h q ), t represents the grid number of the route, and t=1,2,...,n-1,n, n is the total number of grids to which the route belongs. Its objective function mainly includes the airspace risk cost C risk_cost and drone transportation cost C trans_cost , which can be expressed as:

[0059]

[0060] Where: P t-1,t is the transportation operation cost from grid t-1 to adjacent grid t, and the formula is:

[0061] P t-1,t =l t-1,t ×f e ×τ(m cargo )

[0062] Among them:t-1,t is the length from grid i to the adjacent grid t; f e The cost paid for the unit energy consumption of the drone; τ(m cargo ) is the load m cargo The penalty coefficient when , which can be expressed as:

[0063]

[0064] Where: τ max is the maximum cargo loading penalty value, m max The maximum cargo weight.

[0065] In summary, the comprehensive objective function of the UAV path planning model can be expressed as:

[0066] minC cost =(C risk_cost +1)×C trans_cost

[0067] S2: Set constraints, including range constraints, turning angle constraints, and maximum takeoff weight constraints, to ensure that the UAV meets the range, turning angle, and payload limits during flight. The specific method is:

[0068] P1: The calculation formula for range constraint is:

[0069] d aircraft_min ≥len(r SE )≤d aircraft_max

[0070] Among them, the constraint is the flight range limit constraint of the UAV, and the planned route r SE The distance len(r SE ) cannot be less than the minimum range d of the UAV aircraft_min , nor can it be greater than its maximum range d aircraft_max ;

[0071] P2: The calculation formula for the turning angle constraint is:

[0072] 0≤δ j ≤δ max

[0073] Among them, this constraint is the maximum steering angle limit of the UAV, indicating that the UAV

[0074] Steering angle δ j The maximum steering angle δ that it can withstand must not be exceeded max ;

[0075] P3: Maximum Take-Off Weight The calculation formula for the maximum take-off weight constraint is:

[0076] m uav +m cargo <m max

[0077] Among them, this constraint is for the maximum take-off mass of the UAV, indicating that the UAV's own weight m uav and load m cargo The sum of the weights must be less than the maximum take-off weight of the drone m max .

[0078] S3: Initial path planning method based on weight-A* algorithm, calculates the path planning cost function and generates the initial planning path; the specific method is:

[0079] The path planning cost function f(t) is:

[0080] f(t)=g(t)+h(t)

[0081] Where g(t) is the actual cost of moving from the starting grid s to the expanded grid t; h(t) is the estimated cost of moving from grid t to the end grid e. g(t) and h(t) are the actual cost and heuristic function of the Parallel-A* algorithm, respectively, and the formulas are:

[0082]

[0083] in, and x is the actual environmental operation risk cost and route planning risk cost generated from the starting point S to the extended grid corresponding to the extended route path; e and y e is the grid coordinate of grid e; x t and y t is the grid coordinate of grid t.

[0084] S4: Based on the parallelogram turning point optimization method combined with risk calculation, the initial planned path is optimized to reduce the number of turning points and reduce flight risks. The specific method is as follows:

[0085] like Figure 2 As shown in (a), it is easy to know that the segment ab / / cd, which means that the segment ab can be transformed into a parallelogram. When ab or bc satisfies:

[0086] ab / / cd and(len(ab)≤l max or len(bc)≤l max )

[0087] Among them, l max The maximum length of the segment translation needs to be met. len(ab) is the Euclidean distance between turning points a and b, and len(bc) is the Euclidean distance between turning points b and c. The translation operation can be performed on ab and bc as follows: Figure 2 (b) The steps of the parallelogram turning point optimization method considering risk cost are as follows:

[0088] D1: Calculate the initial planned path segment set segment_path_ini = {s1, s2, ..., s n}, s n For the segments included in the planned path; create a new path segment set segment_path_new = {};

[0089] D2: Determine the number of segments in segment_path_ini. If the number of segments is greater than 3, proceed to step D3; otherwise, add the segment in segment_path_ini to segment_path_new and proceed to step D7.

[0090] D3: Find the first set of consecutive segments ab, bc, and cd from route_path. If ab / / cd is satisfied, proceed to step D4; if not, proceed to step D5.

[0091] D4: Get the path grid point set corresponding to the initial segments ab and bc ini , the path grid point set corresponding to the parallel transformation segments ab1 and b1c trans . For set ini and set trans To calculate the risk, the formula is:

[0092]

[0093] Where M is the number of grids in the path grid point set. If grid ρ is a prohibited flying grid, then Risk seg = +∞. Then, the risk of the initial segment and the changed segment can be calculated according to the formula seg_ini and Risk seg_trans , proceed to step D5;

[0094] D5: If Risk seg_ini ≥Risk seg_trans , then perform parallelogram transformation to add ab1 to segment_path_new; if not satisfied, go to step D6;

[0095] D6: Add ab to segment_path_new and remove it from segment_path_ini, and repeat step D2;

[0096] D7: Conversion ends.

Claims

1. A UAV path planning method based on comprehensive risk cost, characterized in that: The method comprises the following steps: S1: Construct a UAV path planning model that includes comprehensive risks and operating costs, including an objective function and constraints. The objective function is composed of the airspace risk cost and the UAV transportation cost. S2: Set three-dimensional constraints including range constraint, turning angle constraint, and maximum takeoff weight constraint to construct a feasible solution space that conforms to the physical characteristics of the UAV; S3: Initial path planning is performed based on the weighted heuristic search weight-A* algorithm. By dynamically adjusting the heuristic function weights to balance path length and risk cost, an initial feasible path that meets the constraints is generated. S4: Based on the parallelogram turning point optimization method, the turning points of the initial planned path are optimized.

2. The UAV path planning method based on comprehensive risk cost according to claim 1 is characterized in that: The specific method of step S1 is: Layer h q There is a directed graph G(N,E) in the grid, where N is h q Layer grid point set, E represents the edge set; When planning the low-altitude UAV route, the center coordinates of the starting grid S are (x s ,y s ,h q ), the center coordinate of the end grid E is (x e ,y e ,h q ), the center coordinate of the path grid t in route planning is (x t ,y t ,h q ), t represents the grid number of the route, and t=1,2,...,n-1,n, n is the total number of grids to which the route belongs; Its objective function includes the airspace risk cost C risk_cost and drone transportation cost C trans_cost , expressed as: Where: P t-1,t is the transportation operation cost from grid t-1 to adjacent grid t, and the formula is: P t-1,t =l t-1,t ×f e ×τ(m cargo ) Among them: t-1,t is the length from grid i to the adjacent grid t; f e The cost paid for the unit energy consumption of the drone; τ(m cargo ) is the load m cargo The penalty coefficient when , is expressed as: Where: τ max is the maximum cargo loading penalty value, m max is the maximum cargo weight; In summary, the comprehensive objective function of the UAV path planning model is expressed as: minC cost =(C risk_cost +1)×C trans_cost 。 3. The UAV path planning method based on comprehensive risk cost according to claim 1 is characterized in that: The specific steps in step S2 are: P1: The calculation formula for range constraint is: d aircraft_min ≥len(r SE )≤d aircraft_max Among them, the range constraint is the flight range limit constraint of the UAV, and the planned route r SE The distance len(r SE ) cannot be less than the minimum range d of the UAV aircraft_min , nor can it be greater than its maximum range d aircraft_max ; P2: The calculation formula for the turning angle constraint is: 0≤δ j ≤δ max Among them, the turning angle constraint is the limit of the maximum turning angle of the UAV, which means that during the path planning process, the turning angle δ of the UAV j The maximum steering angle δ that it can withstand must not be exceeded max ; P3: The maximum takeoff weight constraint is calculated as follows: m uav +m cargo <m max Among them, the maximum take-off weight constraint is the maximum take-off mass of the UAV, indicating that the UAV's own weight m uav and load m cargo The sum of the weights must be less than the maximum take-off weight of the drone m max .

4. The low-altitude route network planning method based on dynamic and static risk collaborative assessment according to claim 1 is characterized in that: The specific method of step S3 is: The path planning cost function f(t) is: f(t)=g(t)+h(t) Where g(t) is the actual cost of moving from the starting grid s to the expanded grid t; h(t) is the estimated cost of moving from grid t to the end grid e; g(t) and h(t) are the actual cost and heuristic function of the Parallel-A* algorithm, respectively, and the formulas are: in, and x is the actual environmental operation risk cost and route planning risk cost generated from the starting point S to the extended grid corresponding to the extended route path; e and y e is the grid coordinate of grid e; x t and y t is the grid coordinate of grid t.

5. The low-altitude route network planning method based on dynamic and static risk collaborative assessment according to claim 1 is characterized in that: In step S4: If segment ab / / cd, segment ab can be transformed into a parallelogram; when ab or bc satisfies: ab / / cd and(len(ab)≤l max or len(bc)≤l max ) Among them, l max The maximum length of the segment translation needs to be met. len(ab) is the Euclidean distance between turning points a and b, and len(bc) is the Euclidean distance between turning points b and c. That is, ab and bc can be translated. The steps of the parallelogram turning point optimization method are as follows: D1: Calculate the initial planned path segment set segment_path_ini = {s1, s2, ..., s n }, s n For the segments included in the planned path; create a new path segment set segment_path_new = {}; D2: Determine the number of segments in segment_path_ini. If the number of segments is greater than 3, proceed to step D3; otherwise, add the segment in segment_path_ini to segment_path_new and proceed to step D7. D3: Find the first set of consecutive segments ab, bc, and cd from route_path. If ab / / cd is satisfied, proceed to step D4; if not, proceed to step D5. D4: Get the path grid point set corresponding to the initial segments ab and bc ini , the path grid point set corresponding to the parallel transformation segments ab1 and b1c trans ; for set ini and set trans To calculate the risk, the formula is: Where M is the number of grids in the path grid point set; if grid ρ is a prohibited flying grid, then Risk seg =+∞; then calculate the risk of the initial segment and the changed segment according to the formula seg_ini and Risk seg_trans , proceed to step D5; D5: If Risk seg_ini ≥Risk seg_trans , then perform parallelogram transformation to add ab1 to segment_path_new; if not satisfied, go to step D6; D6: Add ab to segment_path_new and remove it from segment_path_ini, and repeat step D2; D7: Conversion ends.

Images