CN114924593B - Quick planning method for vehicle and multi-unmanned aerial vehicle combined route - Google Patents
Quick planning method for vehicle and multi-unmanned aerial vehicle combined route Download PDFInfo
- Publication number
- CN114924593B CN114924593B CN202210681994.2A CN202210681994A CN114924593B CN 114924593 B CN114924593 B CN 114924593B CN 202210681994 A CN202210681994 A CN 202210681994A CN 114924593 B CN114924593 B CN 114924593B
- Authority
- CN
- China
- Prior art keywords
- convergence
- solution
- unmanned aerial
- vehicle
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 152
- 230000008569 process Effects 0.000 claims abstract description 58
- 230000009191 jumping Effects 0.000 claims abstract description 12
- 238000005457 optimization Methods 0.000 claims description 18
- 239000002245 particle Substances 0.000 claims description 15
- 238000013459 approach Methods 0.000 claims description 2
- 238000013178 mathematical model Methods 0.000 description 16
- 230000006870 function Effects 0.000 description 15
- 238000004364 calculation method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000005265 energy consumption Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000014509 gene expression Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000012804 iterative process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000010391 action planning Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
- G05D1/104—Simultaneous control of position or course in three dimensions specially adapted for aircraft involving a plurality of aircrafts, e.g. formation flying
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The invention discloses a rapid planning method for a combined route of a vehicle and a plurality of unmanned aerial vehicles, belongs to the field of planning and control of a multi-robot system, and comprises a multi-unmanned aerial vehicle-vehicle combined route planning problem model under the constraint of a plurality of time windows and a heuristic planning model with a two-layer structure. The two-layer structure heuristic planning model has the following characteristics: 1) The method comprises two layers of iteration, wherein a first layer optimizes the combination of convergence sequence and convergence position, and a second layer respectively optimizes the convergence sequence and convergence position; 2) Heuristic strategies containing a state of jumping out of local optimum can attempt to jump out of the solution process out of local optimum by setting a convergence order; 3) Including heuristic methods that generate an initial solution of "convergence order-convergence location". Aiming at real-time configuration of a vehicle and an unmanned aerial vehicle (initial positions of the vehicle and the unmanned aerial vehicle, residual continuous energy of the unmanned aerial vehicle, an actionable time window and the like), the method can rapidly calculate a group of convergence route solutions which can enable the unmanned aerial vehicle to safely return to the vehicle.
Description
Technical Field
The invention belongs to the field of planning and control of a multi-robot system, and particularly relates to a rapid planning method for a combined route of a vehicle and a plurality of unmanned opportunities.
Background
The multi-unmanned plane-vehicle cooperative system (multi-Drones and Truck Cooperative System, mDTCS) is composed of a vehicle and a plurality of unmanned planes mounted on the vehicle. mDTCS are executed in the following modes: the vehicle moves in the scene environment, flexibly combines with the unmanned aerial vehicle, and timely supplies energy for the unmanned aerial vehicle or carries the unmanned aerial vehicle to move for a long distance. Therefore, mDTCS can break through the limit of the limited endurance energy of the unmanned aerial vehicle on the range, and is an ideal execution mode based on a large-range area task of multiple unmanned aerial vehicles. Therefore, mDTCS has received a great deal of attention in recent years. Many institutions and companies, such as amazon, ***, DHL, and courseware, have developed a great deal of discussion, research, and experimentation for their use in multi-objective inspection scenarios and multi-facility deployment scenarios such as last mile cargo delivery, power facility inspection, and the like.
Route planning is the key to the successful application of mDTCS. Only when the route planning of the vehicle and the unmanned aerial vehicle is matched with the scene situation, mDTCS can ensure that all unmanned aerial vehicles can safely take off and land, the number of access target points of the unmanned aerial vehicle in one take-off and landing can be further increased, the take-off and landing times of the unmanned aerial vehicle are reduced, and therefore the task execution time and cost are effectively reduced. However, during the execution of the task mDTCS, unpredictable scene dynamics often occur, such as: the unmanned aerial vehicle energy consumption rate is too fast, unmanned aerial vehicle processing target point task progress lag etc.. These dynamic changes can deviate the scene situation from the original plan, and if the unmanned aerial vehicle performing the task still acts according to the original plan, the unmanned aerial vehicle may crash because it cannot meet the vehicle before the endurance energy is exhausted. At this time, a feasible meeting route of the unmanned aerial vehicle and the vehicle needs to be planned in time, so that all unmanned aerial vehicles performing tasks can be ensured to safely return to the vehicle, and a feasible and conservative reference scheme of recovering all unmanned aerial vehicles first and then considering a subsequent execution route is provided for subsequent action planning of mDTCS.
The convergence of the drone with the vehicle is limited by the drone's actionable time window. The opening time of the movable time window is the time when the unmanned plane finishes the processing task of the target point currently being executed and can move to the meeting place; the closing time is the time when the continuous energy of the unmanned aerial vehicle is exhausted; the time period from the opening time to the closing time is the opening time period of the actionable time window. The drone must be in the same location as the vehicle during the open time period of its actionable window to account for the successful convergence. The actionable time window of the drone introduces complex constraints to the vehicle-multiple drone joint route planning problem: 1) All unmanned aerial vehicles must meet the vehicle before the respective actionable time window closes; 2) All unmanned aerial vehicles can move only after the opening time of each movable time window; 3) If the opening time of the movable time window is not reached, even if the unmanned aerial vehicle and the vehicle are at the same position, the unmanned aerial vehicle does not count as a convergence. A viable solution to the vehicle-to-multi-drone joint routing problem must meet all of the constraints described above.
In recent years, the problem of mDTCS of scheduling and route planning has been paid attention to, but there is no research on the problem of vehicle-to-multi-unmanned-opportunity joint route planning. In mDTCS scheduling and route planning problems, although the route planning results of the scenes of multi-objective patrol and the like are similar to the unmanned aerial vehicle-vehicle convergence route, a plurality of planning models and solving methods are also proposed, but because of the following particularities of the vehicle-multi-unmanned aerial vehicle convergence route planning problems, the planning models and the solving methods cannot be directly applied to the convergence route planning of the multi-unmanned aerial vehicle-vehicle, and the main reasons are as follows: (1) The initial scene condition of the vehicle-multi-unmanned opportunity joint route planning problem is different from other scenes; (2) The planning target of the vehicle-multi-unmanned aerial vehicle combined route planning problem is to find a feasible solution as soon as possible so as to ensure that all unmanned aerial vehicles can safely return to the vehicle, rather than solving a group of optimal routes as in the scenes of multi-target patrol and the like; (3) Compared with the time window constraint which only considers the moment of reaching the target in the scenes such as multi-target patrol, the available time window constraint of all unmanned aerial vehicles in the vehicle-multi-unmanned aerial vehicle combined route planning problem has more limit on the feasible solution, and the occupation ratio of the feasible solution in the solution space is greatly reduced. Before a viable solution is found, the vehicle-multi-drone joint route planning problem evaluates the solution for its merits based on the number of drones that meet over time and the total timeout time, which introduces a significant local optimization phenomenon in the solution space. For other scene applications, the existing method is easy to trap into local optimum of the infeasible solution, and the feasible solution cannot be obtained.
Therefore, the research on the novel method for rapidly planning the combined route of the vehicle and the multiple unmanned opportunities has important application value.
Disclosure of Invention
Aiming at the defects or improvement demands of the prior art, the invention provides a rapid planning method for a combined route of a vehicle and multiple unmanned aerial vehicles, and aims to rapidly solve a group of solution of a combined route which enables the unmanned aerial vehicles to safely return to the vehicle aiming at real-time configuration of the vehicle and the unmanned aerial vehicles. The method comprises a multi-unmanned aerial vehicle-vehicle convergence route planning problem model under the constraint of a multi-time window and a heuristic planning model with a two-layer structure, and can improve the speed and accuracy of route planning.
In order to achieve the purpose of the invention, the invention adopts a two-layer iterative structure so as to reduce the iterative times. In the first layer iteration, an optimal combination of a convergence sequence Q-convergence position P r is obtained; in the second layer iteration, the HPSO algorithm (mixed particle swarm optimization algorithm) and the PSO algorithm (particle swarm optimization algorithm) are adopted to optimize Q and P r respectively. The method specifically comprises the following steps:
(1) Initializing a first layer iteration number recorder i to be 0, and generating initial 'convergence sequence-convergence position' solutions Q 0 and P r0 by adopting a heuristic initial solution generation method;
(2) Starting the first layer iteration of the i+1th time, and updating a first layer iteration number recorder i=i+1; checking whether the first layer iteration solving process is in local optimum, if so, executing the step (3), otherwise, executing the step (4);
(3) Setting a convergence sequence solution Q i of the ith first layer iteration by adopting a heuristic strategy which is out of a local optimal state, correspondingly adopting a method for setting a convergence position in a heuristic initial solution generating method, modifying a convergence position solution P r(i-1) of the ith-1 first layer iteration, and then jumping to the step (5);
(4) Optimizing and solving a second layer iteration of a convergence sequence solution Q i of the ith first layer iteration, generating a convergence sequence solution Q 'r(i-1) by adopting a heuristic initial solution generating method, randomly generating a convergence sequence solution Q' r(i-1), forming an initial particle position of a HPSO algorithm together with a convergence sequence solution Q r(i-1) of the ith-1 first layer iteration, optimizing and solving a convergence sequence solution Q u by adopting a HPSO algorithm according to a convergence position solution P r(i-1) of the ith-1 first layer iteration;
(5) Optimizing and solving a second layer iteration of a convergence position solution P i of the ith first layer iteration, generating a convergence position solution P 'r(i-1) by adopting a heuristic initial solution generating method, randomly generating a convergence position solution P' r(i-1), forming an initial particle position of a PSO algorithm together with the convergence position solution P r(i-1) of the ith-1 first layer iteration, optimizing and solving a convergence position solution P ri by adopting the PSO algorithm according to a convergence sequence solution Q i of the ith first layer iteration;
(6) Updating the optimal "convergence order-convergence location" solutions Q b and P rb for the above solution process and keeping Q i and P i,Qb and P rb for this first layer iteration to the solution result logger;
(7) It is checked whether the solving process satisfies the exit iteration condition,
If not, returning to the step (2), and starting the (i+1) th first layer iteration;
If so, outputs Q b and P rb as a planning solution to the multiple unmanned aerial vehicle-vehicle convergence routing problem under the multiple time window constraint.
Further, in the step (2), the method for checking whether the first layer iterative solving process falls into the local optimum is as follows: a sample window for checking local optimum is introduced to observe a solution result recorder, and if the solutions Q i and P ri in the sample window remain unchanged, the first layer iterative solution process is considered to be trapped in local optimum. The method comprises the following steps: if the current first layer iteration number i meets i not less than i j_1st and the solution Q i and P ri remain unchanged from the (i-w j_1st) th iteration to the (i) th iteration, the first layer iteration solving process is considered to be in the local optimum.
Further, the heuristic strategy of jumping out of the local optimum state in the step (3) specifically comprises: first, the convergence time T ri of all the drones is analyzed using the above constraint expression according to the convergence order and convergence position solution { Q i,Pri } updated in the last iteration. If there is a rendezvous timeout drone, i.e., { Q i,Pri } is an infeasible solution, then the rendezvous timeout drone is selected and its serial number U to and rendezvous serial number N to are obtained. According to U to、Nto、Tri and Q i, a new convergence sequence Q new is generated by a convergence sequence reset method of infeasible solutions that has not been recorded in the solution result recorder. If none of the drones has a convergence timeout, i.e., { Q i,Pri } is a feasible solution, a new convergence order Q new is generated using the convergence order reset method of the feasible solution.
Further, the method for checking whether the iterative solution process falls into the local optimum in the step (2) and the heuristic strategy for jumping out of the local optimum in the step (3) are both used for optimizing the second-layer iterative process for solving the convergence sequence solution Q i of the ith first-layer iteration in the step (4) so as to avoid the convergence sequence solution process falling into the local optimum.
Further, the convergence sequence resetting method of the infeasible solution is specifically as follows, firstly, a convergence timeout unmanned aerial vehicle u tox is randomly selected, and a convergence sequence number n tox is obtained. If the selected unmanned aerial vehicle u tox is the unmanned aerial vehicle which meets the first frame and the vehicle, randomly rearranging the meeting sequence solution Q i to generate a new meeting sequence Q new; otherwise, a drone u ins with a convergence time before the closing time of the actionable time window of the drone u tox is randomly selected, the ordering of the drone serial numbers in the convergence order solution Q i is adjusted, u tox is inserted into the next position of u ins, or u tox is placed in the first position of the convergence order solution, thereby generating a new convergence order Q new.
Further, the convergence sequence resetting method of the feasible solution is as follows:
1) Calculating the convergence idle time of each unmanned aerial vehicle, and randomly selecting an unmanned aerial vehicle which is not in the first position of the convergence sequence Q i by adopting a roulette manner And obtain its serial number n c in meeting order Q i, in this process, meet the unmanned aerial vehicle that idle duration is shorter and more easily selected.
2) Unmanned aerial vehicle is carried out in meeting sequence Q i Is shifted forward, a new candidate sequence Q t is formed for each shift, and the unmanned aerial vehicle/>, in candidate sequence Q t And the convergence number of (2) is m. According to the candidate sequence Q t, analyzing the convergence of all unmanned aerial vehicles, if the following two conditions are satisfied: a) The vehicle does not need to wait for the drone/>, at the meeting pointB) Unmanned planeIf the convergence time of (1) does not exceed the closing time of the movable time window, the candidate sequence Q t is considered to be valid, and the candidate sequence Q t is stored in the candidate convergence sequence set/>Is a kind of medium.
3) If the convergence order is to be selectedIf not empty, the following criteria A) and B) are comprehensively considered, and the roulette method is adopted to collect/> from the alternative meeting sequenceA candidate sequence Q t is randomly selected as the new convergence sequence Q new. Specifically, the criterion a) is: the shorter the distance from the point P rtm to the line segment [ P rt(m-1),Prt(m+1) ], the easier the candidate sequence Q t is selected; the criterion B) is: the easier the candidate sequence Q t satisfying the following inequality is selected. D (P1, P2) is the distance from the point P1 to the point P2, and D is the same as the following.
The inequality described above uses a representation of parameter/variable nesting, such as: p rt(m-1) represents the convergence position of the unmanned aerial vehicle with the convergence number (m-1) in the candidate sequence Q t; Represents the convergence position of the unmanned aerial vehicle with the convergence number n c in the candidate sequence Q t, and other parameters have the same meaning,
The left side of the above inequality represents an unmoved droneA route length corresponding to the pre-ranking convergence order Q i of (a), wherein each represents a route corresponding to the convergence order Q i with the unmanned aerial vehicle/>, respectivelyThe length of the associated route segment; the right side of the above inequality represents mobile drone/>The route length corresponding to the candidate sequence Q t formed after ranking, wherein each item respectively represents the route corresponding to the candidate sequence Q t and the unmanned aerial vehicle/>The length of the associated route segment. The above criterion B) represents: compared to the convergence order Q i, the candidate order Q t that makes the route length shorter is more easily selected,
4) If the order set is converged due to candidateIs empty, without choosing a new convergence order Q new, and the number of attempts is less than the maximum allowed number j max, return to step 1) for a new attempt. If the number of attempts has reached the maximum allowable number j max, a new rendezvous order Q new has not yet been selected, then the rendezvous order solution Q i is randomly rearranged, generating a new rendezvous order Q new.
Further, the heuristic initial solution generating method in the steps (1), (3), (4) and (5) is as follows: specifically, first, an initial position of the vehicle is set as a current convergence position; then, for each unmanned aerial vehicle which does not meet the vehicle, calculating the time for the vehicle to meet the unmanned aerial vehicle from the current meeting position, selecting the unmanned aerial vehicle with the smallest meeting time as the unmanned aerial vehicle with the next meeting, calculating the corresponding meeting position as the next meeting position, and taking the next meeting position as the current meeting position; next, the previous step is repeated until a complete convergence order and convergence location is obtained.
Further, the method for checking whether the solving process meets the exit iteration condition in the step (7) is as follows: and introducing a sample window for checking the iteration exit condition to observe a solving result recorder, and considering that the solving process meets the exit iteration condition if the optimal solutions Q b and P rb in the sample window are kept unchanged or the executed first layer iteration number reaches the maximum allowable number. The method comprises the following steps: setting the opening time and the width of a sample window for checking iteration exit conditions as i e_1st and w e_1st respectively, and if the current first layer iteration number i meets i not less than i e_1st and the optimal solution Q b and P rb are unchanged from the (i-w e_1st) th iteration to the i th iteration, then considering that the solving process meets the exit iteration conditions; or the current first layer iteration number i reaches the maximum allowed execution number i max_1st, and the solving process can be considered to meet the exit iteration condition.
Further, in the steps (4) and (5), a method similar to the method in the step (7) for exiting the iteration condition and checking whether the solving process meets the exiting iteration condition is also adopted, so that the optimization iteration times are reduced on the premise of not affecting the optimization result.
In general, the above technical solutions conceived by the present invention have the following beneficial effects compared with the prior art:
1. The invention is oriented to the scenes of a large-scale regional task such as multi-target patrol, multi-facility arrangement and the like of a multi-unmanned aerial vehicle-vehicle cooperative system, and aims at the situation that the original planning fails due to the dynamic change of the scene, rapidly solves the feasible route capable of safely recovering all the on-the-fly unmanned aerial vehicles, and provides a conservation scheme of recovering all unmanned aerial vehicles first and then considering the follow-up execution route for the re-planning of the follow-up actions of the cooperative system.
2. The heuristic strategy for jumping out of the local optimal state and the heuristic method for generating the initial solution of the convergence sequence and the convergence position designed by the invention improve the probability and the calculation speed for obtaining the feasible solution in the multi-unmanned plane-vehicle convergence route planning problem. The multi-drone-vehicle junction route planning problem under multi-window constraints has the locally optimal nature of a distinct "junction sequence-junction location" combination: in the solution space there are many sets of combinations of "convergence order-convergence locations", which have fitness extrema within each larger neighborhood, and where most of the "convergence order-convergence locations" combinations are not feasible solutions. The existing route planning method aiming at other scenes is very easy to fall into a locally optimal infeasible solution in the iterative solving process, and is difficult to separate from the neighborhood of the infeasible solution, so that the feasible solution is difficult to find in a short time. This problem becomes more serious as the number of unmanned aerial vehicles increases. The heuristic strategy designed by the method can try to jump out the neighborhood of the local optimal solution by setting a convergence sequence solution mode when the iterative solution process falls into the local optimal state, thereby improving the probability of obtaining a feasible solution in the wired iteration times. The heuristic method for generating the initial solution of the convergence sequence-convergence position can lead the optimization process of the convergence sequence-convergence position combination and the solution process of the convergence sequence and the convergence position to start from the neighborhood of the optimal solution, thereby reducing the iteration times required by obtaining the feasible solution and improving the calculation speed.
Drawings
Fig. 1 is a schematic diagram of a heuristic planning flow of a two-layer structure for a method for fast planning a vehicle and a multi-unmanned opportunity approach according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an example of a scenario in which a 10-frame drone meets a vehicle in one embodiment of the present invention;
fig. 3 is a junction route planning result of a scene example of a junction of a certain 10 unmanned aerial vehicle and a vehicle in one embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the invention, the multi-unmanned aerial vehicle-vehicle convergence route planning problem model under the constraint of the multi-time window defines the characteristics of the multi-unmanned aerial vehicle-vehicle convergence route planning problem, defines the application range of the invention, and is a preparation link of the invention; the planning model details the solving process of the multi-unmanned aerial vehicle-vehicle convergence route planning problem, contains the core content of the invention and is an execution link of the invention.
1. Mathematical model
The mathematical model of the multi-unmanned plane-vehicle convergence route planning problem under the constraint of multiple time windows comprises two parts of problem definition and mathematical modeling:
1.1 problem definition
The problem definition of the multiple unmanned aerial vehicle-vehicle convergence route planning problem under the constraint of multiple time windows includes: scene features, initial conditions, simplifying assumptions, planning objectives, planning solutions, known conditions.
The scene features are as follows: starting from the scene initial moment, the vehicle moves in the scene area and goes to the corresponding convergence point according to the designated convergence sequence; any unmanned aerial vehicle moves from the opening moment of the movable time window to the appointed meeting point so as to meet the vehicle at the meeting point before the closing moment of the movable time window. Both the vehicle and the drone may stay at the meeting point waiting for the other to arrive. When the vehicle is successfully integrated with one unmanned aerial vehicle, the vehicle moves to the meeting point of the next unmanned aerial vehicle. When all the unmanned aerial vehicles successfully meet the vehicle, the scene is ended. The vehicle can run without time limit in the scene area; any unmanned aerial vehicle can only move within a movable time window and must meet the vehicle within the time window, otherwise the unmanned aerial vehicle crashes due to energy consumption.
The initial conditions were as follows: at time Jing Chushi, the drones are distributed in the scene area at different locations. The remaining endurance energy of each unmanned aerial vehicle is different, namely the closing time of the movable time window is different. Part of the unmanned aerial vehicle may need to complete the task currently being performed before the moving body meets the vehicle, so the opening time of the movable time window of the unmanned aerial vehicle may be different.
The simplifying assumption is as follows:
1) The positions of the unmanned plane and the vehicle at the initial moment are known;
2) Unmanned aerial vehicles and vehicles may move unimpeded in a scene area;
3) The vehicle and the unmanned aerial vehicle move at constant speed, and the flying speeds of all unmanned aerial vehicles are the same;
4) The landing time of the unmanned aerial vehicle when the unmanned aerial vehicle is engaged with the vehicle is ignored;
5) The continuous energy consumption of the unmanned aerial vehicle is only related to the flight time, and the energy consumption rate of the unmanned aerial vehicle in unit time is consistent and constant;
6) A successful convergence is considered when the drone is co-located with the vehicle within the moveable time window.
The planning solution includes: the order in which each unmanned aerial vehicle meets the vehicle, and the meeting position of each unmanned aerial vehicle. Wherein the feasible solution is required to enable all unmanned aerial vehicles to meet the vehicle within the open time period of the respective actionable time window.
The planning targets are as follows: and finding out a group of convergence sequences and solutions of convergence positions as soon as possible, so that the convergence time of each unmanned aerial vehicle is within the movable time window of the unmanned aerial vehicle.
Known conditions include: the number of unmanned aerial vehicles, the initial position, the actionable time window, the speed of flight, and the initial position and speed of travel of the vehicle.
1.2 Mathematical modeling
Mathematical modeling of a multiple unmanned aerial vehicle-vehicle convergence route planning problem under multiple time window constraints includes: parameters and variables are defined, an objective function is planned, and constraint conditions are met.
Parameters and variables are defined in the following table. The parameters correspond to known conditions in the problem definition, including formal representations of the known conditions, whose values remain unchanged during the problem solving process. The variables comprise decision variables and auxiliary variables, the values of which change in the solving process, wherein the decision variables correspond to planning solutions in the problem definition and comprise formal expressions of convergence sequences and convergence positions; the auxiliary variables then mainly include observables of the planned solution states and performance.
Parameter table parameter and variable definitions
The planning objective function is represented by formula (1), and the constraint conditions include formulas (2) to (13).
S.t.
tc=trn (7)
tr0=0 (8)
Pr0=Pt (9)
The above formula adopts the representation form of parameter/variable nesting, such as: p r(i-1) represents the convergence position of the unmanned aerial vehicle where the (i-1) th frame meets the vehicle; t s(qi) denotes the opening time of the actionable time window of the unmanned aerial vehicle (q i)) with sequence number (q i); t rn represents the convergence time of the nth unmanned aerial vehicle with the vehicle, and the meaning of the similar parameters and the like.
Constraints (2) and (3) ensure the correctness of the convergence order and convergence position: in the convergence sequence, each drone can only converge with a truck once, and all convergence locations must be within the scene area. Constraint conditions (4) - (6) prescribe a method for calculating the unmanned aerial vehicle convergence time. Constraint (4) specifies that the vehicle or drone that arrives at the convergence point first must stay to wait for the other party to arrive before the convergence is completed. Constraint (5) specifies that the drone does not leave its initial position until the moment at which its actionable time window is open; when the time is after the moment when its actionable time window is open, the drone will fly directly from its initial location to the rendezvous location. The constraint (6) specifies that the convergence route of the vehicle starts from its initial position and reaches convergence positions one by one according to the convergence order. In constraint (7), the completion time of all unmanned aerial vehicle convergence tasks is defined as the convergence time of the last unmanned aerial vehicle. Constraint (8) and (9) are supplemented with two parameters t r0 and P r0 and assigned with 0 and vehicle initial position P t, respectively, to ensure proper calculation of equation (5). Finally, constraint conditions (10) - (13) define a binary variable value method related to the planning objective function and determined by the number of unmanned aerial vehicles with overtime convergence.
The planning objective function consists of two parts, corresponding to the two phases where the solution process finds or does not find a feasible solution, and is switched by a 1 and a 2, respectively. When the solution process has not been feasible, i.e., there is one or more unmanned aerial vehicles with overtime convergence, there is a 1=0,A2 =1. In this case, the latter half of the planning objective function works, the objective function value being determined by the number of rendezvous timeout drones and the average and maximum of the timeout times and always being greater than 1. If the solution process finds a feasible solution, i.e. no unmanned aerial vehicle with a convergence timeout, there is a 1=1,A2 =0. In this case, the first half of the objective function works, and the objective function value is determined by the completion time of the task of all the unmanned aerial vehicles, and the convergence idle time of each unmanned aerial vehicle (the interval time between the convergence time of the unmanned aerial vehicle and the closing time of the movable time window thereof), and is always smaller than 1. The latter half of the planning objective function aims to find a feasible solution as soon as possible, while the former half aims to optimize the found feasible solution, so as to increase the convergence idle time of each unmanned aerial vehicle and reduce the completion time of all unmanned aerial vehicle convergence tasks.
2. Planning model
The heuristic planning method of the two-layer structure is used for solving the multi-unmanned aerial vehicle-vehicle convergence route planning problem under the constraint of a plurality of time windows defined by the problem definition of the mathematical model, and the parameter and variable definition of the mathematical model, the planning objective function and the constraint condition are applied to the method. In the heuristic planning method of the two-layer structure, as shown in fig. 1, fig. 1 is a schematic diagram of a heuristic planning flow of the two-layer structure for a rapid planning method of a vehicle and a multi-unmanned aerial vehicle joint route provided by the embodiment of the invention, and as can be seen from the figure, a two-layer iterative structure is adopted to reduce the number of iterations, and the heuristic method comprises a heuristic strategy of jumping out of a local optimal state and a heuristic method of generating an initial solution of a convergence sequence-convergence position, so that the success rate and the calculation speed of obtaining a feasible solution in a limited number of iterations are improved. Specifically, in a first layer iteration, an optimal combination of a convergence sequence Q-convergence position P r is obtained; in the second layer iteration, the HPSO algorithm (mixed particle swarm optimization algorithm) and the PSO algorithm (particle swarm optimization algorithm) are adopted to optimize Q and P r respectively. In the first tier of iterations, a combination of "convergence order Q-convergence position P r" is optimized, and in the second tier of iterations, second tier of iterations of optimizing convergence order Q, convergence position P r, respectively, is nested.
The planning method comprises the following steps:
(1) The first layer iteration number i=0 is initialized and an initial "convergence order-convergence location" solution Q 0 and P r0 are generated using a heuristic initial solution generation method.
(2) Starting the first layer iteration of the ith+1th time, and updating the first layer iteration times i=i+1; checking whether the first layer iteration solving process is in a local optimal state, if so, executing the step (3), otherwise, executing the step (4).
(3) Setting a convergence sequence solution Q i of the ith first layer iteration by adopting a heuristic strategy which is out of a local optimal state, and correspondingly modifying a convergence position solution P r(i-1) of the ith-1 first layer iteration by adopting a convergence position setting method in a heuristic initial solution generating method. And then jumps to step (5).
(4) This step is the optimization of the second-tier iteration that finds the convergence order solution Q i for the ith first-tier iteration. A heuristic initial solution generation method is adopted to generate a convergence sequence solution Q ' r(i-1), a convergence sequence solution Q ' r(i-1) is randomly generated, and the convergence sequence solution Q ' r(i-1) and the convergence sequence solution Q r(i-1) of the i-1 th first layer iteration form an initial solution set for optimizing the convergence sequence solution Q i. The convergence order solution Q i is optimized based on the convergence position solution P r(i-1) of the i-1 st first layer iteration and the initial solution set of the optimization convergence order solution Q i. In the embodiment, a HPSO algorithm is adopted to optimize and solve a convergence sequence solution Q i, wherein an initial solution set of the convergence sequence solution Q i is optimized as an initial particle position of a HPSO algorithm; other algorithms, such as generalized particle swarm optimization, may also be used to accomplish the optimization of the convergence order solution Q i.
(5) This step is a second layer iteration that optimizes the solution P i for the convergence position of the ith first layer iteration. A heuristic initial solution generation method is adopted to generate a convergence position solution P 'r(i-1), a convergence position solution P' r(i-1) is randomly generated, and the convergence position solution P r(i-1) of the i-1 th first layer iteration form an initial solution set of an optimized convergence position solution P i. Optimizing the solution P i of the convergence position according to the convergence sequence solution Q i obtained in the step (4) and the initial solution set of the solution P i of the convergence position. In the embodiment, a Particle Swarm Optimization (PSO) algorithm is adopted to optimize and solve a convergence position solution P i, wherein an initial solution set of the convergence position solution P i is optimized as an initial particle position of the particle swarm algorithm; other algorithms, such as genetic algorithms, ant colony algorithms, etc., may also be employed to accomplish the optimization of the junction location solution P i at this step.
(6) Using the planning objective function and constraint conditions of the mathematical model, calculating a corresponding objective function value F i according to the solution combinations Q i and P i of the 'convergence sequence-convergence position' of the ith first layer iteration; the corresponding objective function values F b are calculated from the optimal "convergence order-convergence location" solution combinations Q b and P rb for the algorithm solving process, and if F i<Fb, the optimal "convergence order-convergence location" solutions Q b and P rb,Qb=Qi,Prb=Pri for the algorithm solving process are updated if the i-th first-layer iteration "convergence order-convergence location" solution combinations Q i and P i are considered to be superior to the optimal "convergence order-convergence location" solution combinations Q b and P rb for the algorithm solving process. And holds the i-th first layer iteration result Q i,Pi, and the optimal solutions Q b and P rb to the solution result logger.
(7) Checking whether the solving process meets the exit iteration condition, if not, returning to the step (2), and starting the (i+1) th first layer iteration.
(8) The optimal "convergence order-convergence location" solutions Q b and P rb for the algorithm solution process are output as the planning solutions for the multiple unmanned aerial vehicle-vehicle convergence route planning problem under the constraint of multiple time windows.
Further, in the step (2), the method for checking whether the first layer iterative solving process falls into the local optimum is as follows: a sample window for checking local optimum is introduced to observe the change condition of the planning solutions Q i and P ri in the solution result recorder, and if the planning solutions Q i and P ri in the sample window remain unchanged, the first layer iterative solution process is considered to be in local optimum. The method comprises the following steps: if the current first layer iteration number i meets i not less than i j_1st and the planning solutions Q i and P ri in the solution result recorder are unchanged from the (i-w j_1st) th iteration to the i th iteration, the first layer iteration solution process is considered to be in the local optimum.
Further, the heuristic strategy of the step (3) of jumping out of the local optimum state is specifically shown in table 1: first, according to the convergence order and convergence position solution { Q i,Pri } updated in the last iteration, the convergence time T ri of all the unmanned aerial vehicles is calculated using the constraint conditions of the mathematical model described above. If there is a rendezvous timeout drone, i.e., { Q i,Pri } is not a viable solution, then pick up the rendezvous timeout drone and acquire its serial number U to and rendezvous serial number N to. According to U to、Nto、Tri and Q i, a new convergence sequence Q new is generated by a convergence sequence reset method of infeasible solutions that has not been recorded in the solution result recorder. If none of the drones has a convergence timeout, i.e., { Q i,Pri } is a feasible solution, a new convergence order Q new is generated using the convergence order reset method of the feasible solution.
TABLE 1 heuristic strategy for jumping out of local optimum
Further, a method similar to the method for checking whether the iterative solution process falls into the local optimum in the step (2) and the heuristic strategy for jumping out of the local optimum in the step (3) is also used for optimizing the second layer iterative process for solving the convergence sequence solution Q i in the step (4) so as to avoid the convergence sequence solution process falling into the local optimum.
Further, the convergence order resetting method of the infeasible solution is shown in table 2, first, a convergence timeout unmanned plane u tox is randomly selected, and a convergence sequence number n tox of the unmanned plane in the convergence order solution Q i is obtained. If the selected unmanned aerial vehicle u tox is the unmanned aerial vehicle which meets the first frame and the vehicle, randomly rearranging the meeting sequence solution Q i to generate a new meeting sequence Q new; otherwise, a drone u ins with a convergence time before the closing time of the actionable time window of the drone u tox is randomly selected, the ordering of the drone serial numbers in the convergence order solution Q i is adjusted, u tox is inserted into the next position of u ins, or u tox is placed in the first position of the convergence order solution, thereby generating a new convergence order Q new.
Table 2 generate_new_sequence_NFFS (N to,Uto,Qi,Tri)
Further, the convergence order resetting method of the feasible solution is shown in table 3, and the steps are as follows:
1) Calculating the convergence idle time length of each unmanned aerial vehicle according to the constraint conditions in the mathematical model, and randomly selecting an unmanned aerial vehicle which is not in the first position of the convergence sequence Q i by adopting a roulette mode And obtain its serial number n c in meeting order Q i, in this process, meet the unmanned aerial vehicle that idle duration is shorter and more easily selected.
2) Unmanned aerial vehicle is carried out in meeting sequence Q i Is shifted forward, a new candidate sequence Q t is formed for each shift, and the unmanned aerial vehicle/>, in candidate sequence Q t And the convergence number of (2) is m. According to the candidate sequence Q t, calculating the convergence condition of all unmanned aerial vehicles according to the constraint conditions in the mathematical model, if the following two conditions are satisfied: a) The vehicle does not need to wait for the drone/>, at the meeting pointB) Unmanned plane/>If the convergence time of (1) does not exceed the closing time of the movable time window, the candidate sequence Q t is considered to be valid, and the candidate sequence Q t is stored in the candidate convergence sequence set/>Is a kind of medium.
3) If the convergence order is to be selectedIf not empty, the following criteria A) and B) are comprehensively considered, and the roulette method is adopted to collect/> from the alternative meeting sequenceA candidate sequence Q t is randomly selected as the new convergence sequence Q new. Specifically, the criterion a) is: the shorter the distance from the point P rtm to the line segment [ P rt(m-1),Prt(m+1) ], the easier the candidate sequence Q t is selected; the criterion B) is: the easier the candidate sequence Q t satisfying the following inequality is selected. /(I)
The inequality described above uses a representation of parameter/variable nesting, such as: p rt(m-1) represents the convergence position of the unmanned aerial vehicle with the convergence number (m-1) in the candidate sequence Q t; The convergence position of the unmanned aerial vehicle with the convergence number n c in the candidate sequence Q t is shown. And so on.
The left side of the above inequality represents an unmoved droneA route length corresponding to the pre-ranking convergence order Q i of (a), wherein each represents a route corresponding to the convergence order Q i with the unmanned aerial vehicle/>, respectivelyThe length of the associated route segment; the right side of the above inequality represents mobile drone/>The route length corresponding to the candidate sequence Q t formed after ranking, wherein each item respectively represents the route corresponding to the candidate sequence Q t and the unmanned aerial vehicle/>The length of the associated route segment. The above criterion B) represents: the candidate sequence Q t that makes the route length shorter is more easily selected than the convergence sequence Q i.
4) If the order set is converged due to candidateIs empty, without choosing a new convergence order Q new, and the number of attempts is less than the maximum allowed number j max, return to step 1) for a new attempt. If the number of attempts has reached the maximum allowable number j max, a new rendezvous order Q new has not yet been selected, then the rendezvous order solution Q i is randomly rearranged, generating a new rendezvous order Q new.
Table 3 generate_new_sequence_FFS (Q i,Pri)
/>
Further, the heuristic initial solution generating method in the steps (1), (3), (4) and (5) is shown in table 4. The initial position of the vehicle is first set to the current convergence position. Then, calculating the time for the vehicle to meet the unmanned aerial vehicle from the current meeting position for each unmanned aerial vehicle which does not meet the vehicle; selecting the minimum convergence time as the unmanned aerial vehicle converged next, and calculating the corresponding convergence position as the convergence position next; the next rendezvous position is taken as the current rendezvous position. Next, the previous step is repeated until a complete convergence order and convergence location is obtained.
TABLE 4 Generation of initial Convergence sequences and positions
Further, the method for checking whether the solving process meets the exit iteration condition in the step (7) is as follows: and introducing a sample window for checking the iteration exit condition to observe a solving result recorder, and considering that the solving process meets the exit iteration condition if the optimal solutions Q b and P rb in the sample window are kept unchanged or the executed first layer iteration number reaches the maximum allowable number. The method comprises the following steps: setting the opening time and the width of a sample window for checking iteration exit conditions as i e_1st and w e_1st respectively, and if the current first layer iteration number i meets i not less than i e_1st and the optimal solutions Q b and P rb recorded in a solution result recorder are unchanged from the (i-w e_1st) th iteration to the i th iteration, considering that the solution process meets the exit iteration conditions; or the current first layer iteration number i reaches the maximum allowed execution number i max_1st, and the solving process can be considered to meet the exit iteration condition.
Further, in the steps (4) and (5), a method similar to the method in the step (7) for exiting the iteration condition and checking whether the solving process meets the exiting iteration condition is also adopted, so that the optimization iteration times are reduced on the premise of not affecting the optimization result.
Further, in the steps (1), (3), (4), (5), and the heuristic strategy and heuristic initial solution generating method for jumping out of the local optimal state, the solution values of the convergence sequence and convergence position are limited according to the constraint conditions (2) and (3) of the mathematical model, so as to ensure the correctness of the planned solution; in the steps (3), (4), (5) and (6), and in the heuristic strategy and the heuristic initial solution generation method which jump out of the local optimal state, constraint conditions (4) to (13) of the mathematical model are adopted, and the convergence condition of all unmanned aerial vehicles corresponding to the planning solution is analyzed, including calculating convergence time of all unmanned aerial vehicles and counting the number of unmanned aerial vehicles with overtime convergence; in the steps (4), (5) and (6), the objective function value of the planning solution is calculated by adopting the planning objective function of the mathematical model so as to evaluate the merits of the planning solution.
Fig. 2 is a schematic diagram of an example of a scenario where a certain 10 unmanned aerial vehicles meet vehicles in an embodiment of the present invention, and taking a modeling and solving process of a meeting route planning problem of a meeting scenario of a certain multi-unmanned aerial vehicle (specifically, 10 unmanned aerial vehicles) and vehicles shown in fig. 2 as an example, a specific application process of a vehicle and multi-unmanned aerial vehicle meeting route rapid planning method of the present invention is described. In fig. 2, the vehicle is free to move in position, and 10 drones are dispersed throughout the vehicle.
The process of modeling and solving the convergence route planning problem of the convergence scene of the multi-unmanned aerial vehicle and the vehicle by using the method is divided into two stages of preparation and solving. In the preparation stage, the problem definition of the mathematical model is contrasted, the elements of the problem to be planned are combed, and whether the problem to be planned is suitable for solving by adopting the method is analyzed and judged; if the present invention can be used for solving, the input parameters required for using the present invention are prepared according to the above mathematical model. In the solving stage, the parameters prepared in the preparation stage are input into the multi-unmanned aerial vehicle-vehicle convergence route planning model, and a feasible solution of the multi-unmanned aerial vehicle-vehicle convergence route is solved.
The process of carrying out the convergence route planning problem of the convergence scene of the multiple unmanned aerial vehicles and the vehicles by using the method of the invention is summarized as follows:
Step (1): in the preparation stage, a convergence scene of the multiple unmanned aerial vehicles and the vehicles and a convergence route planning problem are checked, and whether the convergence route planning problem completely accords with the problem definition of the mathematical model or not is judged. If the two paths completely coincide, the convergence route planning problem of the convergence scene of the multiple unmanned aerial vehicles and the vehicles can be modeled and solved by using the method.
The convergence scene of the multiple unmanned aerial vehicles and the convergence route planning problem thereof completely accord with the problem definition of the mathematical model, so that the following requirements are met:
1) The scene characteristics and the initial conditions of the multi-unmanned aerial vehicle and vehicle meeting scene comprise all scene characteristics and initial conditions defined by the problems;
2) The multi-unmanned aerial vehicle and vehicle meeting scene can carry out all simplified assumptions of the problem definition;
3) The planning targets and the planning solutions of the convergence route planning problem of the convergence scene of the multiple unmanned aerial vehicles and the vehicles are consistent with the planning targets and the planning solutions defined by the problem;
4) The known conditions defined by the above-mentioned problem are all known and remain unchanged in the case of the convergence route planning problem of the convergence scenario of the multiple unmanned aerial vehicle with the vehicle.
Step (2): in the preparation stage, if the convergence route planning problem of the convergence scene of the unmanned aerial vehicle and the vehicle can be solved by adopting the invention, the parameters required by the invention are prepared to be used according to the parameter definition in the mathematical modeling of the mathematical model. The parameters to be prepared are shown in table 1, and the user can refer to the assignment examples in the table to design the data structure and assign values to each parameter.
Table 1 parameter definition of mathematical modeling of multiple drone-vehicle convergence route planning problem under multiple time window constraints
/>
Step (3): to solve the phase. Inputting the parameters prepared in the step (2) into the multi-unmanned aerial vehicle-vehicle convergence route planning model, and solving a planning solution of the multi-unmanned aerial vehicle-vehicle convergence route, wherein the planning solution comprises a convergence sequence and a convergence position, and can also know convergence conditions of each unmanned aerial vehicle corresponding to the planning solution, including convergence time and convergence overtime conditions of each unmanned aerial vehicle, as shown in fig. 3. Fig. 3 is a schematic view showing a result of a convergence route of a scene example where 10 unmanned aerial vehicles converge with a vehicle according to an embodiment of the present invention, where it is known that, under the conditions that initial positions of the unmanned aerial vehicles and the vehicles, moveable time windows of the unmanned aerial vehicles and the like are known, the convergence sequence of the vehicles and the unmanned aerial vehicles, and the convergence positions of the corresponding unmanned aerial vehicles can be obtained by using the present invention, so as to form a convergence route of the vehicles and the unmanned aerial vehicles, and ensure that each unmanned aerial vehicle can converge with the vehicle within the moveable time window.
It will be readily appreciated by those skilled in the art that the foregoing description is merely a preferred embodiment of the invention and is not intended to limit the invention, but any modifications, equivalents, improvements or alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (7)
1. A rapid planning method for a vehicle and multi-unmanned aerial vehicle joint route is characterized in that a two-layer iteration structure is adopted, in a first layer iteration, an optimized combination of a meeting sequence Q-meeting position P r is obtained, in a second layer iteration, Q and P r are respectively and independently optimized,
The method comprises the following steps:
(1) Initializing a first layer iteration number recorder i to be 0, and generating initial 'convergence sequence-convergence position' solutions Q 0 and P r0 by adopting a heuristic initial solution generation method;
(2) Starting the first layer iteration of the i+1th time, and updating a first layer iteration number recorder i=i+1; checking whether the first layer iteration solving process is in local optimum, if so, executing the step (3), otherwise, executing the step (4);
(3) Setting a convergence sequence solution Q i of the ith first layer iteration by adopting a heuristic strategy which is out of a local optimal state, correspondingly adopting a method for setting a convergence position in a heuristic initial solution generating method, modifying a convergence position solution P r(i-1) of the ith-1 first layer iteration, and then jumping to the step (5);
(4) Optimizing and solving a second layer iteration of a convergence sequence solution Q i of the ith first layer iteration, generating a convergence sequence solution Q 'r(i-1) by adopting a heuristic initial solution generating method, randomly generating a convergence sequence solution Q' r(i-1), forming initial particle positions of a mixed particle swarm optimization algorithm together with a convergence sequence solution Q r(i-1) of the ith-1 first layer iteration, and optimizing and solving a convergence sequence solution Q i by adopting a HPSO algorithm according to a convergence position solution P r(i-1) of the ith-1 first layer iteration;
(5) Optimizing and solving a second layer iteration of a convergence position solution P i of the ith first layer iteration, generating a convergence position solution P 'r(i-1) by adopting a heuristic initial solution generating method, randomly generating a convergence position solution P' r(i-1), forming initial particle positions of a particle swarm optimization algorithm together with the convergence position solution P r(i-1) of the ith-1 first layer iteration, optimizing and solving a convergence position solution P ri by adopting a PSO algorithm according to a convergence sequence solution Q i of the ith first layer iteration;
(6) Updating the optimal "convergence order-convergence location" solutions Q b and P rb for the above solution process and keeping Q i and P i,Qb and P rb for this first layer iteration to the solution result logger;
(7) It is checked whether the solving process satisfies the exit iteration condition,
If not, returning to the step (2), and starting the (i+1) th first layer iteration;
If so, Q b and P rb are output as a planning solution to the vehicle and multiple unmanned opportunities joint route planning problem under the constraint of multiple time windows.
2. The method for quickly planning a combined path of a vehicle and multiple unmanned aerial vehicles according to claim 1, wherein the method for checking whether the first layer iterative solution process falls into a local optimum in step (2) is as follows: introducing a sample window for checking local optimum for observing a solution result recorder, if the solutions Q i and P ri in the sample window are kept unchanged, considering that the first layer iterative solution process falls into the local optimum,
Specifically, the opening time and width of a sample window for checking local optimum are i j_1st and w j_1st respectively, if the current iteration number i of the first layer meets i not less than i j_1st and the solution Q i and P ri remain unchanged from the (i-w j_1st) th iteration to the i th iteration, the first layer iteration solving process is considered to be in local optimum.
3. The method for rapid planning of a vehicle-to-multi-unmanned opportunity approach of claim 2, wherein the heuristic strategy of the step (3) for jumping out of the locally optimal state is:
First, according to the convergence order updated in the last iteration and the convergence position solution { Q i,Pri }, the convergence time T ri of all the unmanned aerial vehicles is analyzed, if there is a convergence timeout unmanned aerial vehicle, { Q i,Pri } is an infeasible solution, the convergence timeout unmanned aerial vehicle is selected, and the sequence number U to and the convergence number N to thereof are acquired, and according to U to、Nto、Tri and Q i, a new convergence order Q new which was not recorded in the solution result recorder is generated by the convergence order reset method of the infeasible solution, and if all the unmanned aerial vehicles are not convergence timeout, { Q i,Pri } is a viable solution, a new convergence order Q new is generated by using the convergence order reset method of the viable solution.
4. A method for rapid planning of a vehicle-to-multi-unmanned joint line of claim 3, wherein the convergence sequence resetting method of the infeasible solution is as follows:
Firstly, randomly selecting a convergence timeout unmanned aerial vehicle u tox, obtaining a convergence sequence number n tox of the convergence timeout unmanned aerial vehicle u tox, and randomly rearranging a convergence sequence solution Q i to generate a new convergence sequence Q new if the selected unmanned aerial vehicle u tox is the unmanned aerial vehicle converged with the vehicle; otherwise, a drone u ins with a convergence time before the closing time of the actionable time window of the drone u tox is randomly selected, the ordering of the drone serial numbers in the convergence order solution Q i is adjusted, u tox is inserted into the next position of u ins, or u tox is placed in the first position of the convergence order solution, thereby generating a new convergence order Q new.
5. A method for rapid planning of a vehicle-to-multi-unmanned joint line of road according to claim 3, wherein the convergence sequence resetting method of the feasible solution comprises the following sub-steps:
1) Calculating the convergence idle time of each unmanned aerial vehicle, and randomly selecting an unmanned aerial vehicle which is not in the first position of the convergence sequence Q i by adopting a roulette manner And obtains the sequence number n c of the unmanned aerial vehicle in the convergence sequence Q i, the unmanned aerial vehicle with shorter convergence idle time length is easier to be selected in the process,
2) Unmanned aerial vehicle is carried out in meeting sequence Q i Is shifted forward, a new candidate sequence Q t is formed for each shift, and the unmanned aerial vehicle/>, in candidate sequence Q t According to the candidate sequence Q t, analyzing the convergence of all unmanned aerial vehicles if the following two conditions are satisfied: a) The vehicle does not need to wait for the drone/>, at the meeting pointB) Unmanned plane/>If the convergence time of (1) does not exceed the closing time of the movable time window, the candidate sequence Q t is considered to be valid, and the candidate sequence Q t is stored in the candidate convergence sequence set/>In,
3) If the convergence order is to be selectedIf not empty, the following criterion A) and criterion B) are comprehensively considered, and the roulette method is adopted to collect/> from the alternative meeting sequenceOne candidate sequence Q t is randomly selected as the new rendezvous sequence Q new,
Specifically, the criterion a) is: the shorter the distance from the point P rtm to the line segment [ P rt(m-1),Prt(m+1) ], the easier the candidate sequence Q t is selected; the criterion B) is: the more easily the candidate sequence Q t satisfying the following inequality is selected, D (P1, P2) is the distance from the point P1 to the point P2, D is the same as the following meaning,
The left side of the above inequality represents an unmoved droneA route length corresponding to the pre-ranking convergence order Q i of (a), wherein each represents a route corresponding to the convergence order Q i with the unmanned aerial vehicle/>, respectivelyThe length of the associated route segment; the right side of the above inequality represents mobile drone/>The route length corresponding to the candidate sequence Q t formed after ranking, wherein each item respectively represents the route corresponding to the candidate sequence Q t and the unmanned aerial vehicle/>The length of the relevant route segment, criterion B) above represents: compared to the convergence order Q i, the candidate order Q t that makes the route length shorter is more easily selected,
4) If the convergence order is to be selectedIf the number of attempts is less than the maximum allowable number j max, returning to step 1) to perform a new attempt, if the number of attempts has reached the maximum allowable number j max, and the new convergence sequence Q new has not been selected, then randomly rearranging the convergence sequence solution Q i to generate a new convergence sequence Q new.
6. A method for rapid planning of a combined route of a vehicle and multiple unmanned opportunities as defined in any one of claims 1-5, wherein the heuristic initial solution generating method comprises, first, setting an initial position of the vehicle as a current convergence position; then, for each unmanned aerial vehicle which does not meet the vehicle, calculating the time for the vehicle to meet the unmanned aerial vehicle from the current meeting position, selecting the unmanned aerial vehicle with the shortest meeting time as the unmanned aerial vehicle with the next meeting, calculating the corresponding meeting position as the next meeting position, and taking the next meeting position as the current meeting position; next, the previous step is repeated until a complete convergence order and convergence location is obtained.
7. The method for quickly planning a vehicle-to-multiple unmanned aerial vehicle joint route according to claim 6, wherein the method for checking whether the solving process satisfies the exit iteration condition in step (7) is as follows:
A sample window for checking the iteration exit condition is introduced to observe a solution result recorder, if the optimal solution Q b and the optimal solution P rb of the solution process in the sample window are kept unchanged or the number of executed first layer iterations reaches the maximum allowable number, the solution process is considered to meet the exit iteration condition,
The method comprises the following steps: setting the opening time and the width of a sample window for checking iteration exit conditions as i e_1st and w e_1st respectively, and if the current first layer iteration number i meets i not less than i e_1st and the optimal solution Q b and P rb are unchanged from the (i-w e_1st) th iteration to the i th iteration, then considering that the solving process meets the exit iteration conditions; or the current first layer iteration number i reaches the maximum allowed execution number i max_1st, and the solving process is considered to meet the exit iteration condition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210681994.2A CN114924593B (en) | 2022-06-16 | 2022-06-16 | Quick planning method for vehicle and multi-unmanned aerial vehicle combined route |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210681994.2A CN114924593B (en) | 2022-06-16 | 2022-06-16 | Quick planning method for vehicle and multi-unmanned aerial vehicle combined route |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114924593A CN114924593A (en) | 2022-08-19 |
CN114924593B true CN114924593B (en) | 2024-04-26 |
Family
ID=82815334
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210681994.2A Active CN114924593B (en) | 2022-06-16 | 2022-06-16 | Quick planning method for vehicle and multi-unmanned aerial vehicle combined route |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114924593B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
LU102400A1 (en) * | 2019-08-06 | 2021-02-09 | Nanjing Seawolf Ocean Tech Co Ltd | Path planning method and system for unmanned surface vehicle based on improved genetic algorithm |
AU2020104224A4 (en) * | 2020-10-30 | 2021-03-11 | Guangdong University Of Technology | Indoor Collaborative Path Planning Method and System for Multi-UAVs in High-rise Fire-fighting Tasks |
CN113848970A (en) * | 2021-08-31 | 2021-12-28 | 中南大学 | Multi-target collaborative path planning method for vehicle and unmanned aerial vehicle |
WO2022057700A1 (en) * | 2020-09-18 | 2022-03-24 | 武汉理工大学 | Coverage route planning method for multiple unmanned surface mapping vessels |
-
2022
- 2022-06-16 CN CN202210681994.2A patent/CN114924593B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
LU102400A1 (en) * | 2019-08-06 | 2021-02-09 | Nanjing Seawolf Ocean Tech Co Ltd | Path planning method and system for unmanned surface vehicle based on improved genetic algorithm |
WO2022057700A1 (en) * | 2020-09-18 | 2022-03-24 | 武汉理工大学 | Coverage route planning method for multiple unmanned surface mapping vessels |
AU2020104224A4 (en) * | 2020-10-30 | 2021-03-11 | Guangdong University Of Technology | Indoor Collaborative Path Planning Method and System for Multi-UAVs in High-rise Fire-fighting Tasks |
CN113848970A (en) * | 2021-08-31 | 2021-12-28 | 中南大学 | Multi-target collaborative path planning method for vehicle and unmanned aerial vehicle |
Non-Patent Citations (3)
Title |
---|
Route Optimization Method for Unmanned Air Vehicle Launched from a Carrier;Savuran 等;Lecture Notes on Software Engineering;20151231;全文 * |
基于动态规划的无人机协同侦察路线规划;童玮;李业军;沈焕生;郑翔;蒋海霞;;军事通信技术;20160925(第03期);全文 * |
车辆-多无人机协同巡检的路径规划与调度算法研究;刘卫东;中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑 (月刊);20220515;全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN114924593A (en) | 2022-08-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Liu et al. | An adaptive large neighborhood search metaheuristic for agile satellite scheduling with time-dependent transition time | |
CN111024080B (en) | Unmanned aerial vehicle group-to-multi-mobile time-sensitive target reconnaissance path planning method | |
CN113848970B (en) | Multi-target cooperative path planning method for vehicle-unmanned aerial vehicle | |
CN110852554B (en) | Unmanned aerial vehicle task allocation intelligent decision-making method and device under cooperation of vehicle machine | |
CN113641192A (en) | Route planning method for unmanned aerial vehicle crowd sensing task based on reinforcement learning | |
CN114281104A (en) | Multi-unmanned-aerial-vehicle cooperative regulation and control method based on improved ant colony algorithm | |
CN115016537B (en) | Heterogeneous unmanned aerial vehicle configuration and task planning combined optimization method in SEAD scene | |
Shahzaad et al. | Robust composition of drone delivery services under uncertainty | |
CN116451934A (en) | Multi-unmanned aerial vehicle edge calculation path optimization and dependent task scheduling optimization method and system | |
Geißer et al. | Optimal and heuristic approaches for constrained flight planning under weather uncertainty | |
CN114924593B (en) | Quick planning method for vehicle and multi-unmanned aerial vehicle combined route | |
CN114595871A (en) | Multi-unmanned aerial vehicle task planning method | |
CN113283827A (en) | Two-stage unmanned aerial vehicle logistics path planning method based on deep reinforcement learning | |
CN113687651A (en) | Path planning method and device for delivering vehicles according to needs | |
Zhu et al. | The Path Optimization Algorithm of Car Navigation System considering Node Attributes under Time‐Invariant Network | |
CN116415480B (en) | IPSO-based off-road planning method for aircraft offshore platform | |
Liang et al. | Route planning of truck and multi-drone rendezvous with available time window constraints of drones | |
Xiao et al. | MACNS: A generic graph neural network integrated deep reinforcement learning based multi-agent collaborative navigation system for dynamic trajectory planning | |
CN116859919A (en) | Multi-AGV path planning method based on deep reinforcement learning and cyclic neural network | |
CN116629479A (en) | AGV multi-target traffic control method and system for single-way collision-free AGV | |
Xie et al. | Hybrid AI-based Dynamic Re-routing Method for Dense Low-Altitude Air Traffic Operations | |
Shi et al. | Collaborative Planning of Parking Spaces and AGVs Path for Smart Indoor Parking System | |
WO2021252683A1 (en) | Systems and methods for controlling automated systems using integer programming and column generation techniques | |
CN113487236A (en) | Airplane scheduling method based on genetic algorithm | |
Xiang et al. | Enhanced pigeon inspired optimisation approach for Agile Earth observation satellite scheduling |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |