CN103699135B - The flight path automatic planning in depopulated helicopter pesticide spraying farmland operation region - Google Patents

Abstract

The invention discloses a kind of flight path automatic planning of depopulated helicopter pesticide spraying farmland operation region, the local optimum of planning arbitrary initial vertex, polygon farmland operation region, the local optimum S type trajectory planning of arbitrary original heading and arbitrary initial vertex, arbitrary original heading " returns " font flight path, " returns " font flight path find out global optimum S type flight path respectively and font flight path " returns " in global optimum from local optimum S type trajectory planning and local optimum.The present invention overcomes the deficiency of artificial planning flight path in Traditional Man operating type, make depopulated helicopter under the prerequisite meeting job task, various cost consumption factor can be reduced, improve spraying efficiency.

Description

Automatic planning method for flight path of unmanned helicopter pesticide spraying farmland operation area

Technical Field

The invention relates to the field of unmanned helicopter flight path planning, in particular to an automatic flight path planning method for an unmanned helicopter pesticide spraying farmland operation area.

Background

Under the vigorous push of the department of agriculture, the mechanization level of China in the aspects of ploughing, sowing, harvesting and the like is remarkably improved in recent years, but pesticide spraying (particularly rice pesticide spraying) is basically the traditional manual operation. China is a big agricultural country, how to effectively prevent agricultural pests becomes one of important targets of agricultural production in China, particularly, in the process of vigorously advocating and popularizing green agriculture and precision agriculture in China, low-cost, precise and high-environment-friendly pesticide spraying mechanization and automation which are suitable for the current situation of rural areas in China become an indispensable technology, and pesticide spraying by using a small unmanned helicopter is the best choice of pesticide spraying mechanization.

Under the current rural conditions in China, the method for spraying pesticide by using the small unmanned helicopter is a practical and feasible method in China, particularly in the southern region. The sand agricultural department in 2012 has bought 50 unmanned plant protection planes for pesticide spraying.

The unmanned pesticide spraying helicopter is high in speed, ultra-low-capacity pesticide spraying is used, pesticide and water resources are saved, pesticide residues and environmental pollution of crops are reduced, and remote operation can also reduce harm to pesticide applying personnel. The device is suitable for various terrains, accords with the current situation of rural roads in our city, and can realize cross-regional operation by matching with a table-board trolley.

At present, civil unmanned helicopters are in a rapid development stage in China, the utilization rate of the civil unmanned helicopters is higher and higher, and particularly in the field of agriculture, the operation efficiency can be greatly improved by using the helicopters. However, the unmanned helicopter has a long controllable distance, so that the specific flight state, such as flight direction and flight distance, cannot be judged by human eyes. The lack of a method for reasonably and effectively planning the flight path of the unmanned helicopter in a farmland operation area causes that the flight path selected by an operator is not optimal; and the spraying efficiency is reduced and the operation cost is increased due to the missed spraying, the mistaken spraying and the repeated spraying caused by the visual error.

During the task completion process of the unmanned aerial vehicle, how to effectively and safely complete the task process of the unmanned aerial vehicle needs to be planned, which is called task planning. In the task planning process, the most important and complex is to plan a flight path required for completing a flight task for the unmanned aerial vehicle, namely the unmanned aerial vehicle path planning.

Unmanned aerial vehicle flight path planning is to comprehensively consider various factors, such as: the arrival time, the flight distance, the fuel consumption, the threat, the flight area and the like, an optimal or most satisfactory flight path is planned for the unmanned aerial vehicle, so that the flight mission can be completed satisfactorily.

The unmanned helicopter flight path planning has various methods, such as an A-star search method, a Voronoi graph algorithm, a genetic algorithm, an ant colony algorithm, a particle swarm optimization algorithm, a heuristic search and the like. The algorithm A is a classical optimal heuristic search algorithm, is generally used for solving the static planning problem and has wide application in path planning and graph search. The algorithm guides search through heuristic information, and achieves the purposes of reducing the search range and improving the calculation speed. When a conventional a-x algorithm is used to perform a track search, a planning environment is usually represented in a grid form, and then a minimum-cost track is found according to a predetermined cost function. The method calculates the cost of each possible grid unit of the current position, and then selects the grid unit with the lowest cost to be added into the search space for exploration. This new grid cell added to the search space is in turn used to generate more possible paths. Voronoi diagrams are an important geometry in computer geometry. McLain and Beard et al propose a collaborative multi-aircraft flight path planning method based on Voronoi diagrams. Firstly, a Voronoi diagram is constructed through known local radar or threats, the boundaries of the Voronoi diagram are all flyable tracks, then the weight values of the boundaries are given, and finally the optimal tracks are searched. The genetic algorithm is a calculation model of a biological evolution process for simulating natural selection and genetic mechanism of Darwinian biological evolution theory, and is a method for searching an optimal solution by simulating the natural evolution process. The general steps for planning a flight path using a genetic algorithm are: a) encoding the flight path; b) constructing a proper track evaluation function; c) selecting a genetic operator suitable for flight path planning; d) operators are calculated and fine tuned to obtain the final solution. The ant colony algorithm realizes path search through information exchange and mutual cooperation of ants, and has good universality and robustness. The process of searching the path by the ant colony algorithm comprises the following steps: a) initializing pheromones at all nodes on a flight area diagram to form an initial pheromone matrix; b) m ants are positioned at the starting point A to wait for starting; c) each ant selects the next point on the grid graph according to the state transition rule and finally reaches the target point to form a feasible route; d) calculating an objective function of a feasible route of each ant, and storing an optimal route solution; e) adjusting the pheromone of each point according to the target function and the pheromone adjusting criterion; f) checking the optimal solution, judging whether to adjust the information evaporation prime factor P, and if so, correspondingly adjusting according to a certain rule; g) judging whether an iteration condition is met (namely whether a set iteration number or a minimum objective function is met), and if so, finishing the search; if not, returning to the step b), and repeatedly executing until the iteration condition is met. Heuristic search is that the search in the state space evaluates each searched position to get the best position, and then searches from this position to the target. Therefore, a large number of unnecessary search paths can be omitted, and the efficiency is improved. In heuristic search, the valuation of the location is very important. Different valuations may be used with different results.

Traditional heuristic search and other path search methods have respective advantages in processing the shortest reachable path and the optimal flight path (short time, low oil consumption and high safety), avoiding obstacles and the like, but are not suitable for flight path planning in farmland operation. This is due to the specificity of the field work. The main points are as follows:

a) the flight path of an unmanned helicopter in farm operations must first cover the entire farm operation area, which is different from the conventional shortest or optimal flight path.

b) On the basis of covering the whole operation area, the flight path and the flight mode in the whole area are further considered, namely, how to fly can enable the operation efficiency to be the highest.

c) In the flight process, oil consumption and residual pesticide amount are also considered, when the oil consumption or pesticide amount is insufficient, the operation is continued after return voyage addition, and the distance from the base station to the base station also needs to be calculated in the whole planning method.

The invention aims to apply the traditional heuristic search algorithm to the farmland operation track planning of the unmanned helicopter.

The terms used in the present invention are explained as follows:

planning a flight path: the aircraft can meet flight tasks and meet flight trajectories of constraint conditions.

An unmanned helicopter: an unmanned helicopter.

S-type flight path: the unmanned helicopter flies along a preset course to spray pesticides, flies for a certain distance at the rear side after reaching a boundary point, and flies in the direction opposite to the original course, so that a curved S-shaped flight path is formed.

The 'Hui' shape flight path: the unmanned helicopter flies along a preset course to spray pesticide, hovers to turn after reaching a turning point, and continuously flies forwards along the next course, thereby forming a 'return' flight track.

A base station: the base station is a basic workstation, and when fuel oil or pesticide of the unmanned helicopter is insufficient, the base station can return to the base station to add fuel oil and pesticide to the unmanned helicopter and then start to continue operation.

Disclosure of Invention

The invention aims to solve the technical problem that the prior art is insufficient, and provides an automatic flight path planning method for a farmland operation area sprayed with pesticide by an unmanned helicopter.

In order to solve the technical problems, the technical scheme adopted by the invention is as follows: an automatic planning method for flight paths of an unmanned helicopter pesticide spraying farmland operation area comprises the following steps:

1) digitizing the polygonal farmland operation area by using a grid method, and dividing the polygonal farmland operation area into a plurality of grids; the number of vertexes of the polygonal farmland operation area is n;

2) establishing a node index matrix for the digitized polygonal farmland operation area, taking each grid as a node, and storing the terrain information of each grid into the node index matrix;

3) defining the following cost function h (n):

$h\left(n\right)=\left\{\begin{array}{cc}1,& \stackrel{\→}{v}\∈I\\ 2,& \stackrel{\→}{v}\∈J\\ 3,& \stackrel{\→}{v}\∈K\end{array}\right.;$

wherein,a velocity vector for the unmanned helicopter to fly from the current node to the next target node; i is a direct flight course vector set of the unmanned helicopter; j is a set of unmanned helicopter lateral flight direction vectors; k is an unmanned helicopter oblique flight vector set;

4) carrying out unmanned helicopter track search by utilizing an heuristic memory method to obtain a local S-shaped optimal flight reference track of any initial course of all initial vertexes of a polygonal farmland operation area and a local 'returning' character-shaped optimal flight reference track of any initial course of all initial vertexes of the polygonal farmland operation area, wherein the local S-shaped optimal flight reference track and the local 'returning' character-shaped optimal flight reference track are respectively 2 n;

5) respectively solving the cost values of the 2n local S-shaped optimal flight reference tracks and the cost values of the 2n local 'back' shaped optimal flight reference tracks in the step 4); determining the local S-type optimal flight reference track with the minimum cost value in the 2n local S-type optimal flight reference tracks as a global optimal S-type track; and determining the local 'Hui' font optimal flight reference track with the minimum cost value in the 2n local 'Hui' font optimal flight reference tracks as the global optimal 'Hui' font track.

The method for acquiring the local S-shaped optimal flight reference track comprises the following steps:

1) selecting one vertex of the polygonal farmland operation area as an initial vertex, selecting one course as an initial course, and taking a node closest to the initial vertex as a current node.

2) Searching all nodes adjacent to the current node flying by the unmanned helicopter, searching the nodes meeting the constraint conditions, and taking the adjacent nodes meeting the constraint conditions as next possible target nodes; satisfying the constraint condition means: searching a numerical value corresponding to each adjacent node in the node index matrix in the nodes adjacent to the current node, wherein the adjacent node with the minimum numerical value meets the constraint condition;

3) judging whether the current node is a turning point, if so, sequentially calculating the speed vector from the current node to the next possible target node, selecting the node with the smallest included angle with the current direct flight course vector as the next target node, and entering the step 6); if the current node is not the turning point, entering the step 4);

4) calculating cost values from the current node to all the next possible target nodes by using a cost function h (n);

5) selecting a possible target node with the minimum cost value as a next target node;

6) adding 1 to the value of the next target node in the node index matrix to indicate that the node is visited once, and putting the node into a track node sequence list;

7) taking the obtained next target node as the current node of the next cycle, and repeating the steps 2) to 6) until all nodes in the digitized polygonal farmland operation area are traversed to obtain a local S-shaped optimal flight reference track;

8) and repeating the steps 1) to 7), traversing all vertexes of the polygonal farmland operation area, and respectively planning the local S-shaped optimal flight reference flight path of each vertex to obtain 2n local S-shaped optimal flight reference flight paths.

The method for acquiring the local 'back' font optimal flight reference track comprises the following steps:

1) selecting one vertex of the polygonal farmland operation area as an initial vertex, selecting one course as an initial course, and taking a node closest to the initial vertex as a current node.

2) Searching all nodes adjacent to the current node flying by the unmanned helicopter, searching the nodes meeting the constraint conditions, and taking the adjacent nodes meeting the constraint conditions as next possible target nodes; satisfying the constraint condition means: searching a numerical value corresponding to each adjacent node in the node index matrix in the nodes adjacent to the current node, wherein the adjacent node with the minimum numerical value meets the constraint condition;

3) judging whether the current node is a turning point, if so, sequentially calculating the speed vector from the current node to the next possible target node, selecting the node with the smallest included angle with the current direct flight course vector as the next target node, updating a direct flight course vector set I, a side flight course vector set J and an oblique flight course vector set K, and entering the step 6); if the current node is not the turning point, entering the step 4);

4) calculating cost values from the current node to all the next possible target nodes by using a cost function h (n);

5) selecting a possible target node with the minimum cost value as a next target node;

6) adding 1 to the value of the next target node in the node index matrix to indicate that the node is visited once, and putting the node into a track node sequence list;

7) taking the obtained next target node as the current node of the next circulation, and repeating the steps 2) to 6) until all the nodes in the digitized polygonal farmland operation area are traversed to obtain a local 'back' font optimal flight reference track;

8) and repeating the steps 1) to 7), traversing all vertexes of the polygonal farmland operation area, and respectively planning out the local 'returning' font optimal flight reference flight path of each vertex to obtain 2n local 'returning' font optimal flight reference flight paths.

The method for acquiring the global optimal S-shaped flight path comprises the following steps:

1) calculating the direct flight distance x of the unmanned helicopter in each local S-shaped optimal flight reference track_p1Side flight distance x_p5Distance x between takeoff point and track starting point_p2Distance x between intermediate return point and base station_p3(ii) a And the distance x of the unmanned helicopter returning to the position of the unmanned helicopter operator after the operation of the unmanned helicopter is finished_p4；

2) Calculating a cost value of each local S-shaped optimal flight reference track, wherein the cost value f (p) of the local S-shaped optimal flight reference track p is calculated according to the formula:wherein w₁Coefficient, w, of the sum of all direct flight distances of the unmanned helicopter₂Coefficient, w, of the sum of all lateral flight distances of the unmanned helicopter₁+w₂=1；

3) And determining the local S-shaped optimal flight reference track with the minimum cost value as a global optimal S-shaped track.

The method for acquiring the global optimal 'Hui' font track comprises the following steps:

1) calculating the direct flight distance x of the unmanned helicopter in each local 'Hui' shaped optimal flight reference track_q1Turning times t, distance x between takeoff point and track starting point_q2Distance x between intermediate return point and base station_q3(ii) a To be provided withAnd the distance x of the position where the unmanned helicopter operator is located after the unmanned helicopter finishes the operation_q4；

2) Calculating the cost value of each local 'hui' font optimal flight reference track, wherein the calculation formula of the cost value f (q) of the 'hui' font optimal flight reference track q is as follows:

3) the local 'return' font optimal flight reference track with the minimum cost value is determined as the global optimal S-shaped track

Compared with the prior art, the invention has the beneficial effects that: the method can automatically plan the flight path of the unmanned helicopter in a farmland operation area to generate the globally optimal S-shaped flight path and the 'return' shaped flight path, and an operator can randomly select one of the paths to plan pesticide spraying operation according to the operation habit and the operation area condition; the method realizes the local optimal S-shaped track planning and the local optimal 'go back' shaped track planning of any initial vertex and any initial course of any convex polygon farmland operation area; the invention fully considers cost factors such as the flight distance, the turning times, the pesticide amount, the fuel amount and the like of the unmanned helicopter in the operation area, selects the global optimal flight path from the local optimal S-shaped flight path and the local optimal 'return' shaped flight path based on the cost function minimization principle, and has simple calculation and easy realization; the method can be popularized to farmlands with any shapes, including convex polygons and concave polygons, wherein the concave polygons can be divided into a plurality of convex polygons, and then the method is used for planning the flight path of each convex polygon farmland operation area.

Drawings

FIG. 1 is a schematic diagram of a rectangular field area ABCD in an embodiment of the present invention;

FIG. 2 is a diagram illustrating a field area after rotation in an embodiment of the present invention;

FIG. 3 is a schematic view of a field work area after rasterization in an embodiment of the present invention;

FIG. 4 is a schematic diagram illustrating the establishment of a node index matrix according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of neighboring nodes of a current node according to an embodiment of the present invention;

FIG. 6 is a diagram illustrating a next possible target node in an embodiment of the invention;

FIG. 7 is a current node diagram of a next cycle in an embodiment of the present invention;

FIG. 8 is a node index matrix obtained by adding 1 to the next target node value in an embodiment of the present invention;

FIG. 9 is an S-shaped trajectory obtained by an embodiment of the present invention;

FIG. 10 is an S-shaped trajectory line of the original reference coordinate system of an embodiment of the present invention;

FIG. 11 is a "Hui" shaped track of the original reference coordinate system in accordance with an embodiment of the present invention;

FIG. 12 is a first schematic view of a flight path in which the takeoff point S, the base station point M, and the operation completion return point E of the unmanned helicopter are not at the same point;

FIG. 13 is a second schematic view of a flight path in which the takeoff point S, the base station point M, and the operation completion return point E of the unmanned helicopter are not at the same point;

FIG. 14 is a flow chart of the global optimal path planning of the present invention;

FIG. 15 is a flow chart of the local S-shaped optimal trajectory planning of the present invention;

FIG. 16 is a flow chart of the local "go back" type optimal path planning of the present invention.

Detailed Description

The method comprises the following steps:

local optimal S-shaped track planning

1. Digital farmland operation area

Digital maps are the main information source for planning tracks, most of domestic and foreign research digital maps adopt a grid method to divide the terrain into grids with equal intervals, and the grid intervals are selected according to the required actual conditions.

The invention utilizes a grid method to digitize farmland operation areas. Namely, for the farmland with the region ((x, y) |0 ≦ x ≦ a,0 ≦ y ≦ b), the x-axis and the y-axis are divided into small cells according to a given interval, and the discretized farmland region is obtained.

Without loss of generality, for any polygonal farmland area, assuming that the number of vertexes is n, the starting point which can be used as a flight path planning is equal to n. For each starting vertex, in order to ensure that the track can cover the whole area, the optimal starting course is to fly along the left and right sides of the starting vertex in a straight line. There are thus two flight directions available for searching for each starting vertex. There are a total of 2n S-shaped possible tracks to search.

2. Track planning initialization

Initializing the content includes:

1) and creating topographic information of the digital map, and dividing a working area and a non-working area.

The specific implementation manner is to establish a node index matrix (NodeList) for the created digital area. Each grid is a node (hereinafter, each grid is represented by a node), the terrain information of each grid cell is stored in the index matrix, the working area is represented by 0, and the non-working area is represented by #. Thus, a node index matrix storing topographic information is established. The node index value is used for representing the visiting priority of the current node, and the smaller the value is, the higher the priority is. The node index matrix has memory capacity, when a node is visited in the searching process, the value in the index matrix is +1, and the access priority is reduced. When a node in the working area is visited, the value of the node is changed from 0 to 1, and the priority of the node is lowered. Therefore, in the searching process, the nodes which are not visited have the highest priority of the visits all the time, and the heuristic searching is guided to traverse all the nodes in the operation area, so that the planned flight path meets the first requirement of farmland operation.

2) Specifying heuristic information for searching

In heuristic search, the valuation of the location is very important. Different valuations may be used with different results. The valuations in the heuristic are expressed by valuation functions, such as: (n) = g (n) + h (n). Where f (n) is the valuation function of node n, g (n) is the actual cost in state space from the initial node to n nodes, and h (n) is the estimated cost of the best path from n to the target node. Here mainly h (n) embodies the heuristic of the search, since g (n) is known. When h (n) > > g (n), g (n) may be omitted, thereby improving efficiency.

The flight path of the unmanned helicopter operation area must pass through all nodes in the operation area to ensure that all the operation area is covered, so that from a starting point to an operation end of each flight path, a cost function h (n) from a current point to a next target point needs to be specified.

Before defining the cost function, several flight regimes need to be known. The flight mode when unmanned helicopter operation has: straight flight, side flight and oblique flight. In consideration of the actual operation complexity of the unmanned helicopter, the priority order of the three is from high to low. In practice, it is desirable to operate with straight flights as much as possible, and only with side or oblique flights when a boundary is reached or a curve is needed. Therefore, the cost of adopting different flight modes from the current node to the next target node is different.

A set of straight flight heading vectors of the unmanned helicopter is defined as I (the set I contains all straight flight heading vectors), a set of lateral flight heading vectors is J, and a set of oblique flight heading vectors is K. The velocity from the current node to the next target node is a vectorSince the x-axis coincides with the starting heading, I is the set of unit vectors parallel to the x-axis (e.g., vector (1, 0) belongs to I and (-1, 0) also belongs to I), J is the set of unit vectors parallel to the y-axis, and K is the set of unit vectors parallel to line y = x.

Definition of $h\left(n\right)=\left\{\begin{array}{cc}1,& \stackrel{\→}{v}\∈I\\ 2,& \stackrel{\→}{v}\∈J\\ 3,& \stackrel{\→}{v}\∈K\end{array}\right.,$ That is, the velocity vector from the current node to the next target nodeIf the target vector belongs to the direct flight course vector set I, the required cost is 1, namely the minimum estimation cost; if the target vector belongs to a side flight vector set J, the required cost is 2; if the vector belongs to the oblique flight vector set K, the required cost is 3. This is defined to correspond to actual operation. Meanwhile, the priority sequence of visiting the next target node is reflected, so that the next target node can visit the direction with the minimum estimation cost.

3. And (4) performing track search by utilizing a heuristic memory algorithm to obtain 2n S-shaped local optimal flight reference tracks. The specific implementation method is as follows (see fig. 15):

and step1, selecting a vertex as an initial vertex, selecting a heading as an initial heading, and taking a node closest to the initial vertex as a current node.

And step2, searching all nodes adjacent to the current node, searching the nodes meeting the constraint conditions, and taking the adjacent nodes meeting the conditions as possible target nodes of the next step. The constraint condition is the adjacent node with the minimum value in the node index matrix.

And step3, judging whether the current node is a turning point. If the node is a turning point, the velocity vector from the current node to the next possible node is calculated in sequenceAnd selecting the node with the smallest included angle with the current straight flight course vector as the next target node. Step6 is executed.

If not, jump to Step4.

After determining the possible target nodes, all cost values f (n) from the current node to the next possible node are calculated. f (n) = h (n), h (n) velocity vector from current node to next possible nodeAnd (6) determining.

And step5, sorting f (n), and selecting the node with the minimum value of f (n) as a next target node.

And step6, taking the next target node as the current node of the next cycle, and indicating that the node is visited once by using the value +1 in the node index matrix. And put into the track node sequence list.

And Step7, repeating the steps 2 to 6, and searching the next track point until all the nodes in the operation area are traversed, so that the searching process is finished. And obtaining a local S-shaped optimal flight reference track.

And Step 8) traversing all vertexes of the polygonal farmland operation area by utilizing the steps 1) to 7), and respectively planning the local S-shaped optimal flight reference flight path of each vertex to obtain all possible 2n local S-shaped optimal flight reference flight paths.

(II) local optimal 'Hui' shaped track planning

The digitalized farmland operation area of the 'Hui' shaped track planning and the S-shaped track planning is the same as the track planning initialization. The difference lies in that compared with the S type, the unmanned helicopter can hover and turn after reaching a turning point, and the straight flight course of the unmanned helicopter is changed. Therefore, the track search process using the heuristic memory algorithm is different from the S-type. The following description mainly deals with differences.

3. And (3) performing track search by utilizing a heuristic memory algorithm to obtain an optimal flight reference track in a shape of 'hui' with any initial course at any initial vertex. The specific implementation method is as follows (see fig. 16):

and step1, selecting a vertex as an initial vertex, selecting a heading as an initial heading, and taking a node closest to the initial vertex as a current node.

And step2, searching all nodes adjacent to the current node, searching the nodes meeting the constraint conditions, and taking the adjacent nodes meeting the conditions as possible target nodes of the next step. The constraint condition is the adjacent node with the minimum value in the node index matrix.

And step3, judging whether the current node is a turning point. If the node is a turning point, the velocity vectors from the current node to the next possible node are calculated in sequenceAnd selecting the node with the smallest included angle with the current straight flight course vector as the next target node. Step6 is executed, and the straight flight vector set I, the side flight vector set J and the oblique flight vector set K are updated. Wherein the set I comprises velocity vectors associated with the next determined target node to the current nodeAnd all parallel unit vectors, the lateral flight vector set J comprises all unit vectors which are perpendicular to the set I, and the oblique flight vector set K comprises all unit vectors which form an included angle of 45 degrees with the set I.

If not, jump to Step4.

After determining the possible target nodes, all cost values f (n) from the current node to the next possible node are calculated. f (n) = h (n), h (n) velocity vector from current node to next possible nodeAnd (6) determining.

And step5, sorting f (n), and selecting the node with the minimum value of f (n) as a next target node.

And step6, taking the next target node as the current node of the next cycle, and indicating that the node is visited once by using the value +1 in the node index matrix. And put into the track node sequence list.

And Step7, repeating the steps from Step2 to Step6, searching the next track point until all the nodes in the operation area are traversed, and finishing the searching process. And obtaining a local optimal flight reference track in a shape of Chinese character hui.

And Step 8) traversing all vertexes of the polygonal farmland operation area by utilizing the steps 1) to 7), and respectively planning out the local 'returning' font optimal flight reference flight path of each vertex to obtain all possible 2n local 'returning' font optimal flight reference flight paths.

(III) Global optimal S-shaped track and Hui-shaped track planning (see FIG. 14)

The global optimal path planning needs to comprehensively consider various cost problems on the basis of local optimization, and the global optimal path planning of the S-shaped path and the 'Hui' shaped path is discussed below respectively.

1. Global optimal S-shaped track planning

Cost factors to be considered for the global optimal S-shaped track are: the distance between the unmanned helicopter and the starting point of a flight path is the direct flight distance and the side flight distance of the unmanned helicopter in an operation area, the distance between the take-off point of the unmanned helicopter and the starting point of a flight path, the distance between the unmanned helicopter and a base station in a midway return position due to insufficient fuel oil or pesticide, and the return distance after the operation of the unmanned helicopter is completed. In order to comprehensively consider the factors, a global optimal S-shaped track decision method needs to be provided. The method is used for judging and selecting the global optimal flight path from all the local optimal S-shaped flight paths.

The specific implementation method comprises the following steps:

step1, calculating the direct flight distance x of the unmanned helicopter in each flight path p_p1Side flight distance x_p5Starting the process; distance x between takeoff point and track starting point_p2(ii) a Distance x between midway return point and base station_p3(ii) a And the distance x returned after the operation of the unmanned helicopter is finished_p4。

Step2, calculating the cost value f (p) of each track p, wherein the calculation formula of f (p) is as followsWherein w₁Coefficient being the sum of all direct flight distances, w₂Coefficient being the sum of all side flight distances, w₁+w₂=1。

And step3, sequencing all cost values f (p), wherein the lowest cost value is the global optimal S-shaped track.

2. Global optimal Hui-shaped track planning

Cost factors to be considered for the globally optimal 'go back' font track are: the flight distance and the turning times of the unmanned helicopter in the operation area; the distance between the takeoff point of the unmanned helicopter and the starting point of a certain flight path; the distance between the midway return position of the unmanned helicopter and the base station due to insufficient fuel oil or pesticide; the distance returned after the operation of the unmanned helicopter is completed. In order to comprehensively consider the factors, a global optimal 'go back' font track decision method needs to be provided. The method is used for judging and selecting the global optimal track from all local optimal 'go back' type tracks.

The specific implementation method comprises the following steps:

step1, calculating the direct flight distance x of the unmanned helicopter in each flight path p_p1Turning times t; distance x between takeoff point and track starting point_p2(ii) a Distance x between midway return point and base station_p3(ii) a And the distance x returned after the operation of the unmanned helicopter is finished_p4。

Step2, calculating the cost value f (p) of each track p, wherein the calculation formula of f (p) is as followsWherein w₁Coefficient being the sum of all direct flight distances, w₂Coefficient of number of turns, w₁+w₂=1。

And step3, sequencing all cost values f (p), wherein the lowest cost value is the global optimal 'go back' type track.

Therefore, the optimal flight path planning of the unmanned helicopter in the farmland operation area is realized.

The technical solution is further described in detail below with reference to the drawings and the embodiments.

Local optimal S-shaped track planning

As shown in fig. 1, a rectangular farmland area ABCD, an unmanned helicopter can select any vertex as a starting vertex and fly with two sides connected by any vertex as a starting heading. The following describes how to obtain all 8 locally optimal S-shaped track plans for a rectangular farmland area ABCD.

1. Digital farmland operation area

And digitalizing the farmland operation area by using a grid method. Namely, for a farmland with the region ((x, y) |0 ≦ x ≦ a,0 ≦ y ≦ b), the x-axis and the y-axis are divided into small cells according to a given distance to obtain a discretized farmland region, and the distance is determined by the pesticide spraying width of the unmanned helicopter. Here, the x-axis and the y-axis are divided into small cells at a pitch of 5 meters per unit length, as shown in fig. 3.

2. Track planning initialization

Initializing the content includes:

1) and creating topographic information of the digital map, and dividing a working area and a non-working area.

The specific implementation manner is to establish a node index matrix (NodeList) for the created digital area. Each grid is a node (hereinafter, each grid is represented by a node), the terrain information of each grid cell is stored in the index matrix, the working area is represented by 0, and the non-working area is represented by #. Thus, a node index matrix storing topographic information is established.

2) Specifying heuristic information for searching

Defining a set of direct flight heading vectors of the unmanned helicopter as I (the set I contains all direct flightsCourse vector), the set of lateral flight direction vectors is J, and the set of oblique flight direction vectors is K. The velocity from the current node to the next target node is a vector

Definition of $h\left(n\right)=\left\{\begin{array}{cc}1,& \stackrel{\→}{v}\∈I\\ 2,& \stackrel{\→}{v}\∈J\\ 3,& \stackrel{\→}{v}\∈K\end{array}\right.$

That is, the velocity vector from the current node to the next target nodeIf the target heading belongs to the straight-flight heading set I, the required cost is 1, namely the minimum estimation cost. And so on.

3. And (4) performing track search by using a heuristic memory algorithm to obtain all 8 local S-shaped optimal reference flight tracks of the rectangular farmland area ABCD. The specific implementation method comprises the following steps:

and step1, selecting a vertex A as an initial vertex, selecting a heading AB as an initial heading, and taking a node closest to the initial vertex as a current node.

And step2, searching all nodes adjacent to the current node (indicated by black dots in fig. 5), searching nodes meeting constraint conditions, and taking the adjacent nodes meeting the conditions as possible target nodes of the next step. The constraint condition is the adjacent node with the minimum value in the node index matrix.

According to Step2, the next possible target node is shown in fig. 6, and the three black dots with numbers are the next possible target points.

And step3, judging whether the current node is a turning point. If the node is a turning point, the velocity vector from the current node to the next possible node is calculated in sequenceAnd selecting the node with the smallest included angle with the current straight flight course vector as the next target node. Step6 is executed.

If not, jump to Step4.

After determining the possible target nodes, all cost values f (n) from the current node to the next possible node are calculated. f (n) = h (n), h (n) velocity vector from current node to next possible nodeAnd (6) determining.

According to Step4, velocity vectors from the current node point to the three black points with numbers are calculated respectivelyAccording to the velocity vectorTo calculate the arrival at eachThe cost value of the point can be targeted. The cost value to dot No. 1 is 1, the cost value to dot No. 2 is 2, and the cost value to dot No. 3 is 3.

And step5, sorting f (n), and selecting the node with the minimum value of f (n) as a next target node.

The dot No. 1 can be determined as the next target node according to Step5.

And step6, taking the next target node as the current node of the next cycle, and indicating that the node is visited once by using the value +1 in the node index matrix of the next target node (as shown in fig. 8). And put into the track node sequence list.

As shown in fig. 7, dot No. 1 has become the current node for the next loop.

And Step7, repeating the steps 2 to 6, and searching the next track point until all the nodes in the operation area are traversed, so that the searching process is finished. And obtaining a local S-shaped optimal flight reference track.

And obtaining an S-shaped optimal flight reference flight path line taking A as a starting vertex AB as a starting heading according to Step7.

And Step 8) sequentially traversing four vertexes A, B, C and D of the rectangular farmland operation area by utilizing the steps 1) to 7), and respectively planning the local S-shaped optimal flight reference flight path of each vertex to obtain all possible 8 local S-shaped optimal flight reference flight paths.

Fig. 9 shows the whole process of searching, and finally obtains an S-shaped flight path line. The black dots in the region represent points in the track node sequence list.

(II) local optimal 'Hui' shaped track planning

The process is consistent with the process (a), only after the turning point is reached in the step3, the straight flight heading needs to be changed, and the final result is shown in fig. 11.

(III) Global optimal S-shaped track and 'Hui' shaped track planning

1. Global optimal S-shaped track planning

Cost factors to be considered for the global optimal S-shaped track are: the distance between the unmanned helicopter and the starting point of a flight path is the direct flight distance and the side flight distance of the unmanned helicopter in an operation area, the distance between the take-off point of the unmanned helicopter and the starting point of a flight path, the distance between the unmanned helicopter and a base station in a midway return position due to insufficient fuel oil or pesticide, and the return distance after the operation of the unmanned helicopter is completed. In order to comprehensively consider the factors, a global optimal S-shaped track decision method needs to be provided. The method is used for judging and selecting the global optimal flight path from all the local optimal S-shaped flight paths.

The system specifically realizes the method as follows:

step1, calculating the direct flight distance x of the unmanned helicopter in each S-shaped local optimal flight path p of the rectangular ABCD_p1Side flight distance x_p5(ii) a Distance x between takeoff point and track starting point_p2(ii) a Distance x between midway return point and base station_p3(ii) a And the distance x returned after the operation of the unmanned helicopter is finished_p4。

The cost value f (p) is calculated by taking the vertex A as the starting point and AB as the S-shaped track of the starting course.

Suppose that the takeoff point S of the unmanned helicopter, the point M of the base station, and the point E of the work completion return point are all different points, as shown in fig. 12.

Firstly, calculating the direct flight distance x of the unmanned helicopter in the flight path_p1Side flight distance x_p5。x_p1Equal to the sum of all straight flight segments in the flight path. x is the number of_p5Equal to the sum of all the side flight segments in the flight path.

Then calculating the distance between the flying point S and the track starting point A₁Distance x_p2Equal to line segment SA₁Length of (d); distance x between midway return point and base station M_p3Equal to line segment MA₂2 times the length; distance x from return point E after completion of the job_p4Equal to line segment EA₃Length of (d).

Step2, calculating the cost value f (p) of the track, wherein the calculation formula of f (p) is as followsWherein w₁Coefficient being the sum of all direct flight distances, w₂Coefficient being the sum of all side flight distances, w₁+w₂=1。

Coefficient w₁、w₂Calculating according to the ratio of the direct flight speed and the side flight speed

w₁=0.1，w₂=0.9

From this, the cost value f (p) for the track can be calculated.

And step3, repeating the steps to calculate the cost value of each local optimal S-shaped track, sequencing all the cost values f (p), and obtaining the global optimal S-shaped track with the minimum cost value.

Therefore, global optimal S-shaped operation reference track planning is realized.

2. Global optimal Hui-shaped track planning

Cost factors to be considered for the globally optimal 'go back' font track are: the flight distance and the turning times of the unmanned helicopter in the operation area; the distance between the takeoff point of the unmanned helicopter and the starting point of a certain flight path; the distance between the midway return position of the unmanned helicopter and the base station due to insufficient fuel oil or pesticide; the distance returned after the operation of the unmanned helicopter is completed. In order to comprehensively consider the factors, the invention provides a global optimal 'Hui' font track decision method. The method is used for judging and selecting the global optimal track from all local optimal 'go back' type tracks.

The specific implementation method comprises the following steps:

step1, calculating the direct flight distance x of the unmanned helicopter in each local optimal 'return' shaped flight path p of the rectangular ABCD_p1Turning times t; distance x between takeoff point and track starting point_p2(ii) a Distance x between midway return point and base station_p3(ii) a And the distance x returned after the operation of the unmanned helicopter is finished_p4。

The cost value f (p) is calculated by taking the vertex A as the starting point and AB as the back-shaped track of the starting course.

Suppose that the takeoff point S of the unmanned helicopter, the point M of the base station, and the point E of the work completion return point are all different points, as shown in fig. 13.

Firstly, calculating the direct flight distance x of the unmanned helicopter in the flight path_p1And the number of turns t. x is the number of_p1Equal to the sum of all straight flight segments in the flight path. t is equal to the sum of the number of all turning points in the track (4 black dots with number numbers (1-4) are turning points).

Then calculating the distance between the flying point S and the track starting point A₁Distance x_p2Equal to line segment SA₁Length of (d); distance x between midway return point and base station M_p3Equal to line segment MA₂2 times the length; distance x from return point E after completion of the job_p4Equal to line segment EA₃Length of (d).

Step2, calculating the cost value f (p) of the track, wherein the calculation formula of f (p) is as followsWherein w₁Coefficient being the sum of all direct flight distances, w₂Coefficient of number of turns, w₁+w₂=1。

Coefficient w₁、w₂According to the ratio of the complexity of the direct flight operation and the turning operation, the calculation is carried out

w₁=0.25，w₂=0.75

From this, the cost value f (p) for the track can be calculated.

And step3, repeating the steps to calculate the cost value of each local optimal 'go back' font track, sequencing all the cost values f (p), and obtaining the globally optimal 'go back' font track with the minimum cost value.

Therefore, the global optimal 'return' font operation reference track planning is realized.

Claims (4)

1. An automatic planning method for flight paths of an unmanned helicopter pesticide spraying farmland operation area is characterized by comprising the following steps:

1) digitizing the polygonal farmland operation area by using a grid method, and dividing the polygonal farmland operation area into a plurality of grids; the number of vertexes of the polygonal farmland operation area is n;

2) establishing a node index matrix for the digitized polygonal farmland operation area, taking each grid as a node, and storing the terrain information of each grid into the node index matrix;

3) defining the following cost function h (n):

wherein,a velocity vector for the unmanned helicopter to fly from the current node to the next target node; i is a direct flight course vector set of the unmanned helicopter; j is a set of unmanned helicopter lateral flight direction vectors; k is an unmanned helicopter oblique flight vector set;

4) carrying out unmanned helicopter track search by utilizing an heuristic memory method to obtain a local S-shaped optimal flight reference track of any initial course of all initial vertexes of a polygonal farmland operation area and a local 'returning' character-shaped optimal flight reference track of any initial course of all initial vertexes of the polygonal farmland operation area, wherein the local S-shaped optimal flight reference track and the local 'returning' character-shaped optimal flight reference track are respectively 2 n;

5) respectively solving the cost values of the 2n local S-shaped optimal flight reference tracks and the cost values of the 2n local 'back' shaped optimal flight reference tracks in the step 4); determining the local S-type optimal flight reference track with the minimum cost value in the 2n local S-type optimal flight reference tracks as a global optimal S-type track; determining the local 'Hui' font optimal flight reference track with the minimum cost value in the 2n local 'Hui' font optimal flight reference tracks as the global optimal 'Hui' font track;

the method for acquiring the local S-shaped optimal flight reference track comprises the following steps:

1) selecting one vertex of a polygonal farmland operation area as an initial vertex, selecting one course as an initial course, and taking a node closest to the initial vertex as a current node;

2) searching all nodes adjacent to the current node flying by the unmanned helicopter, searching the nodes meeting the constraint conditions, and taking the adjacent nodes meeting the constraint conditions as next possible target nodes; satisfying the constraint condition means: searching a numerical value corresponding to each adjacent node in the node index matrix in the nodes adjacent to the current node, wherein the adjacent node with the minimum numerical value meets the constraint condition;

3) judging whether the current node is a turning point, if so, sequentially calculating the speed vector from the current node to the next possible target node, selecting the node with the smallest included angle with the current direct flight course vector as the next target node, and entering the step 6); if the current node is not the turning point, entering the step 4);

4) calculating cost values from the current node to all the next possible target nodes by using a cost function h (n);

5) selecting a possible target node with the minimum cost value as a next target node;

6) adding 1 to the value of the next target node in the node index matrix to indicate that the node is visited once, and putting the node into a track node sequence list;

7) taking the obtained next target node as the current node of the next cycle, and repeating the steps 2) to 6) until all nodes in the digitized polygonal farmland operation area are traversed to obtain a local S-shaped optimal flight reference track;

8) and repeating the steps 1) to 7), traversing all vertexes of the polygonal farmland operation area, and respectively planning the local S-shaped optimal flight reference flight path of each vertex to obtain 2n local S-shaped optimal flight reference flight paths.

2. The automatic flight path planning method for the unmanned helicopter pesticide spraying farmland operation area according to claim 1, characterized in that the method for obtaining the local 'back' shaped optimal flight reference flight path is as follows:

1) selecting one vertex of a polygonal farmland operation area as an initial vertex, selecting one course as an initial course, and taking a node closest to the initial vertex as a current node;

2) searching all nodes adjacent to the current node flying by the unmanned helicopter, searching the nodes meeting the constraint conditions, and taking the adjacent nodes meeting the constraint conditions as next possible target nodes; satisfying the constraint condition means: searching a numerical value corresponding to each adjacent node in the node index matrix in the nodes adjacent to the current node, wherein the adjacent node with the minimum numerical value meets the constraint condition;

3) judging whether the current node is a turning point, if so, sequentially calculating the speed vector from the current node to the next possible target node, selecting the node with the smallest included angle with the current direct flight course vector as the next target node, updating a direct flight course vector set I, a side flight course vector set J and an oblique flight course vector set K, and entering the step 6); if the current node is not the turning point, entering the step 4);

4) calculating cost values from the current node to all the next possible target nodes by using a cost function h (n);

5) selecting a possible target node with the minimum cost value as a next target node;

6) adding 1 to the value of the next target node in the node index matrix to indicate that the node is visited once, and putting the node into a track node sequence list;

7) taking the obtained next target node as the current node of the next circulation, and repeating the steps 2) to 6) until all the nodes in the digitized polygonal farmland operation area are traversed to obtain a local 'back' font optimal flight reference track;

8) and repeating the steps 1) to 7), traversing all vertexes of the polygonal farmland operation area, and respectively planning out the local 'returning' font optimal flight reference flight path of each vertex to obtain 2n local 'returning' font optimal flight reference flight paths.

3. The automatic flight path planning method for the unmanned helicopter pesticide spraying farmland operation area according to claim 1, characterized in that the method for obtaining the global optimal S-shaped flight path is as follows:

1) calculating the direct flight distance x of the unmanned helicopter in each local S-shaped optimal flight reference track_p1Side flight distance x_p5Takeoff point is from trackStarting point distance x_p2Distance x between intermediate return point and base station_p3(ii) a And the distance x of the unmanned helicopter returning to the position of the unmanned helicopter operator after the operation of the unmanned helicopter is finished_p4；

2) Calculating a cost value of each local S-shaped optimal flight reference track, wherein the cost value f (p) of the local S-shaped optimal flight reference track p is calculated according to the formula:wherein w₁Coefficient, w, of the sum of all direct flight distances of the unmanned helicopter₂Coefficient, w, of the sum of all lateral flight distances of the unmanned helicopter₁+w₂＝1；

3) And determining the local S-shaped optimal flight reference track with the minimum cost value as a global optimal S-shaped track.

4. The automatic flight path planning method for the unmanned helicopter pesticide spraying farmland operation area according to claim 2, characterized in that the method for obtaining the global optimal 'return' shaped flight path is as follows:

1) calculating the direct flight distance x of the unmanned helicopter in each local 'Hui' shaped optimal flight reference track_q1Turning times t, distance x between takeoff point and track starting point_q2Distance x between intermediate return point and base station_q3(ii) a And the distance x of the unmanned helicopter returning to the position of the unmanned helicopter operator after the operation of the unmanned helicopter is finished_q4；

2) Calculating the cost value of each local 'hui' font optimal flight reference track, wherein the calculation formula of the cost value f (q) of the 'hui' font optimal flight reference track q is as follows:

3) and determining the local 'return' font optimal flight reference track with the minimum cost value as the global optimal S-shaped track.