CN111461464A - Plant protection unmanned aerial vehicle cluster operation task allocation method and device - Google Patents

Abstract

The invention discloses a plant protection unmanned aerial vehicle cluster operation task allocation method which comprises the steps of dividing an operation field into N sub-fields, allocating all the sub-fields to all plant protection unmanned aerial vehicles participating in operation according to a preset allocation principle, searching an optimal operation allocation matrix by adopting an MOSE L A algorithm, determining an operation time matrix of the sub-fields according to the optimal operation allocation matrix and an aircraft operation time matrix, defining an operation sequence matrix of the sub-fields by combining a preset assembly rule, and optimizing the operation sequence matrix of the sub-fields by adopting a genetic algorithm.

Description

Plant protection unmanned aerial vehicle cluster operation task allocation method and device

Technical Field

The invention relates to the technical field of cluster operation of plant protection unmanned aerial vehicles, in particular to a method and a device for distributing cluster operation tasks of plant protection unmanned aerial vehicles.

Background

The plant protection unmanned aerial vehicle has the remarkable characteristics of high speed, high efficiency, wide adaptability and the like, is flexible in operation, does not need a take-off and landing runway, can adapt to complex terrains such as hills, mountainous areas, sloping fields, paddy fields and the like, is rapidly developed in China in recent years, and the annual accumulated operation area exceeds 5 hundred million mu times.

However, the plant protection unmanned aerial vehicle is limited by the carrying capacity of the machine (less than 20kg), the single-frame operation area generally does not exceed 30 mu, and in the actual operation process, the machine needs to waste 40% of the operation time for frequently taking off and landing, replacing batteries and filling pesticides. In order to improve the operation efficiency, part of enterprises and scientific research teams develop the operation technical research of one control of multiple machines, the current research work mainly focuses on networking communication and system hardware development, the most representative of the research work is that a P-30XP type plant protection unmanned aerial vehicle is released by Guangzhou polar flight science and technology company in 2019, and a T-20 type plant protection unmanned aerial vehicle released by Dajiang innovation company in 2019 supports that one person controls more than two machines to operate simultaneously. But the operation path planning mode is single, the field blocks are simply divided according to the operation capacity of the machines, and each machine adopts an independent air route to perform sequencing operation. Only the full-coverage path planning is considered, the cooperation and the coordination among machines are rarely considered, and the situations of lifting conflict and unreasonable task allocation are easily generated in the actual operation.

Disclosure of Invention

The invention aims to provide a method and a device for distributing operation tasks of a plant protection unmanned aerial vehicle cluster, which are used for integrating two constraint conditions of time and energy consumption on the basis of ensuring the full coverage of an airway, establishing a double-layer decision method for distributing the shortest flight distance and the optimal time, reducing time conflict and improving operation efficiency.

In order to achieve the above purpose, with reference to fig. 1, the present invention provides a plant protection unmanned aerial vehicle cluster operation task allocation method, where the allocation method includes the following steps:

s1, dividing the operation field into N sub-fields according to the input field parameters and the plant protection unmanned aerial vehicle parameters, wherein the operation requirements of each sub-field are the same, and only the sub-field is required to be operatedCorresponding to one unmanned aircraft, defining the set of N sub-fields as H_j，j＝1,2,3,…,N；

S2, distributing all the sub-fields to all the plant protection unmanned aerial vehicles participating in the operation according to a preset distribution principle, wherein the preset distribution principle is as follows: a single plant protection unmanned aerial vehicle is allocated with at least one sub-field block, each sub-field block can be allocated only once, and all the sub-field blocks are allocated;

s3, setting the number of unmanned planes, the number of field blocks K, the number of sub-field blocks N and the operation time matrix (T)_K×NAnd the distance matrix of the in/out field is (L)¹)_K×N、(L²)_K×NBased on the operation, an MOSE L A algorithm is adopted to find an optimal operation distribution matrix (X)_K×N；

S4, distributing matrix (X) according to the optimal operation_K×NAnd the aircraft operation time matrix (T)_K×NDetermining an operation time matrix (T') of the sub-field block_N×1(ii) a Defining a sequence of operations matrix (SE) for the sub-fields in combination with predetermined assembly rules_N×1The preset assembly rule is that after each unmanned aircraft lands, assembly tasks including battery replacement and medicine chest replacement are carried out, but only one unmanned aircraft executes the assembly tasks at the same time;

s5, adopting genetic algorithm to work sequence matrix (SE) of the sub-farmland blocks_N×1And (6) optimizing.

As a preferred example, in step S1, the process of dividing the working field into N sub-fields according to the input field parameters and plant protection unmanned aerial vehicle parameters includes the following steps:

s11, putting the fields with the same operation requirements into the same set, wherein the number of the fields is M;

s12, further field processing is carried out on all the fields in the M sets according to the one-time lifting/landing operation area of the plant protection unmanned aerial vehicle, and the field processing comprises the operations of adjacent small field merging and large field dividing;

s13, extracting all the sub-field blocks in the M sets, setting the number of the sub-field blocks to be N, and defining the N sub-field block sets to be H_j，j＝1,2,3,…,N。

As a preferred example, in step S1 and step S2, the allocating all the sub-fields to all the plant protection unmanned aerial vehicles participating in the operation according to the preset allocation principle includes:

defining the operation distribution matrix of plant unmanned plane as (X)_K×NWherein a parameter x is assigned_ijThe assignment is defined as:

where j is 1,2,3, …, N.

As a preferred example, in step S1 and step S3, the MOSE L A algorithm is used to find the optimal job distribution matrix (X)_K×NComprises the following steps:

s31, inputting population scale S, sub-population number S, individual frog num of the population, total population iteration number D and population iteration number I;

s32, initializing the frog group P ═ X¹,X²,…,X^S}；

S33, calculating adaptive values, arranging the adaptive values in a descending order, and dividing the group;

s34, intra-population evolution: each sub-population makes a first jump according to the following formula:

wherein: x^bAnd X^wRespectively, the optimal individual and the worst individual of the sub-group;

s35, merging of first population: and (4) performing descending arrangement according to the adaptive value, and combining the following formula, and performing second jump on the whole group:

wherein: x^gAnd X^wRespectively, the optimal individual and the worst individual of the whole population;

s36, second population merging: and (4) performing descending arrangement according to the adaptive value, and performing third jump on the weakest individuals of the population by combining the following formula:

(X^w)′＝rand(X),X∈Q^M×N

wherein: rand (X) represents randomly generating a decision matrix in the feasible solution space;

s37, judging whether the current cycle times are less than the total group iteration times D, if so, turning to the step S33, otherwise, outputting the optimal individual X¹；

Wherein, the related parameters of population update are defined as follows:

let matrix X¹∈Q^M×N、X²∈Q^M×NWherein Q is^M×NFor a feasible solution space, a matrix is defined

Represents X²And X¹The matrix of the differences between the two,

indicating an ectopic or operational point between matrices;

the meaning of z ═ cut (z, k) is: intercepting the front k columns of the one-dimensional array z and storing the front k columns into a new one-dimensional array z';

(X¹) ' -F (B, Φ) means: updating an approximation matrix B for the matrix X according to the difference matrix B, counting element sums of all columns of B, if the element sums are larger than 0, storing the column number into a one-dimensional array z, and defining an operator phi as follows:

φ＝～(B,X¹,cut(z,ri(n)))

wherein: n is the length of the array z, ri (n) denotes the interval [1, n ]]Any integer within; "-" indicates that X is assigned to the column number ri (n) designated by z¹Element correction in the column is consistent with the arrangement of the B element;

the meaning of (A) is: for matrix X¹According to the operator

Performing mutation operation, operator

Is defined as:

in the formula: switch (r)₁,r₂) Representing r in a switching matrix₁Row and r₂Line, swicth (c)₁,c₂) Representing c in the switching matrix₁Column sum c₂And (4) columns.

As a preferred example, in step S1, in step S33, the process of dividing the population includes the following steps:

s331, defining an adaptive value function FV of an individual t₁As follows below, the following description will be given,

s332, taking the adaptive value of the individual as a reference for dividing the group according to FV₁ ^tIn descending order, the individuals in the ith subgroup are:

{FV₁ ^k|k∈[(i-1)×num+1,i×num]}

wherein i is 1,2,3, …, s.

As a preferred example, in step S1 and step S34, the process of intra-population evolution includes the following steps:

s341, jumping to the optimal individual of the sub-population according to the following formula:

s342, introducing a mutation operator for detecting whether the plant protection unmanned aerial vehicle which is not distributed to the field exists, and performing mutation detection and processing by adopting the mutation operator:

(X^w)″＝F((X^w)′,switch([r₀,r_m],c))

in the formula: r is₀Representation matrix (X)^w) ' line element and line number 0; r is_mRepresentation matrix (X)^w) ' line element sum is the largest line number; c represents r_mAny column number with a median of 1; switch ([ r ]₀,r_m]And c) is represented in matrix (X)^w) In column c of `, r is exchanged₀And r_mWhen r is an element of₀When the value of (A) is not unique, a jump formula corresponding to the first jump is adopted for carrying out multiple variations;

s343, updating the optimal and worst individuals;

and S344, repeating the steps S341 to S343 until the cycle number reaches I times.

As a preferred example, in step S1, in step S35, the process of merging the first population includes the following steps:

s351, jumping to the population-optimized individual using the following formula:

s352, introducing a mutation operator for detecting whether the plant protection unmanned aerial vehicle which is not distributed to the field exists, and performing mutation detection and processing by adopting the mutation operator:

(X^w)″＝F((X^w)′,switch([r₀,r_m],c))

in the formula: r is₀Representation matrix (X)^w) ' line element and line number 0; r is_mRepresentation matrix (X)^w) ' line element sum is the largest line number; c represents r_mAny column number with a median of 1; switch ([ r ]₀,r_m]And c) is represented in matrix (X)^w) In column c of `, r is exchanged₀And r_mWhen r is an element of₀When the value of (A) is not unique, a jump formula corresponding to the first jump is adopted for carrying out multiple variations;

and S353, updating the optimal and worst individuals.

As a preferable example of these, in step S4, the operation sequence matrix (SE) defining the sub-fields_N×1Comprises the following steps:

s41, according to the airplane operation time matrix, the time matrix is (T)_K×NDetermining a time matrix (T') of the operation of the field_N×1，t_j' is:

wherein j is 1,2,3, …, N;

s42, defining the operation sequence matrix of the field block as (SE)_N×1，se_jIs [1, N ]]A non-repeating set of integers within the interval;

s42, according to the operation Sequence (SE)_N×1The time matrix after defining the sequence is

It is composed of

Comprises the following steps:

s43, based on (SE)_N×1And

defining the operation time of the plant protection unmanned plane-field block as (TA)_K×N’，ta_ij′Indicates the time of j ' th operation of the ith aircraft, j ' is 1,2,3, …, N ' is max { sum [ x (i,:)]}；(TA)_K×N’Each column of the non-assigned items is completed by 0;

s44, considering the overlapping time of assembling batteries and medicine boxes simultaneously on a plurality of plant protection unmanned planes, defining the landing time matrix of the unmanned planes as (T L)_K×N’Takeoff time matrix is (TR)_K×N’，tl_ij′And tr_ij′Are respectively provided withComprises the following steps:

wherein, ts_ij′The time required for the aircraft to fit to the next takeoff after landing, here set to a fixed length, ts_ij′＝c。

As a preferred example, in step S5, the operation sequence matrix (SE) of the sub-field block is implemented by using genetic algorithm_N×1The process of performing the optimization includes the steps of:

s51, evaluating the adaptive value of the sub-field operation:

order of operation (SE)_N×1As chromosome, in determining (SE)_N×1Under the condition of (3), determining the final landing time of different plant protection unmanned aerial vehicles to be tl_KN’Defining an fitness value selection function FV of a chromosome₂Comprises the following steps:

wherein g (SEt) max [ tl [ ]_KN′(SEt,c)](ii) a t represents the position of the chromosome in the population; wt is a time adaptive value parameter, the unit is min, and the value range is [300,600,]；

s52, genetic operator definition:

a. chromosome selection operations

Selecting initial population by roulette method, if fitness of a certain individual t is f_tThen the probability that the individual is selected is denoted as P_t：

Wherein N represents the chromosome scale;

b. chromosome crossing operations

Carrying out chromosome crossing operation by adopting a partial matching crossing method: randomly generating two cross points and determining a matching section; generating two sub-individuals according to the mapping relation between the matching segments and the non-matching segments in the parent;

c. operation of chromosomal variation

Randomly selecting N positions, and interchanging genes corresponding to the N positions, wherein N is more than 1;

s53, updating chromosome based on GA algorithm to obtain optimal operation sequence matrix (SE) of sub-field_N×1The method comprises the following substeps:

s531, inputting population size N, maximum iteration times max (Gen) of evolution, and cross rate p_cThe rate of variation p_m；

S532, generating an initial population { SE }_PThe total population size is P;

s533, calculating the adaptive value of the population individuals;

s534, determining the following conditions for stopping genetic calculation: (1) the iteration number reaches max (Gen), (2) the requirement of an adaptive value is met, (3) the optimal chromosome is continuously kept for 10 generations; if any one of the conditions is satisfied, the iteration process is terminated, and the optimal individual { SE } is output_max(f)If not, go to step S535;

s535, selecting and calculating other chromosomes except the optimal chromosome;

s536, performing cross operation on other chromosomes except the optimal chromosome;

s537, performing mutation operation on other chromosomes except the optimal chromosome, and returning to step S533.

Based on the distribution method, the invention also provides a plant protection unmanned aerial vehicle cluster operation task distribution device, which comprises an ARM embedded system, and a GPS positioning system, a temperature and humidity sensor, a data transmission module, a power supply module, an L CD touch screen and a cache module which are respectively connected with the ARM embedded system;

the temperature and humidity sensor is used for detecting the real-time temperature and the real-time humidity of the operation area and feeding back the detection result to the ARM embedded system;

the L CD touch screen is used for inputting field parameters and plant protection unmanned aerial vehicle parameters to the ARM embedded system, and the field parameters comprise boundary coordinates of each field;

the ARM embedded system combines input field parameters, plant protection unmanned aerial vehicle parameters and environment information fed back by a temperature and humidity sensor, adopts the distribution method, jointly calculates all plant protection unmanned aerial vehicle cooperative task distribution models and corresponding scheduling schemes of plant protection unmanned aerial vehicle cluster operation, sends the scheduling schemes to the flight control system of each plant protection unmanned aerial vehicle through a data transmission module, and stores the input parameters and the corresponding scheduling schemes to a cache module;

the power supply module is used for providing electric energy required by normal work of the ARM embedded system, the GPS positioning system, the temperature and humidity sensor, the data transmission module, the power supply module, the L CD touch screen and the buffer module;

the GPS is used for acquiring the position information of each plant protection unmanned aerial vehicle and displaying the acquired position information to a user through an L CD touch screen.

Compared with the prior art, the technical scheme of the invention has the following remarkable beneficial effects:

(1) on the basis of ensuring the full coverage of the air route, two constraint conditions of time and energy consumption are integrated, a double-layer decision method for distributing the shortest flight distance and the optimal time is established, time conflict is reduced, and the operation efficiency is improved.

(2) The whole operation process of all plant protection unmanned aerial vehicles is automatically planned, the manpower loss is reduced, the maintenance and the monitoring of all unmanned aerial vehicles can be completed by only one worker, and the true one-control-multiple-machine operation is realized.

It should be understood that all combinations of the foregoing concepts and additional concepts described in greater detail below can be considered as part of the inventive subject matter of this disclosure unless such concepts are mutually inconsistent. In addition, all combinations of claimed subject matter are considered a part of the presently disclosed subject matter.

The foregoing and other aspects, embodiments and features of the present teachings can be more fully understood from the following description taken in conjunction with the accompanying drawings. Additional aspects of the present invention, such as features and/or advantages of exemplary embodiments, will be apparent from the description which follows, or may be learned by practice of specific embodiments in accordance with the teachings of the present invention.

Drawings

The drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component that is illustrated in various figures may be represented by a like numeral. For purposes of clarity, not every component may be labeled in every drawing. Embodiments of various aspects of the present invention will now be described, by way of example, with reference to the accompanying drawings, in which:

fig. 1 is a flowchart of a plant protection unmanned aerial vehicle cluster work task allocation method of the present invention.

Fig. 2 is a flow chart of a distribution method according to a first embodiment of the present invention.

FIG. 3 is a flow diagram illustrating the division of a field of the present invention.

FIG. 4 is a MOSF L A based search for an optimal job assignment matrix (X)_K×NAnd (4) a flow chart.

FIG. 5 is a schematic diagram of chromosome crossover preparation.

FIG. 6 is a schematic diagram of chromosome crossing completion.

FIG. 7 is a schematic diagram of a chromosomal mutation process.

FIG. 8 is an operation order matrix (SE) for a sub-field block using a genetic algorithm_N×1And carrying out an optimization flow chart.

Fig. 9 is a schematic structural diagram of a plant protection unmanned aerial vehicle cluster operation task distribution device.

Fig. 10 is a schematic view of a target work field of a plant protection unmanned aerial vehicle according to a third embodiment.

Fig. 11 is a schematic diagram of a target field block preprocessing method in the third embodiment.

FIG. 12 is a comparison of the results of the third embodiment.

Detailed Description

In order to better understand the technical content of the present invention, specific embodiments are described below with reference to the accompanying drawings.

Detailed description of the preferred embodiment

With reference to fig. 1, the present invention provides a plant protection unmanned aerial vehicle cluster work task allocation method, which includes the following steps:

s1, dividing the operation field into N sub-fields according to the input field parameters and the plant protection unmanned aerial vehicle parameters, wherein the operation requirements of each sub-field are the same, and only corresponding to one unmanned aerial vehicle, the set of N sub-fields is defined as H_j，j＝1,2,3,…,N。

S2, distributing all the sub-fields to all the plant protection unmanned aerial vehicles participating in the operation according to a preset distribution principle, wherein the preset distribution principle is as follows: a single plant protection drone is assigned at least one sub-field, each sub-field can only be assigned once and all sub-fields are assigned.

S3, setting the number of unmanned planes, the number of field blocks K, the number of sub-field blocks N and the operation time matrix (T)_K×NAnd the distance matrix of the in/out field is (L)¹)_K×N、(L²)_K×NBased on the operation, an MOSE L A algorithm is adopted to find an optimal operation distribution matrix (X)_K×N。

S4, distributing matrix (X) according to the optimal operation_K×NAnd the aircraft operation time matrix (T)_K×NDetermining an operation time matrix (T') of the sub-field block_N×1(ii) a Defining a sequence of operations matrix (SE) for the sub-fields in combination with predetermined assembly rules_N×1And after each unmanned aircraft lands, the preset assembly rule is that assembly tasks including battery replacement and medicine box replacement are carried out, but only one unmanned aircraft executes the assembly tasks at the same moment.

And S5, optimizing a work sequence matrix (SE) N × 1 of the sub-field block by adopting a genetic algorithm.

Referring to fig. 1 and 2, assuming known conditions of ① known field positions, the number of plant protection unmanned planes and take-off and landing positions, ② known flight and operation performance of the planes and unmanned plane operation parameters of each field, ③ only one labor force (one control machine) when the unmanned planes land for medicine box and battery assembly, the process of unmanned plane distribution and field operation sequencing of the operation fields under the coordination of a plurality of plant protection unmanned planes comprises the following steps:

first, repartitioning of the field pieces

Assuming that n blocks of the common target operation field exist, the dividing steps are as follows: (I) putting the field blocks with the same operation requirements (pesticide type, pesticide solution concentration, airplane operation parameters and the like) into the same set, wherein the total number of the field blocks is M; (II) according to the one-time lifting/landing operation area of the plant protection unmanned aerial vehicle, further dividing all the fields in the M sets into fields, merging adjacent small fields and dividing large fields; (III) extracting all sub-field blocks in the M sets, counting as N, redefining the N field block sets into H_j(j ═ 1,2,3, …, N). The dividing step is shown in fig. 3 below.

Secondly, distributing unmanned aerial vehicle operation field blocks based on MOSF L A

A single plant protection drone may be assigned to multiple fields (< N), but each field may only be assigned once.

Defining the operation distribution matrix of plant unmanned plane as (X)_K×NWherein a parameter x is assigned_ijAn assignment is defined as

Where j is 1,2,3, …, N. The unmanned aircraft allocation parameters follow the constraint conditions: each field can only be operated once and all fields are assigned to plant protection drones.

Third, job allocation optimization based on MOSF L A

Given the number of airplanes, the field size K, N, and the operation time matrix (T)_K×NThe distance matrix of the in/out field is (L)¹)_K×N、(L²)_K×NThe MOSF L A algorithm is adopted to find the optimal operation distribution matrix as (X)_K×N。

Through the co-evolution of multi-clan groups, the mixed frog leaping algorithm can realize the update of the group. To optimize the distribution parameter X in equation (1), the definition of the population update method is as follows.

Definition 1 sets matrix X¹∈Q^M×N、X²∈Q^M×NWherein Q is^M×NFor a feasible solution space, a matrix is defined

Represents X²And X¹The matrix of the differences between the two,

indicating an ectopic or operator-by-matrix.

Definition 2z ═ cut (z, k) is defined as: and intercepting the first k columns of the one-dimensional array z and storing the first k columns into a new one-dimensional array z'.

Definition 3 (X)¹) ' - < F (B, φ), based on the difference matrix B (i.e.

) The matrix X is updated to approximate the matrix B. And counting the element sum of each column B, and if the element sum is greater than 0, storing the column number into a one-dimensional array z. The definition of the operator φ is:

φ＝～(B,X¹,cut(z,ri(n))). (2)

wherein: n is the length of the array z, ri (n) denotes the interval [1, n ]]Any integer; "-" indicates: according to the column number ri (n) designated by z' (i.e. cut (z, ri (n))), X is added¹Element correction in the column is consistent with B element arrangement

Definition 4

Is a pair matrix X¹According to the operator

Performing mutation operation, operator

Is defined as:

in the formula: switch (r)₁,r₂) Representing r in a switching matrix₁Row and r₂Line, swicth (c)₁,c₂) Representing c in the switching matrix₁Column sum c₂And (4) columns.

Frog leaping rule of MOSF L A

The rules related to the frog leaping designed in the invention are as follows

The number 1 of the jumps is the number one jump,

wherein: x^bAnd X^wRespectively, the best individual and the worst individual of the subgroup.

The 2 nd hop, the first hop,

wherein: x^gAnd X^wRespectively the best and worst individuals of the entire population.

The 3 rd time of the jump is the jump,

(X^w)′＝rand(X),X∈Q^M×N. (6)

wherein: rand (x) represents the random generation of a decision matrix in the feasible solution space.

Therefore, the optimization capability of the worst individual in the sub-population and the total population to the optimal individual in the population and the global updating process of the total population are embodied.

The new individuals generated may have a problem with the aircraft not being assigned to the field (i.e., the aircraft is not assigned to the field)

). The invention introduces mutation operators as follows to carry out mutation processing on the partial components:

(X^w)″＝F((X^w)′,switch([r₀,r_m],c)). (7)

in the formula: r is₀Representation matrix (X)^w) ' line element and line number 0; r is_mRepresentation matrix (X)^w) ' line element sum is the largest line number; c represents r_mAny column number with a median of 1; switch ([ r ]₀,r_m]And c) is represented in matrix (X)^w) In column c of `, r is exchanged₀And r_mWhen r is an element of₀The value of (3) is not exclusive and may be varied several times according to the expression (4).

Group division of individuals

Assuming that the total number of the frog-jump population is S, the number of the sub-populations is S, and the number of the sub-populations is num (S is S × num), an adaptive value function FV of the individual t is defined based on the formula (6)₁As follows below, the following description will be given,

the present invention uses the fitness value of an individual as a reference for population division. According to FV₁ ^tIn descending order, the individuals in the ith subgroup are respectively,

{FV₁ ^k|k∈[(i-1)×num+1,i×num]}. (9)

wherein i is 1,2,3, …, s.

According to the algorithm design, the specific steps of the modified MOSF L a algorithm for the multi-machine cooperative allocation problem proposed by the present invention are shown in fig. 4.

Sixth, field operation ordering model based on GA

Based on the MOSF L A algorithm, we have found a superior job assignment matrix as (X)_K×N. According to the aircraft operation time matrix as (T)_K×NWe can determine the operation time matrix (T') of the field block_N×1T of which_j' to (a) is,

where j is 1,2,3, …, N. Define the operation sequence matrix of the field as (SE)_N×1，j＝1,2,3,…,N，se_jIs [1, N ]]Non-repeating whole within intervalA set of numbers. According to the operation Sequence (SE)_N×1The time matrix after defining the sequence is

It is composed of

In order to realize the purpose,

based on (SE)_N×1And

defining the operation time of the plant protection unmanned plane-field block as (TA)_K×N’，ta_ij′Indicates the time of j ' th operation of the ith aircraft, j ' is 1,2,3, …, N ' is max { sum [ x (i,:)]}。(TA)_K×N’Each column of unassigned entries is completed with 0.

Considering the overlapping time of assembling batteries and medicine boxes simultaneously for a plurality of plant protection unmanned planes, the landing time matrix of the unmanned plane is defined as (T L)_K×N’Takeoff time matrix is (TR)_K×N’，tl_ij′And tr_ij′Respectively as follows:

wherein, ts_ij′The time required for assembly to the next takeoff after landing of the aircraft, irrespective of the assembly time differences of the assemblers, may be set to a fixed length, ts, here_ij′C, simplifying the computational complexity.

Seventh, GA-based field block operation sequencing optimization

To reduce multiple unmannedThe overall time of simultaneous machine operations requires determining the sequence of operations (SE) of a field_N×1. The invention optimizes the operation sequence by GA algorithm.

7.1 evaluation of adaptive value for field work

The invention combines the operation Sequence (SE)_N×1As chromosome, in determining (SE)_N×1Under the condition, the final landing time tl of different plant protection unmanned aerial vehicles can be determined_KN’Defining an fitness value selection function FV for the chromosome₂The definition is that,

wherein g (SEt) max [ tl [ ]_KN′(SEt,c)](ii) a t represents the position of the chromosome in the population; wt is a time adaptive value parameter (min) with a value range of [300,600]。

7.2 genetic operator definition

a. Chromosome selection operations

Selecting initial population by roulette method, if fitness of a certain individual t is f_tThe probability that the individual is selected can be expressed as P_t：

Where N represents the chromosome size.

b. Chromosome crossing operations

With reference to fig. 5 and fig. 6, a partial matching Crossover method (PMX) is used to perform Crossover operation of chromosomes: randomly generating two cross points and determining a matching section; and generating two sub-individuals according to the mapping relation between the matching segment and the non-matching segment in the parent. For example, there are two parents, and the repeated numbers are marked with a symbol, assuming that the matching segment generated randomly is [4,6 ]. Based on the intermediate mapping, the number of repeated portions may be determined, the first occurrence of font a1 may be assigned an initial value of 1, the occurrence of intersections of 1 originating from 6, and 6 participating in intersections of which the initial value is 5, 5 no longer participating in intersections, and the first occurrence of font a1 may be assigned a value of 5; the second occurrence of a1 is assigned a value of 4 in the same way. The assignment at child B1 is done in the same way.

c. Operation of chromosomal variation

And (3) randomly selecting N (>1) positions, and interchanging the corresponding genes. For chromosome A, the selected gene positions are 3, 5 and 6, and the compiled chromosome is shown in FIG. 7.

Eighthly, GA-based chromosome renewal process

The conditions for stopping genetic calculation of the chromosome in the invention are ① iteration to max (Gen) times, the algorithm is terminated, ② meets the requirement of an adaptive value, the algorithm is terminated, ③ the optimal chromosome is continuously maintained for 10 generations, and the algorithm is terminated, and the GA-based chromosome updating process is shown in figure 8.

In the invention, the population specification P is 20-100; the iteration number max (Gen) of genetic evolution is 100-500; cross probability p_c0.4-0.99; probability of variation p_m＝0.0001～0.1。

Detailed description of the invention

With reference to fig. 9, based on the foregoing allocation method, the present invention further provides a plant protection unmanned aerial vehicle cluster operation task allocation device, where the allocation device includes an ARM embedded system, and a GPS positioning system, a temperature and humidity sensor, a data transmission module, a power supply module, an L CD touch screen, and a cache module, which are respectively connected to the ARM embedded system.

The temperature and humidity sensor is used for detecting the real-time temperature and the real-time humidity of the operation area and feeding back the detection result to the ARM embedded system.

The L CD touch screen is used for inputting field parameters and plant protection unmanned aerial vehicle parameters to the ARM embedded system, and the field parameters comprise boundary coordinates of each field.

The ARM embedded system combines the input field parameters, the plant protection unmanned aerial vehicle parameters and the environment information fed back by the temperature and humidity sensor, adopts the distribution method, jointly calculates all plant protection unmanned aerial vehicle cooperative task distribution models and the corresponding scheduling schemes of the plant protection unmanned aerial vehicle cluster operation, and sends the scheduling schemes to the flight control system of each plant protection unmanned aerial vehicle through the data transmission module, so that the flight control system controls machines and tools to perform cluster operation according to the received instructions, and the input parameters and the corresponding scheduling schemes are stored in the cache module.

The power supply module is used for providing electric energy required by normal work of the ARM embedded system, the GPS positioning system, the temperature and humidity sensor, the data transmission module, the power supply module, the L CD touch screen and the buffer module.

The GPS is used for acquiring the position information of each plant protection unmanned aerial vehicle and displaying the acquired position information to a user through an L CD touch screen.

Detailed description of the preferred embodiment

In order to determine the optimization performance of the multi-plant protection unmanned aerial vehicle operation planning based on the MOSF L A + GA collaborative task allocation algorithm, the target multi-field block with the total area of 330 mu is selected by the invention as shown in the following figures 10 and 11.

The maximum cruising time of the selected plant protection unmanned aerial vehicle is 25min, the drug loading is 15L, the plant protection unmanned aerial vehicle is divided again according to the multiple fields in the figure 10 based on the unmanned aerial vehicle, the division result is shown in figure 11, the area of the divided 22 fields is less than or equal to 15 mu, so that the plant protection unmanned aerial vehicle can complete spraying operation once for one frame, the rising/landing area of the unmanned aerial vehicle is positioned near the center of the fields, the area simultaneously meets the requirements of being convenient to take off, transporting liquid medicine batteries and the like, when multiple aircrafts land at the same time, the maximum distance between the aircrafts is 100m, so that the assembling work of the liquid medicine and the batteries is ensured to be completed within the specified time, in the invention, the position of 100m at the center of the fields is selected as the rising and landing assembling point of the aircrafts, the distance between the adjacent landing points is 25m, the assembling time c is 6min, and the.

In order to compare the weighted distance and the total flight time of the airplane entering/exiting the field under the operation of the unmanned aerial vehicle under the normal condition, the invention compares the following two traditional multi-machine operation modes:

in the conventional multi-aircraft operation mode ①, the aircraft is not sure of the field of operation, but the aircraft will perform the next field of operation after completing the previous field of operation.

Conventional multi-aircraft operation mode ②. unmanned aircraft have defined fields, each with a fixed sequence of operations.

The results of two conventional multi-machine operation modes are as follows:

compared with the traditional mode, the method can save 25min at most by adopting the MOSFA L + GA algorithm, and the working efficiency and the total time of the unmanned aerial vehicle cluster are shown in figure 12.

In this disclosure, aspects of the present invention are described with reference to the accompanying drawings, in which a number of illustrative embodiments are shown. Embodiments of the present disclosure are not necessarily defined to include all aspects of the invention. It should be appreciated that the various concepts and embodiments described above, as well as those described in greater detail below, may be implemented in any of numerous ways, as the disclosed concepts and embodiments are not limited to any one implementation. In addition, some aspects of the present disclosure may be used alone, or in any suitable combination with other aspects of the present disclosure.

Although the present invention has been described with reference to the preferred embodiments, it is not intended to be limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, the protection scope of the present invention should be determined by the appended claims.

Claims (10)

1. A plant protection unmanned aerial vehicle cluster operation task distribution method is characterized by comprising the following steps:

s1, dividing the operation field into N sub-fields according to the input field parameters and the plant protection unmanned aerial vehicle parameters, wherein the operation requirements of each sub-field are the same, and only corresponding to one unmanned aerial vehicle, the set of N sub-fields is defined as H_j，j＝1,2,3,…,N；

S2, distributing all the sub-fields to all the plant protection unmanned aerial vehicles participating in the operation according to a preset distribution principle, wherein the preset distribution principle is as follows: a single plant protection unmanned aerial vehicle is allocated with at least one sub-field block, each sub-field block can be allocated only once, and all the sub-field blocks are allocated;

s3, setting the number of unmanned planes, the number of field blocks K and the number of sub-field blocks in the plant protection unmanned planeQuantity N, and the operation time matrix (T)_K×NAnd the distance matrix of the in/out field is (L)¹)_K×N、(L²)_K×NBased on the operation, an MOSE L A algorithm is adopted to find an optimal operation distribution matrix (X)_K×N；

S4, distributing matrix (X) according to the optimal operation_K×NAnd the aircraft operation time matrix (T)_K×NDetermining an operation time matrix (T') of the sub-field block_N×1(ii) a Defining a sequence of operations matrix (SE) for the sub-fields in combination with predetermined assembly rules_N×1The preset assembly rule is that after each unmanned aircraft lands, assembly tasks including battery replacement and medicine chest replacement are carried out, but only one unmanned aircraft executes the assembly tasks at the same time;

s5, adopting genetic algorithm to work sequence matrix (SE) of the sub-farmland blocks_N×1And (6) optimizing.

2. The plant protection unmanned aerial vehicle cluster operation task allocation method according to claim 1, wherein in step S1, the process of dividing the operation field into N sub-fields according to the input field parameters and plant protection unmanned aerial vehicle parameters comprises the following steps:

s11, putting the fields with the same operation requirements into the same set, wherein the number of the fields is M;

s12, further field processing is carried out on all the fields in the M sets according to the one-time lifting/landing operation area of the plant protection unmanned aerial vehicle, and the field processing comprises the operations of adjacent small field merging and large field dividing;

s13, extracting all the sub-field blocks in the M sets, setting the number of the sub-field blocks to be N, and defining the N sub-field block sets to be H_j，j＝1,2,3,…,N。

3. The method for assigning task tasks to cluster plant protection unmanned aerial vehicle as claimed in claim 1, wherein in step S1 and step S2, the assigning all sub-fields to all plant protection unmanned aerial vehicles participating in operation according to the preset assignment rule is:

defining the operation distribution matrix of plant unmanned plane as (X)_K×NWherein a parameter x is assigned_ijThe assignment is defined as:

where j is 1,2,3, …, N.

4. The plant protection unmanned aerial vehicle cluster task allocation method of claim 3, wherein in step S1 and step S3, the MOSE L A algorithm is used to find the optimal task allocation matrix (X)_K×NComprises the following steps:

s31, inputting population scale S, sub-population number S, individual frog num of the population, total population iteration number D and population iteration number I;

s32, initializing the frog group P ═ X¹,X²,…,X^S}；

S33, calculating adaptive values, arranging the adaptive values in a descending order, and dividing the group;

s34, intra-population evolution: each sub-population makes a first jump according to the following formula:

wherein: x^bAnd X^wRespectively, the optimal individual and the worst individual of the sub-group;

s35, merging of first population: and (4) performing descending arrangement according to the adaptive value, and combining the following formula, and performing second jump on the whole group:

wherein: x^gAnd X^wRespectively, the optimal individual and the worst individual of the whole population;

s36, second population merging: and (4) performing descending arrangement according to the adaptive value, and performing third jump on the weakest individuals of the population by combining the following formula:

(X^w)′＝rand(X),X∈Q^M×N

wherein: rand (X) represents randomly generating a decision matrix in the feasible solution space;

s37, judging whether the current cycle times are less than the total group iteration times D, if so, turning to the step S33, otherwise, outputting the optimal individual X¹；

Wherein, the related parameters of population update are defined as follows:

let matrix X¹∈Q^M×N、X²∈Q^M×NWherein Q is^M×NFor a feasible solution space, a matrix is defined

Represents X²And X¹The matrix of the differences between the two,

indicating an ectopic or operational point between matrices;

the meaning of z ═ cut (z, k) is: intercepting the front k columns of the one-dimensional array z and storing the front k columns into a new one-dimensional array z';

(X¹) ' -F (B, Φ) means: updating an approximation matrix B for the matrix X according to the difference matrix B, counting element sums of all columns of B, if the element sums are larger than 0, storing the column number into a one-dimensional array z, and defining an operator phi as follows:

φ＝～(B,X¹,cut(z,ri(n)))

wherein: n is the length of the array z, ri (n) denotes the interval [1, n ]]Any integer within; "-" indicates that X is assigned to the column number ri (n) designated by z¹Element correction in the column is consistent with the arrangement of the B element;

the meaning of (A) is: for matrix X¹According to the operator

Performing mutation operation, operator

Is defined as:

in the formula: switch (r)₁,r₂) Representing r in a switching matrix₁Row and r₂Line, swicth (c)₁,c₂) Representing c in the switching matrix₁Column sum c₂And (4) columns.

5. The plant protection unmanned aerial vehicle cluster task allocation method of claim 4, wherein in step S1, in step S33, the process of performing cluster division comprises the steps of:

s331, defining an adaptive value function FV of an individual t₁As follows below, the following description will be given,

s332, taking the adaptive value of the individual as a reference for dividing the group according to FV₁ ^tIn descending order, the individuals in the ith subgroup are:

{FV₁ ^k|k∈[(i-1)×num+1,i×num]}

wherein i is 1,2,3, …, s.

6. The method of claim 4, wherein in step S1 and step S34, the process of evolving inside the population comprises the following steps:

s341, jumping to the optimal individual of the sub-population according to the following formula:

s342, introducing a mutation operator for detecting whether the plant protection unmanned aerial vehicle which is not distributed to the field exists, and performing mutation detection and processing by adopting the mutation operator:

(X^w)″＝F((X^w)′,switch([r₀,r_m],c))

in the formula: r is₀Representation matrix (X)^w) ' line element and line number 0; r is_mRepresentation matrix (X)^w) ' line element sum is the largest line number; c represents r_mAny column number with a median of 1; switch ([ r ]₀,r_m]And c) is represented in matrix (X)^w) In column c of `, r is exchanged₀And r_mWhen r is an element of₀When the value of (A) is not unique, a jump formula corresponding to the first jump is adopted for carrying out multiple variations;

s343, updating the optimal and worst individuals;

and S344, repeating the steps S341 to S343 until the cycle number reaches I times.

7. The method for assigning task tasks to cluster unmanned aerial vehicle for plant protection as claimed in claim 4, wherein in step S1, step S35, the process of merging the first population comprises the following steps:

s351, jumping to the population-optimized individual using the following formula:

s352, introducing a mutation operator for detecting whether the plant protection unmanned aerial vehicle which is not distributed to the field exists, and performing mutation detection and processing by adopting the mutation operator:

(X^w)″＝F((X^w)′,switch([r₀,r_m],c))

in the formula: r is₀Representation matrix (X)^w) ' line element and line number 0; r is_mRepresentation matrix (X)^w) ' line element sum is the largest line number; c represents r_mAny column number with a median of 1; switch ([ r ]₀,r_m]And c) is represented in matrix (X)^w) In column c of `, r is exchanged₀And r_mWhen r is an element of₀When the value of (A) is not unique, a jump formula corresponding to the first jump is adopted for carrying out multiple variations;

and S353, updating the optimal and worst individuals.

8. The plant protection unmanned aerial vehicle cluster task allocation method of claim 4, wherein in step S4, the operation sequence matrix (SE) defining the sub-fields_N×1Comprises the following steps:

s41, according to the airplane operation time matrix, the time matrix is (T)_K×NDetermining a time matrix (T') of the operation of the field_N×1，t_j' is:

wherein j is 1,2,3, …, N;

s42, defining the operation sequence matrix of the field block as (SE)_N×1，se_jIs [1, N ]]A non-repeating set of integers within the interval;

s42, according to the operation Sequence (SE)_N×1The time matrix after defining the sequence is

It is composed of

Comprises the following steps:

s43, based on (SE)_N×1And

defining the operation time of the plant protection unmanned plane-field block as (TA)_K×N’，ta_ij′Indicates the time of j ' th operation of the ith aircraft, j ' is 1,2,3, …, N ' is max { sum [ x (i,:)]}；(TA)_K×N’Each column of the non-assigned items is completed by 0;

s44, considering the overlapping time of assembling batteries and medicine boxes simultaneously on a plurality of plant protection unmanned planes, defining the landing time matrix of the unmanned planes as (T L)_K×N’Takeoff time matrix is (TR)_K×N’，tl_ij′And tr_ij′Respectively as follows:

wherein, ts_ij′The time required for the aircraft to fit to the next takeoff after landing, here set to a fixed length, ts_ij′＝c。

9. The method for assigning task tasks to cluster unmanned aerial vehicle for plant protection as claimed in claim 8, wherein in step S5, the application of genetic algorithm to operation sequence matrix (SE) of sub-field block_N×1The process of performing the optimization includes the steps of:

s51, evaluating the adaptive value of the sub-field operation:

order of operation (SE)_N×1As chromosome, in determining (SE)_N×1Under the condition of (3), determining the final landing time of different plant protection unmanned aerial vehicles to be tl_KN’Defining an fitness value selection function FV of a chromosome₂Comprises the following steps:

wherein, g (SE)^t)＝max[tl_KN′(SE^t,c)](ii) a t represents the position of the chromosome in the population; wt is a time adaptive value parameter, the unit is min, and the value range is [300,600,]；

s52, genetic operator definition:

a. chromosome selection operations

Selecting initial population by roulette method, if fitness of a certain individual t is f_tThen the probability that the individual is selected is denoted as P_t：

Wherein N represents the chromosome scale;

b. chromosome crossing operations

Carrying out chromosome crossing operation by adopting a partial matching crossing method: randomly generating two cross points and determining a matching section; generating two sub-individuals according to the mapping relation between the matching segments and the non-matching segments in the parent;

c. operation of chromosomal variation

Randomly selecting N positions, and interchanging genes corresponding to the N positions, wherein N is more than 1;

s53, updating chromosome based on GA algorithm to obtain optimal operation sequence matrix (SE) of sub-field_N×1The method comprises the following substeps:

s531, inputting population size N, maximum iteration times max (Gen) of evolution, and cross rate p_cThe rate of variation p_m；

S532, generating an initial population { SE }_PThe total population size is P;

s533, calculating the adaptive value of the population individuals;

s534, determining the following conditions for stopping genetic calculation: (1) the iteration number reaches max (Gen), (2) the requirement of an adaptive value is met, (3) the optimal chromosome is continuously kept for 10 generations; if any one of the conditions is satisfied, the iteration process is terminated, and the optimal individual { SE } is output_max(f)Otherwise, go to step S535;

s535, selecting and calculating other chromosomes except the optimal chromosome;

s536, performing cross operation on other chromosomes except the optimal chromosome;

s537, performing mutation operation on other chromosomes except the optimal chromosome, and returning to step S533.

10. The plant protection unmanned aerial vehicle cluster operation task distribution device is characterized by comprising an ARM embedded system, and a GPS positioning system, a temperature and humidity sensor, a data transmission module, a power supply module, an L CD touch screen and a cache module which are respectively connected with the ARM embedded system;

the temperature and humidity sensor is used for detecting the real-time temperature and the real-time humidity of the operation area and feeding back the detection result to the ARM embedded system;

the L CD touch screen is used for inputting field parameters and plant protection unmanned aerial vehicle parameters to the ARM embedded system, and the field parameters comprise boundary coordinates of each field;

the ARM embedded system combines input field parameters, plant protection unmanned aerial vehicle parameters and environment information fed back by a temperature and humidity sensor, adopts the distribution method as any one of claims 1 to 9, jointly calculates all plant protection unmanned aerial vehicle cooperative task distribution models and corresponding scheduling schemes of plant protection unmanned aerial vehicle cluster operation, sends the scheduling schemes to the flight control system of each plant protection unmanned aerial vehicle through a data transmission module, and stores the input parameters and the corresponding scheduling schemes to a cache module;

the power supply module is used for providing electric energy required by normal work of the ARM embedded system, the GPS positioning system, the temperature and humidity sensor, the data transmission module, the power supply module, the L CD touch screen and the buffer module;

the GPS is used for acquiring the position information of each plant protection unmanned aerial vehicle and displaying the acquired position information to a user through an L CD touch screen.

Images