Method and device for newly adding search task in autonomous auction of unmanned aerial vehicle group

Abstract

The invention provides a new search task method and device for autonomous auction of an unmanned aerial vehicle group. The method comprises the following steps: step 1, numbering each member in the unmanned aerial vehicle group in a range of 1-M; step 2, loading an initial task area position database {Dj}, j = 1,..., N; step 3, loading the longitude TOLoni and the like of a take-off point; step 4, a bidding time threshold value THSubmit and a decision time threshold value THDecide are loaded; step 5, if the information of the newly added task area DN + 1 is received, calculating a bid bid Picei and sending the bid bid Picei to other unmanned aerial vehicle members in the unmanned aerial vehicle group; step 6, if the time for entering the step 5 is greater than the bidding time threshold THSubmit, performing a preliminary decision according to the bidding bid {Picej} (j = 1,..., M) of the unmanned aerial vehicle and the other unmanned aerial vehicle members in the group, and sending a preliminary decision result Resulti to the other unmanned aerial vehicle members in the group; and step 7, for each member i (i = 1,..., M) in the unmanned aerial vehicle group, if the time for entering the step 6 is greater than the decision time threshold THDecide, performing a final decision according to the initial decision result {Resultj}, j = 1,..., M of the local unmanned aerial vehicle and the received other unmanned aerial vehicle members in the group.

Description

Technical Field

[0001] The present invention belongs to flight control technology, and in particular relates to a method and device for autonomously auctioning new search tasks by a group of unmanned aerial vehicles. Background Art

[0002] For drone swarms, a blanket search of unknown areas is a common task requirement. Drones can establish a comprehensive understanding of the environmental situation and enemy deployment through carpet reconnaissance in the mission area. In this mission scenario, drones have no prior knowledge support, and task allocation can be regarded as a drone resource optimization and task scheduling problem. On the basis of the task allocation of the drone swarm, if there are new tasks, online reallocation is required.

[0003] The autonomous task auction of drone swarms is a distributed task allocation algorithm, which is particularly suitable for the online allocation of new search tasks. After the auction is started, each member of the drone swarm can decide which member will take on the task by comparing prices and negotiating. Compared with the centralized task allocation method, the advantage of this method is that it distributes the computing requirements among the onboard computing resources (such as the onboard computers of each drone), which improves the computing efficiency and is particularly suitable for the online task allocation of large-scale drone swarms.

[0004] However, this method also has shortcomings in practical applications. The most important one is that it has high requirements for the quality of inter-machine communication. If there is loss or delay in inter-machine communication data, it is easy to lead to inconsistent decision-making results of each drone member, which may cause serious consequences such as chaos in the swarm. For example, drones A and B both have the ability to undertake new tasks, and drone B is better at undertaking the task, but drone A did not receive the bid from drone B due to communication reasons, so it mistakenly believed that it should undertake the task. At this time, drones A and B have inconsistent decisions, which may lead to a collision between A and B. Therefore, it is necessary to design a method for drone swarms to autonomously auction new search tasks to ensure decision consistency. Summary of the invention

[0005] The purpose of the present invention is to provide a method and device for autonomously auctioning new search tasks for a drone swarm, introduce a bidding time threshold and a decision time threshold, and a dual decision mechanism, so as to overcome the problem of inconsistent decisions caused by communication data delay, loss, etc. in actual applications of existing task auction methods.

[0006] The technical solution of the present invention:

[0007] In a first aspect, the present application provides a method for autonomously auctioning a new search task by a drone swarm, the method comprising:

[0008] Step 1: number each member in the drone group in the range of 1 to M, where M is the number of drone members;

[0009] Step 2: For each member i (i = 1, ..., M) in the drone group, load the initial mission area location database {D j}, j = 1, ..., N; wherein N is the number of initial task areas;

[0010] Step 3: For each member i (i=1,…,M) in the drone group, load the take-off point longitude TOLon i , take-off point latitude TOLat i , optimal cruising speed V i , load scanning field of view width W i , member load capacity C i , member initial task sequence T i ;

[0011] Step 4: For each member i (i=1,…,M) in the drone group, load the bidding time threshold THSubmit and the decision time threshold THDecide;

[0012] Step 5: For each member i (i = 1, ..., M) in the drone group, if a new task area D is received N+1 If the member is able to undertake the task, the bid price is calculated. i And send it to other drone members in the drone group;

[0013] Step 6: For each member i (i = 1, ..., M) in the drone group, if the time to enter step 5 is greater than the bidding time threshold THSubmit, then the bid price {Price j},j=1,…,M makes a preliminary decision and sends the preliminary decision result Result i Send to other drone members in the group; if the bid of a drone member k is not received, its bid Price k Set to positive infinity;

[0014] Step 7: For each member i (i = 1, ..., M) in the drone group, if the time to enter step 6 is greater than the decision time threshold THDecide, then the initial decision result {Result j}, j = 1, ..., M makes the final decision; if the preliminary decision result of a drone member k is not received, the preliminary decision result Result k Set to 0.

[0015] Furthermore, the calculated bid price in step 5 i The process includes:

[0016] Step 51: For each original and newly added task area D j ,j=1,…,N+1, based on the longitude and latitude of the take-off point (TOLon i ,TOLat i ) is the coordinate origin, and D j The longitude and latitude of each vertex k Convert to the coordinates of the geodetic coordinate system

[0017] Step 52: For each original and newly added task area D j ,j=1,…,N+1, according to the coordinates in the geodetic coordinate system Calculate D j Area DS j ;

[0018] Step 53: For each original and newly added task area D j ,j=1,…,N+1, calculate D j The coordinates of the center of gravity in the geodetic coordinate system (DGX j ,DGY j ):

[0019]

[0020] Step 54: Pre-set the first new task sequence According to the task numbers in the task sequence, a task area centroid matrix and a task area area matrix are formed;

[0021] Step 55: According to the task area matrix and the task area centroid matrix, the execution task sequence T is calculated according to the following formula: i1 The time required, i.e. the bid price i (1):

[0022]

[0023] Step 56: Repeat steps 54 to 55 to insert the newly added task number N+1 into the initial task sequence T i 2nd~TN i Before the task, the 2nd to TN i A new task sequence to calculate the bid price i (2)~Price i (TN i ), and insert into TN i After the task, the TNi +1 new task sequence, calculate the bid price i (TN i +1); if TN i =0, then directly insert the new task number N+1 into the initial task sequence T i The first item in the list is sufficient;

[0024] Step 57: Compare Bids {Price i (j),j=1,…,TN i +1}, select the minimum value as the bid price of this member i , record the corresponding task sequence as the optimal task sequence TB i .

[0025] Further, step 52 includes:

[0026] Step 521: According to the coordinates in the geodetic coordinate system Get the matrix

[0027]

[0028] Step 522: According to DM j calculate

[0029]

[0030] Step 523: According to Calculate D j Area DS j ,

[0031] Further, step 54 includes:

[0032] Step 541: Based on each original and newly added task area D j Area DS j and Forming a task area matrix

[0033] Step 542: Based on each original and newly added task area D j The coordinates of the center of gravity in the geodetic coordinate system (DGX j ,DGY j )and Forming the Task Area Gravity Matrix

[0034]

[0035] Furthermore, the preliminary decision method in step 6 is: select the first one among j=1,…,M such that {Price j The smallest j is recorded as the initial decision result Result i , which is the drone member number with the smallest bid decided by drone member i.

[0036] Furthermore, the final decision method in step 7 is: compare {Result j}, j = 1, ..., M, if the non-zero results exist and are the same, the decision is considered consistent and taken as the final decision result; if the non-zero results exist and are not the same, the smallest non-zero result is taken as the final decision result; if the non-zero result does not exist, the decision is considered to have failed; compare the drone number with the final decision result, if they are consistent, it is considered that the drone member has won the task through the auction, and then execute the optimal task sequence TB recorded in step 5 i .

[0037] Furthermore, in step 2:

[0038] The jth initial task area D j is a polygon with no less than 3 and no more than 8 sides, represented by It is stated For mission area D j The k-th vertex position, including the vertex longitude and latitude The DN j For mission area D j The number of vertices is in the range [3,8].

[0039] Furthermore, in step 3:

[0040] The load capacity of the i-th member is C i The value range is 1 or 2, 1 means that the member payload has the ability to search for vehicles and larger targets, and 2 means that the member payload has the ability to search for people and larger targets.

[0041] The initial task sequence T of the i-th member i It consists of several tasks, represented as It is stated is the number of the jth mission area undertaken by UAV member i. The third mission undertaken by UAV crew member 1 is mission area D2; the TN i is the number of mission areas undertaken by UAV member i, which can be 0.

[0042] Furthermore, in step 4:

[0043] The bidding time threshold THSubmit is equal to the sum of the calculation period, the inter-machine communication period, and the inter-machine communication maximum delay × 2;

[0044] The decision time threshold THDecide is equal to the sum of the inter-machine communication cycle and the inter-machine communication maximum delay×2.

[0045] In a second aspect, the present application provides a device for autonomously auctioning new search tasks for a swarm of drones, and the device is used to implement the above-mentioned method for autonomously auctioning new search tasks for a swarm of drones.

[0046] The present invention has the following technical features:

[0047] (1) For new search tasks, drone swarms can achieve task allocation through autonomous auctions, and obtain consistent auction results without human intervention;

[0048] (2) The results of autonomous auctions by drone swarm members are made consistent through the following means: introducing bidding time thresholds and decision time thresholds. Drones must reach the time thresholds before they can take the next action, eliminating the impact of communication delays and communication asynchrony; introducing a dual decision-making mechanism to resolve the inconsistency of initial decisions caused by communication delays and losses through final decisions;

[0049] (3) This method is simple to implement, easy to operate, and has practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A flowchart of a method for autonomously auctioning new search tasks by a drone swarm provided by the present invention. DETAILED DESCRIPTION

[0051] The present invention proposes a method for autonomously auctioning new search tasks for drone swarms, introducing a bidding time threshold and a decision time threshold, as well as a dual decision mechanism, to overcome the problem of inconsistent decisions caused by communication data delay, loss, etc. in the existing task auction method in practical applications. The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.

[0052] like Figure 1 As shown, the present application provides a method for autonomously auctioning a new search task by a drone swarm, comprising:

[0053] Step 1: number each member in the drone group in the range of 1 to M; wherein M is the number of drone members.

[0054] Step 2: For each member i (i = 1, ..., M) in the drone group, load the initial mission area location database {D j}, j = 1, ..., N; wherein N is the number of initial task areas.

[0055] The jth initial task area D j is a polygon with no less than 3 and no more than 8 sides, represented by It is stated For mission area D j The k-th vertex position, including the vertex longitude and latitude The DN j For mission area D j The number of vertices is in the range [3,8].

[0056] Step 3: For each member i (i=1,…,M) in the drone group, load the take-off point longitude TOLon i , take-off point latitude TOLat i , optimal cruising speed V i , load scanning field of view width W i , member load capacity C i , member initial task sequence T i .

[0057] The load capacity of the i-th member is C i The value range is 1 or 2, 1 means that the member payload has the ability to search for vehicles and larger targets, and 2 means that the member payload has the ability to search for people and larger targets.

[0058] The initial task sequence T of the i-th member i It consists of several tasks, represented as It is stated is the number of the jth mission area undertaken by UAV member i. The third mission undertaken by UAV crew member 1 is mission area D2; the TN i is the number of mission areas undertaken by UAV member i, which can be 0.

[0059] Step 4: For each member i (i=1,…,M) in the drone swarm, load the bidding time threshold THSubmit and the decision time threshold THDecide.

[0060] The bidding time threshold THSubmit should be equal to the sum of the calculation period, the inter-machine communication period, and the maximum inter-machine communication delay × 2, and the decision time threshold THDecide should be equal to the sum of the inter-machine communication period and the maximum inter-machine communication delay × 2; wherein, the calculation period is 0.01s, and the inter-machine communication period and the maximum inter-machine communication delay are determined by the communication equipment, and are generally 0.1s.

[0061] Step 5: For each member i (i = 1, ..., M) in the drone group, if a new task area D is received N+1 If the member is able to undertake the task, the bid price is calculated. i And send it to other drone members in the drone group.

[0062] The newly added task area D N+1 The information comes from the ground station, including the payload capacity requirements of the new mission area. N+1 and vertex set The CN N+1 The value is 1 or 2. 1 means that the area requires the UAV crew to have the ability to search for vehicles and larger targets, and 2 means that the area requires the UAV crew to have the ability to search for people and larger targets. For mission area D N+1 The k-th vertex position, including the vertex longitude and latitude The DN N+1 For mission area D N+1 The number of vertices is in the range [3,8].

[0063] The member capability is derived from the onboard computer of the drone member, including the member's flight capability and mission capability; wherein the flight capability is 0 or 1, 0 represents no flight capability, 1 represents flight capability; the mission capability is 0 or 1, 0 represents no mission capability, 1 represents mission capability.

[0064] The method for evaluating whether the member can undertake the mission is as follows: if the member has the flying capability and mission capability, and the load capacity C i Not less than the load capacity requirement of the newly added mission area CN N+1 , then it is considered that the member can undertake the new task, otherwise it is considered that the member cannot undertake the new task.

[0065] The bid price i It refers to the time required for member i to complete the original tasks and the new tasks.

[0066] Specifically, the bid price calculated in step 5 is i The process includes:

[0067] Step 51: For each original and newly added task area D j ,j=1,…,N+1, based on the longitude and latitude of the take-off point (TOLon i ,TOLati ) is the coordinate origin, and D j The longitude and latitude of each vertex k Convert to the coordinates of the geodetic coordinate system

[0068] Step 52: For each original and newly added task area D j ,j=1,…,N+1, according to the coordinates in the geodetic coordinate system Calculate D j Area DS j ;

[0069] More specifically, step 52 includes:

[0070] Step 521: According to the coordinates in the geodetic coordinate system Get the matrix

[0071]

[0072] Step 522: According to DM j calculate

[0073]

[0074] Step 523: According to Calculate D j Area DS j ,

[0075] Step 53: For each original and newly added task area D j ,j=1,…,N+1, calculate D j The coordinates of the center of gravity in the geodetic coordinate system (DGX j ,DGY j ):

[0076]

[0077] Step 54: Pre-set the first new task sequence According to the task numbers in the task sequence, a task area centroid matrix and a task area area matrix are formed;

[0078] Among them, the first new task sequence is pre-set The method is:

[0079] Insert the new task number N+1 into the initial task sequence T i Before the first task in .

[0080] More specifically, step 54 includes:

[0081] Step 541: Based on each original and newly added task area D j Area DS j and Forming a task area matrix

[0082] Step 542: Based on each original and newly added task area D j The coordinates of the center of gravity in the geodetic coordinate system (DGX j ,DGY j )and Forming the Task Area Gravity Matrix

[0083]

[0084] Step 55: According to the task area matrix and the task area centroid matrix, the execution task sequence T is calculated according to the following formula: i1 The time required, i.e. the bid price i (1):

[0085]

[0086] Step 56: Repeat steps 54 to 55 to insert the newly added task number N+1 into the initial task sequence T i 2nd~TN i Before the task, the 2nd to TN i A new task sequence to calculate the bid price i (2)~Price i (TN i ), and insert into TN i After the task, the TN i +1 new task sequence, calculate the bid price i (TN i +1); if TN i =0, then directly insert the new task number N+1 into the initial task sequence T i The first item in the list is sufficient;

[0087] Step 57: Compare Bids {Price i (j),j=1,…,TN i +1}, select the minimum value as the bid price of this member i , record the corresponding task sequence as the optimal task sequence TB i .

[0088] Step 6: For each member i (i = 1, ..., M) in the drone group, if the time to enter step 5 is greater than the bidding time threshold THSubmit, then the bid price {Price j},j=1,…,M makes a preliminary decision and sends the preliminary decision result Result i Send to other drone members in the group; if the bid of a drone member k is not received, its bid Price k Set to positive infinity.

[0089] The preliminary decision method is: select the first one among j=1,…,M such that {Price j The smallest j is recorded as the initial decision result Result i , which is the drone member number with the smallest bid decided by drone member i.

[0090] Step 7: For each member i (i = 1, ..., M) in the drone group, if the time to enter step 6 is greater than the decision time threshold THDecide, then the initial decision result {Result j}, j = 1, ..., M makes the final decision; if the preliminary decision result of a drone member k is not received, the preliminary decision result Result k Set to 0.

[0091] The final decision method is: Compare {Result j}, j = 1, ..., M, if the non-zero results exist and are the same, the decision is considered consistent and taken as the final decision result; if the non-zero results exist and are not the same, the smallest non-zero result is taken as the final decision result; if the non-zero result does not exist, the decision is considered to have failed; compare the drone number with the final decision result, if they are consistent, it is considered that the drone member has won the task through the auction, and then execute the optimal task sequence TB recorded in step 5 i .

[0092] Example 1

[0093] Step 1: M=3, the drones are numbered 1 to 3.

[0094] Step 2: N=0, the initial mission area location database is empty.

[0095] Step 3: TOLon i =108.60093709°,TOLat i =34.15587989°, V i =20m / s,Wi = 250m, C1 = 1, C2 = 2, C3 = 2, the initial task sequence T i is empty.

[0096] Step 4: THSubmit = 0.31s, THDecide = 0.3s.

[0097] Step 5: The new task area is D1, CN1 = 2, D1 is a hexagon, and the vertices DP1 k (k = 1, …, 6) latitude and longitude are shown in the following table:

[0098]

[0099]

[0100] Member 1 has flight ability and task ability, but C1 < CN1, so it cannot undertake task D1 and thus cannot send a tender offer.

[0101] Member 2 has flight ability and task ability, and C2 ≥ CN1, so it can undertake task D1. The process of calculating the tender offer is as follows:

[0102]

[0103] DS1 = 523620.45

[0104] DGX1 = 769.15, DGY1 = 7060.13

[0105] Price2(1) = 459.82

[0106] Price2 = 459.82

[0107] Member 2 sends the tender offer Price2 = 459.82 to other members in the UAV swarm.

[0108] The judgment and calculation processes of Member 3 and Member 2 are similar, and Member 3 sends the tender offer Price3 = 459.82 to other members in the UAV swarm.

[0109] Step 6: For Members 1 - 3, Price1 = +∞, Price2 = 459.82, Price3 = 459.82, and the preliminary decision result Result1 = 2.

[0110] Step 7: For Members 1 - 3, Result1 = 2, Result2 = 2, Result3 = 2. It is determined that the decision results are consistent. Member 2 undertakes the new task D1, and Members 1 and 3 do not undertake the new task D1.

[0111] Embodiment 2

[0112] Steps 1 to 5: Same as Example 1.

[0113] Step 6: For member 1 and member 2, Price1=+∞, Price2=459.82, Price3=459.82, and the preliminary decision result Result1=2; for member 3, due to communication failure, the bid of member 2 was not received, Price1=+∞, Price2=+∞, Price3=459.82, and the preliminary decision result Result1=3.

[0114] Step 7: For members 1 to 3, Result1 = 2, Result2 = 2, Result3 = 3, and the decision results are determined to be inconsistent. The final decision result is 2, and member 2 undertakes the new task D1. Members 1 and 3 do not undertake the new task D1.

Claims

1. A method for autonomous auction of new search tasks by drone swarms, characterized in that: Methods include: Step 1: number each member in the drone group in the range of 1 to M, where M is the number of drone members; Step 2: For each member i (i = 1, ..., M) in the drone group, load the initial mission area location database {D j }, j = 1, ..., N; wherein N is the number of initial task areas; Step 3: For each member i (i=1,…,M) in the drone group, load the take-off point longitude TOLon i , take-off point latitude TOLat i , optimal cruising speed V i , load scanning field of view width W i , member load capacity C i , member initial task sequence T i ; Step 4: For each member i (i=1,…,M) in the drone group, load the bidding time threshold THSubmit and the decision time threshold THDecide; Step 5: For each member i (i = 1, ..., M) in the drone group, if a new task area D is received N+1 If the member is able to undertake the task, the bid price is calculated. i And send it to other drone members in the drone group; Step 6: For each member i (i=1,…,M) in the drone group, if the time to enter step 5 is greater than the bidding time threshold THSubmit, a preliminary decision is made based on the bid price {Pricej}, j=1,…,M received from the local drone and other drone members in the group, and the preliminary decision result Result is displayed. i Send to other drone members in the group; if the bid of a drone member k is not received, its bid Price k Set to positive infinity; Step 7: For each member i (i = 1, ..., M) in the drone group, if the time to enter step 6 is greater than the decision time threshold THDecide, then the initial decision result {Result j }, j = 1, ..., M makes the final decision; if the preliminary decision result of a drone member k is not received, the preliminary decision result Result k Set to 0.

2. The method according to claim 1, characterized in that: Calculate the bid price in step 5 i The process includes: Step 51: For each original and newly added task area D j ,j=1,…,N+1, based on the longitude and latitude of the take-off point (TOLon i ,TOLat i ) is the coordinate origin, and D j The longitude and latitude of each vertex k Convert to the coordinates of the geodetic coordinate system Step 52: For each original and newly added task area D j ,j=1,…,N+1, according to the coordinates in the geodetic coordinate system Calculate D j Area DS j ; Step 53: For each original and newly added task area D j ,j=1,…,N+1, calculate D j The coordinates of the center of gravity in the geodetic coordinate system (DGX j ,DGY j ): Step 54: Pre-set the first new task sequence According to the task numbers in the task sequence, a task area centroid matrix and a task area area matrix are formed; Step 55: According to the task area matrix and the task area centroid matrix, the execution task sequence T is calculated according to the following formula: i1 The time required, i.e. the bid price i (1): Step 56: Repeat steps 54 to 55 to insert the newly added task number N+1 into the initial task sequence T i 2nd~TN i Before the task, the 2nd to TN i A new task sequence to calculate the bid price i (2)~Price i (TN i ), and insert into TN i After the task, the TN i +1 new task sequence, calculate the bid price i (TN i +1); if TN i =0, then directly insert the new task number N+1 into the initial task sequence T i The first item in the list is sufficient; Step 57: Compare Bids {Price i (j),j=1,…,TN i +1}, select the minimum value as the bid price of this member i , record the corresponding task sequence as the optimal task sequence TB i .

3. The method according to claim 2, characterized in that Step 52 includes: Step 521: According to the coordinates in the geodetic coordinate system Get the matrix Step 522: According to DM j calculate Step 523: According to Calculate D j Area DS j , 4. The method according to claim 2, characterized in that: Step 54 includes: Step 541: Based on each original and newly added task area D j Area DS j and Forming a task area matrix Step 542: Based on each original and newly added task area D j The coordinates of the center of gravity in the geodetic coordinate system (DGX j ,DGY j )and Forming the task area center of gravity matrix 5. The method according to claim 1, characterized in that The preliminary decision method in step 6 is: select the first j that minimizes {Pricej} among j=1,…,M, and record it as the preliminary decision result Result i , which is the drone member number with the smallest bid decided by drone member i.

6. The method according to claim 1, characterized in that The final decision method in step 7 is: Compare {Result j }, j = 1, ..., M, if non-zero results exist and are the same, the decision is considered consistent and is taken as the final decision result; If non-zero results exist and are not the same, take the smallest non-zero result as the final decision result; If there is no non-zero result, the decision is considered to have failed; compare the drone number with the final decision result. If they are consistent, it is considered that the drone member has won the task at the auction, and then execute the optimal task sequence TB recorded in step 5 i .

7. The method according to claim 1, characterized in that In step 2: The jth initial task area D j is a polygon with no less than 3 and no more than 8 sides, represented by It is stated For mission area D j The k-th vertex position, including the vertex longitude and latitude The DN j For mission area D j The number of vertices is in the range [3,8].

8. The method according to claim 1, characterized in that In step 3: The load capacity of the i-th member is C i The value range is 1 or 2, 1 means that the member payload has the ability to search for vehicles and larger targets, and 2 means that the member payload has the ability to search for people and larger targets. The initial task sequence T of the i-th member i It consists of several tasks, represented as It is stated is the number of the jth mission area undertaken by UAV member i. The third mission undertaken by UAV crew member 1 is mission area D2; the TN i is the number of mission areas undertaken by UAV member i, which can be 0.

9. The method according to claim 1, characterized in that: In step 4: The bidding time threshold THSubmit is equal to the sum of the calculation period, the inter-machine communication period, and the inter-machine communication maximum delay × 2; The decision time threshold THDecide is equal to the sum of the inter-machine communication cycle and the inter-machine communication maximum delay×2.

10. A drone swarm autonomous auction for new search tasks, characterized in that: The device is used to implement the method for autonomously auctioning new search tasks by a swarm of drones as described in claim 1.

Images