Unmanned aerial vehicle group scheduling method and device based on brain-like calculation

Abstract

The invention discloses an unmanned aerial vehicle group scheduling method and device based on brain-like computing, the method is realized by using a central computing server and computing nodes, the computing nodes are arranged on each unmanned aerial vehicle, and the method comprises the following steps: each unmanned aerial vehicle in an unmanned aerial vehicle group receives and obtains a to-be-completed task information set; the computing node of each unmanned aerial vehicle carries out fusion evaluation processing on the to-be-completed task information set and an own technical state information set to obtain a task participation information set and an updated technical state information set; each unmanned aerial vehicle sends the task participation information set and the updated technical state information set to a central computing server; the central computing server performs resource allocation computing processing on the received all task participation information set and the updated technical state information set to obtain unmanned aerial vehicle resource allocation information; the unmanned aerial vehicle resource allocation information comprises task information participated by each unmanned aerial vehicle.

Description

Technical Field

[0001] The present invention relates to the fields of brain-inspired computing and unmanned cluster control processing, and specifically to a method and device for scheduling unmanned aerial vehicle (UAV) swarms based on brain-inspired computing. Background Art

[0002] With the rapid development of drone technology, drone swarms have been widely used in many fields such as logistics and distribution, agricultural plant protection, and disaster relief. However, in large-scale drone swarm applications, how to efficiently and in real time dispatch them to complete complex tasks has become a key issue that needs to be addressed.

[0003] Traditional drone swarm scheduling methods typically rely on centralized computing platforms, which suffer from high computational latency, limited communication bandwidth, and excessive dependence on central nodes. Furthermore, existing methods often lack flexibility and adaptability when dealing with complex and changing mission environments and emergencies. Summary of the Invention

[0004] The present invention mainly solves the problem of how to flexibly and quickly allocate and schedule drone clusters. The present invention discloses a drone cluster scheduling method and device based on brain-like computing.

[0005] In a first aspect, an embodiment of the present invention discloses a method for scheduling a swarm of drones based on brain-inspired computing, which is implemented using a central computing server and computing nodes, wherein the computing nodes are provided on each drone. The method includes:

[0006] S1, each drone in the drone swarm receives a set of information about the task to be completed;

[0007] S2, the computing node of each UAV performs a fusion evaluation process on the pending task information set and its own technical status information set to obtain a corresponding task participation information set and an updated technical status information set; the task participation information set includes a set of sequence numbers of the UAV's participating tasks and a set of sequence numbers of the UAV's non-participating tasks;

[0008] S3, each UAV sends its corresponding mission participation information set and updated technical status information set to the central computing server;

[0009] S4, the central computing server performs resource allocation calculation processing on all received task participation information sets and updated technical status information sets to obtain drone resource allocation information; the drone resource allocation information includes the task information in which each drone participates.

[0010] The computing nodes are implemented using a computing platform based on a brain-like chip;

[0011] The set of information about pending tasks includes a sequence number and a set of required index values for each pending task; the set of required index values includes a communication bandwidth requirement value, a communication channel number requirement value, a communication remaining duration requirement value, a communication frequency band requirement value, a flight time requirement value, a flight altitude requirement value, a positioning accuracy requirement value, a positioning distance requirement value, and a remaining computing resource requirement value;

[0012] The computing node of each UAV performs a fusion evaluation process on the to-be-completed task information set and its own technical status information set to obtain a corresponding task participation information set and an updated technical status information set, including:

[0013] S21, the computing node of each UAV collects a set of technical status information of the UAV; the set of technical status information includes a channel bandwidth information sequence, a communication channel number information sequence, a communication remaining duration information sequence, a communication frequency band information sequence, a flight endurance information sequence, a flight altitude information sequence, a positioning accuracy information sequence, a positioning distance information sequence, and a remaining computing resource information sequence;

[0014] S22, performing usability evaluation processing on the technical status information set to obtain a usability evaluation value;

[0015] S23, determining whether the availability evaluation value is greater than a set evaluation threshold, and obtaining a first determination result; if the first determination result is greater than, executing S24; if the first determination result is not greater than, determining that the participating task sequence number set in the task participation information set is empty, determining that the non-participating task sequence number set in the task participation information set is the sequence number of all tasks, and not updating the technical status information set;

[0016] S24, obtaining a set of upper limit values for drone resource capabilities; constructing a single-machine resource allocation model using the set of upper limit values for drone resource capabilities, the set of technical status information of the drone, and the set of information about pending tasks; the set of upper limit values for drone resource capabilities includes an upper limit value for channel bandwidth, an upper limit value for the number of communication channels, an upper limit value for communication duration, an upper limit value for communication frequency band, an upper limit value for flight time, an upper limit value for flight altitude, an upper limit value for positioning accuracy, an upper limit value for positioning distance, and an upper limit value for computing resources;

[0017] S25, solving the single-machine resource allocation model to obtain a resource allocation vector;

[0018] S26: Based on the resource allocation vector, construct a mission participation information set and an updated technical status information set of the UAV.

[0019] The performing usability evaluation processing on the technical status information set to obtain a usability evaluation value includes:

[0020] S221, obtaining a standard value for each type of technical status information;

[0021] S222, subtracting each information sequence of the technical status information set from the standard value of the corresponding technical status information to obtain a corresponding difference information sequence;

[0022] S223, constructing a difference matrix using all difference information sequences;

[0023] S224, performing singular value decomposition on the difference matrix to obtain a left decomposition matrix, a singular matrix, and a right decomposition matrix;

[0024] S225, respectively extracting the diagonal vectors of the left decomposition matrix, the singular matrix, and the right decomposition matrix, and concatenating and fusing the diagonal vectors of the three matrices to obtain a fused diagonal vector;

[0025] S226, performing Gram angular field transform on the fused diagonal vector to obtain a corresponding two-dimensional image matrix;

[0026] S227, performing cross-correlation calculation on the two-dimensional image matrix to obtain a corresponding cross-correlation matrix;

[0027] S228, performing transformation feature calculation on the cross-correlation matrix and the two-dimensional image matrix to obtain a transformation matrix;

[0028] S229: Calculate the fluctuation characteristics of the transformation matrix to obtain an availability evaluation value.

[0029] The expression for calculating the transformation feature is:

[0030] Z=(R T R) -1 R T E,

[0031] Where E is the two-dimensional image matrix, R is the cross-correlation matrix, and Z is the transformation matrix;

[0032] The expression for calculating the fluctuation characteristics is:

[0033]

[0034] Among them, conj() means finding the conjugate, Z i and Z j They represent the i-th row vector and the j-th row vector of the transformation matrix, respectively, || represents the modulus value, ρ is the rank value of the transformation matrix, FFT represents the fast Fourier transform, z ij Represents the element in the i-th row and j-th column of the transformation matrix, i ijrepresents the complex cross-correlation value between the i-th row vector and the j-th row vector of the transformation matrix, M and N are the column dimension and row dimension of the transformation matrix respectively, and ky is the availability evaluation value.

[0035] The expression of the single-machine resource allocation model is:

[0036]

[0037] Where CV(l) is the objective function, l is the resource allocation vector to be solved, I represents the total number of technical status information, J represents the total number of tasks to be completed, K represents the total number of elements contained in the information sequence of the technical status information set, and zy ik represents the kth element of the i-th information sequence of the technical status information set, l j represents the jth element of the resource allocation vector to be solved, l j When the value is 1, it means that the drone participates in the task to be completed with sequence number j. j When the value is 0, it means that the drone does not participate in the task to be completed with sequence number j. ji represents the i-th demand indicator value of the j-th task to be completed, z i0 Represents the i-th upper limit value of the drone resource capability upper limit value set.

[0038] The step of constructing a task participation information set and an updated technical status information set of the UAV based on the resource allocation vector includes:

[0039] S261, using the sequence number of each element in the resource allocation vector with a value of 1 as the sequence number of the drone's participating mission; using the sequence number of each element in the resource allocation vector with a value of 0 as the sequence number of the drone's non-participating mission;

[0040] S262: Using all participating task serial numbers, construct a participating task serial number set in the task participating information set of the UAV; using all non-participating task serial numbers, construct a non-participating task serial number set in the task participating information set of the UAV;

[0041] S263, performing update processing on each information sequence of the technical status information set to obtain an updated information sequence;

[0042] S264: Utilize all updated information sequences to construct an updated technical status information set.

[0043] The central computing server performs resource allocation calculation processing on all received task participation information sets and updated technical status information sets to obtain UAV resource allocation information, including:

[0044] S41, finding the union of the non-participating task sequence number sets in all received task participation information sets to obtain a set of task sequence numbers to be assigned;

[0045] S42, constructing a group resource allocation model based on the set of sequence numbers of the tasks to be assigned and the updated technical status information set of all drones;

[0046] S43, solving the group resource allocation model to obtain drone resource allocation information.

[0047] In a second aspect of an embodiment of the present invention, a drone swarm scheduling device based on brain-inspired computing is disclosed, the device comprising:

[0048] a memory storing executable program code;

[0049] a processor coupled to the memory;

[0050] The processor calls the executable program code stored in the memory to execute the drone swarm scheduling method based on brain-like computing.

[0051] According to a third aspect of an embodiment of the present invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions. When the computer instructions are called by a computer, they are used to execute the drone swarm scheduling method based on brain-like computing.

[0052] In a fourth aspect of an embodiment of the present invention, an information data processing terminal is disclosed, which is used to implement the drone swarm scheduling method based on brain-like computing.

[0053] The beneficial effects of the present invention are:

[0054] By deploying brain-inspired chips and a distributed architecture across unmanned clusters, this approach significantly reduces computational latency and communication overhead, enabling fast and efficient scheduling decisions and reducing reliance on central nodes. The distributed computing and communication model reduces reliance on a single central node, improving the reliability and fault tolerance of the entire system. This approach dynamically allocates drone resources based on their real-time status and mission requirements, maximizing resource utilization.

[0055] In the drone swarm scheduling method based on brain-inspired computing provided by an embodiment of the present invention, the computing node of each drone is implemented based on a brain-inspired chip. By leveraging the powerful parallel processing and adaptive learning capabilities of brain-inspired computing, it can quickly and accurately collect and process its own technical status information set and pending task information set. When performing availability assessment processing on the technical status information set, a series of complex and effective mathematical transformations and calculation methods such as singular value decomposition and Gram angular field transform are used to deeply analyze the characteristics of the technical status information from multiple dimensions to obtain accurate availability assessment values. Compared with traditional assessment methods, this can more comprehensively and accurately judge the working status of the drone, avoiding task allocation errors caused by inaccurate assessments.

[0056] During task allocation, a single-machine resource allocation model is constructed and solved based on availability assessment results and a set of upper limits for drone resource capabilities. This model fully considers the actual technical status and mission requirements of the drones, enabling customized mission participation plans for each drone. This ensures that drones are optimally assigned to tasks that match their capabilities, effectively avoiding resource waste and mission failures. A central computing server comprehensively processes the mission participation information sets and updated technical status information sets for all drones, achieving global resource optimization and further improving the collaborative operation capabilities and overall efficiency of the drone swarm. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 4 is an implementation flow chart of the method of the present invention. DETAILED DESCRIPTION

[0058] In order to better understand the content of the present invention, an embodiment is given here.

[0059] Figure 1 4 is an implementation flow chart of the method of the present invention.

[0060] In a first aspect, an embodiment of the present invention discloses a method for scheduling a swarm of drones based on brain-inspired computing, which is implemented using a central computing server and computing nodes, wherein the computing nodes are provided on each drone. The method includes:

[0061] S1, each drone in the drone swarm receives a set of information about pending tasks; the set of information about pending tasks includes a sequence number and a set of requirement index values for each pending task; the set of requirement index values includes a communication bandwidth requirement value, a communication channel number requirement value, a communication remaining duration requirement value, a communication frequency band requirement value, a flight time requirement value, a flight altitude requirement value, a positioning accuracy requirement value, a positioning distance requirement value, and a remaining computing resource requirement value;

[0062] S2, the computing node of each UAV performs a fusion evaluation process on the pending task information set and its own technical status information set to obtain a corresponding task participation information set and an updated technical status information set; the task participation information set includes a set of sequence numbers of the UAV's participating tasks and a set of sequence numbers of the UAV's non-participating tasks;

[0063] S3, each UAV sends the mission participation information set and the updated technical status information set to the central computing server;

[0064] S4, the central computing server performs resource allocation calculation processing on all received task participation information sets and updated technical status information sets to obtain drone resource allocation information; the drone resource allocation information includes the task information in which each drone participates.

[0065] The computing nodes can be implemented using a computing platform based on brain-like chips;

[0066] The brain-like chip can be implemented using Tianjic chip, PAICore chip, etc.

[0067] The brain-like chip can be implemented using a RISC-V computing platform based on the brain-like chip.

[0068] The application of brain-like chips gives this scheduling method greater adaptability and intelligence. It can adjust the scheduling strategy in real time according to task requirements and dynamic changes in drone status in complex and changing operating environments, ensuring that the drone swarm always maintains efficient operation, greatly expanding the application potential of drone swarms in various complex scenarios.

[0069] The computing node of each UAV performs a fusion evaluation process on the to-be-completed task information set and its own technical status information set to obtain a task participation information set and an updated technical status information set, including:

[0070] S21, the computing node of each UAV collects a set of technical status information of the UAV; the set of technical status information includes a channel bandwidth information sequence, a communication channel number information sequence, a communication remaining duration information sequence, a communication frequency band information sequence, a flight endurance information sequence, a flight altitude information sequence, a positioning accuracy information sequence, a positioning distance information sequence, and a remaining computing resource information sequence;

[0071] S22, performing usability evaluation processing on the technical status information set to obtain a usability evaluation value;

[0072] S23, determining whether the availability evaluation value is greater than a set evaluation threshold, and obtaining a first determination result; if the first determination result is greater than, executing S24; if the first determination result is not greater than, determining that the participating task sequence number set in the task participation information set is empty, determining that the non-participating task sequence number set in the task participation information set is the sequence number of all tasks, and not updating the technical status information set;

[0073] S24, obtaining a set of upper limit values for drone resource capabilities; constructing a single-machine resource allocation model using the set of upper limit values for drone resource capabilities, the set of technical status information of the drone, and the set of information about pending tasks; the set of upper limit values for drone resource capabilities includes an upper limit value for channel bandwidth, an upper limit value for the number of communication channels, an upper limit value for communication duration, an upper limit value for communication frequency band, an upper limit value for flight time, an upper limit value for flight altitude, an upper limit value for positioning accuracy, an upper limit value for positioning distance, and an upper limit value for computing resources;

[0074] S25, solving the single-machine resource allocation model to obtain a resource allocation vector;

[0075] S26, constructing a mission participation information set and an updated technical status information set of the UAV based on the resource allocation vector;

[0076] The performing usability evaluation processing on the technical status information set to obtain a usability evaluation value includes:

[0077] Obtain standard values for each type of technical status information;

[0078] Subtracting each information sequence of the technical status information set from the standard value of the corresponding technical status information to obtain a corresponding difference information sequence;

[0079] Using all the difference information sequences, a difference matrix is constructed;

[0080] Performing singular value decomposition on the difference matrix to obtain a left decomposition matrix, a singular matrix, and a right decomposition matrix;

[0081] The diagonal vectors of the left decomposition matrix, the singular matrix and the right decomposition matrix are extracted respectively, and the diagonal vectors of the three matrices are concatenated and fused to obtain the fused diagonal vectors;

[0082] Performing a Gram angular field transform on the fused diagonal vector to obtain a corresponding two-dimensional image matrix;

[0083] Performing cross-correlation calculation on the two-dimensional image matrix to obtain a corresponding cross-correlation matrix;

[0084] Calculate the transformation features of the cross-correlation matrix and the two-dimensional image matrix to obtain a transformation matrix;

[0085] Calculating the fluctuation characteristics of the transformation matrix to obtain an availability evaluation value;

[0086] The expression for calculating the transformation feature is:

[0087] Z=(R T R) -1 R T E,

[0088] Among them, E is the two-dimensional image matrix, R is the cross-correlation matrix, and Z is the transformation matrix.

[0089] The expression for calculating the fluctuation characteristics is:

[0090]

[0091]

[0092] Among them, conj() means finding the conjugate, Z i and Z j They represent the i-th row vector and the j-th row vector of the transformation matrix, respectively, || represents the modulus value, ρ is the rank value of the transformation matrix, FFT represents the fast Fourier transform, z ij Represents the element in the i-th row and j-th column of the transformation matrix, u ij represents the complex cross-correlation value between the i-th row vector and the j-th row vector of the transformation matrix, M and N are the column dimension and row dimension of the transformation matrix respectively, and ky is the availability evaluation value.

[0093] In traditional drone swarm scheduling, the processing of drone technical status information is relatively rough, making it difficult to explore the deep features behind the data, resulting in an inability to accurately evaluate drone availability. The transformation feature calculation expression in the present invention achieves deep extraction of technical status information features by constructing a mathematical connection between the cross-correlation matrix R and the two-dimensional image matrix E. The cross-correlation matrix R can reflect the correlation between different technical status information dimensions, and the two-dimensional image matrix E is a visual expression of the technical status information after the Gram angular field transformation. On this basis, the transformation matrix Z calculated by this expression can project the original technical status information into a new feature space, remove redundant information in the data, and at the same time enhance features that are valuable for availability evaluation. Compared with traditional methods, this calculation method can more accurately characterize the essential characteristics of drone technical status information, provide a more representative data basis for subsequent availability evaluation, avoid evaluation errors caused by insufficient feature extraction, and thus improve the accuracy of drone status judgment before task assignment.

[0094] Traditional UAV technical status assessment methods often lack consideration of the dynamic changes and interrelationships of data, making it difficult to capture subtle fluctuations in technical status, and easily causing the assessment results to be out of touch with the actual situation. The fluctuation characteristic calculation expression analyzes the transformation matrix Z from two dimensions: frequency domain and correlation, through fast Fourier transform (FFT) and complex cross-correlation calculation. Fast Fourier transform can convert technical status data in the time domain to the frequency domain, revealing the hidden periodic fluctuations and frequency components in the data, and helping to discover the changing patterns of UAV technical status at different frequencies; complex cross-correlation calculation quantifies the similarity between different row vectors of the transformation matrix, reflecting the coordinated change relationship between various technical status indicators. On this basis, by calculating the difference zz ij -ρu ij The maximum cumulative value is then taken to obtain the availability assessment value, ky. This process fully considers the dynamic and multi-dimensional correlation of technical status information, and is able to sensitively capture subtle fluctuations and abnormal changes in the drone's technical status. Compared with traditional evaluation methods, this expression can more comprehensively and meticulously assess drone availability, providing a more reliable basis for task allocation decisions. It effectively avoids task allocation failures caused by inaccurate technical status assessments, and improves the reliability and stability of the drone swarm scheduling system.

[0095] The element in the i-th row and j-th column of the cross-correlation matrix is the cross-correlation value between the i-th row vector and the j-th row vector of the two-dimensional image matrix;

[0096] The calculation expression of the singular value processing is:

[0097] X=UTV,

[0098] Among them, U is the left decomposition matrix, X is the difference matrix, T is the singular matrix, V is the right decomposition matrix, U and V are both orthogonal matrices, and T is a diagonal matrix;

[0099] The categories of the technical status information include channel bandwidth information, number of communication channels, remaining communication duration information, communication frequency band information, flight time information, flight altitude information, positioning accuracy information, positioning distance information, and remaining computing resource information;

[0100] The expression of the single-machine resource allocation model is:

[0101]

[0102] Where CV(l) is the objective function, l is the resource allocation vector to be solved, I represents the total number of technical status information, J represents the total number of tasks to be completed, K represents the total number of elements contained in the information sequence of the technical status information set, and zy ik represents the kth element of the i-th information sequence of the technical status information set, lj represents the jth element of the resource allocation vector to be solved, l j When the value is 1, it means that the drone participates in the task to be completed with sequence number j. j When the value is 0, it means that the drone does not participate in the task to be completed with sequence number j. ji represents the i-th demand indicator value of the j-th task to be completed, z i0 represents the i-th upper limit value of the set of upper limit values of the drone resource capability;

[0103] The first to ninth requirement indicator values of the task to be completed are respectively the communication bandwidth requirement value, the communication channel number requirement value, the communication remaining duration requirement value, the communication frequency band requirement value, the flight time requirement value, the flight altitude requirement value, the positioning accuracy requirement value, the positioning distance requirement value, and the remaining computing resource requirement value;

[0104] The first to ninth information sequences of the technical status information set are, respectively, a channel bandwidth information sequence, a communication channel number information sequence, a communication remaining duration information sequence, a communication frequency band information sequence, a flight time information sequence, a flight altitude information sequence, a positioning accuracy information sequence, a positioning distance information sequence, and a remaining computing resource information sequence;

[0105] The first to ninth upper limits of the UAV resource capability upper limit set are, respectively, the channel bandwidth upper limit, the communication channel number upper limit, the communication duration upper limit, the communication frequency band upper limit, the endurance time upper limit, the flight altitude upper limit, the positioning accuracy upper limit, the positioning distance upper limit, and the computing resource upper limit;

[0106] Solving the single-machine resource allocation model to obtain the resource allocation vector may be performed by using a numerical solution algorithm or a swarm optimization algorithm, such as a bee colony algorithm.

[0107] The step of constructing a task participation information set and an updated technical status information set of the UAV based on the resource allocation vector includes:

[0108] S261, using the sequence number of each element in the resource allocation vector with a value of 1 as the sequence number of the drone's participating mission; using the sequence number of each element in the resource allocation vector with a value of 0 as the sequence number of the drone's non-participating mission;

[0109] S262: Using all participating task serial numbers, construct a participating task serial number set in the task participating information set of the UAV; using all non-participating task serial numbers, construct a non-participating task serial number set in the task participating information set of the UAV;

[0110] S263, performing update processing on each information sequence of the technical status information set to obtain an updated information sequence;

[0111] S264: Utilize all updated information sequences to construct an updated technical status information set.

[0112] The expression of the update process is:

[0113]

[0114] in, The kth element of the i-th information sequence is updated.

[0115] The central computing server performs resource allocation calculation processing on all received task participation information sets and updated technical status information sets to obtain UAV resource allocation information, including:

[0116] S41, finding the union of the non-participating task sequence number sets in all received task participation information sets to obtain a set of task sequence numbers to be assigned;

[0117] S42, constructing a group resource allocation model based on the set of sequence numbers of the tasks to be assigned and the updated technical status information set of all drones;

[0118] S43, solving the group resource allocation model to obtain drone resource allocation information.

[0119] The expression of the group resource allocation model is:

[0120]

[0121] Among them, ZCV() is the group resource benefit function, F represents the drone resource allocation matrix to be solved, B1 is the total number of drones in the drone group, J1 is the total number of elements in the set of task numbers to be assigned, wzy bik represents the kth element of the i-th information sequence of the updated technical status information set of the b-th UAV, I represents the total number of technical status information, K represents the total number of elements contained in the information sequence of the technical status information set, and f bj represents the element in the bth row and jth column of the UAV resource allocation matrix to be solved, f bj When the value is 1, it means that the b-th UAV participates in the task to be assigned with the sequence number j in the set of task sequence numbers to be assigned, f bj When the value is 0, it means that the b-th UAV does not participate in the task to be assigned with the sequence number j in the set of task sequence numbers to be assigned. ji It represents the i-th demand index value of the task to be assigned with sequence number j in the set of task sequence numbers to be assigned.

[0122] Solving the group resource allocation model to obtain drone resource allocation information includes:

[0123] Using a numerical solution algorithm to solve the group resource allocation model, obtain a UAV resource allocation matrix, and confirm that the UAV resource allocation matrix is UAV resource allocation information;

[0124] The numerical solution algorithm may be an optimization algorithm, an LMS algorithm, a particle filter algorithm, or the like.

[0125] Each drone in the drone swarm includes a communication module, which is used to send information to the central computing server and receive mission information at the same time.

[0126] The central computing server is implemented using a GPU server;

[0127] The computing nodes are implemented using various edge computing platforms, including NVIDIA's Jetson, Raspberry Pi computing platform, brain-like computing platform, etc.

[0128] In a second aspect of an embodiment of the present invention, a drone swarm scheduling device based on brain-inspired computing is disclosed, the device comprising:

[0129] a memory storing executable program code;

[0130] a processor coupled to the memory;

[0131] The processor calls the executable program code stored in the memory to execute the drone swarm scheduling method based on brain-like computing.

[0132] According to a third aspect of an embodiment of the present invention, a computer-storable medium is disclosed, wherein the computer-storable medium stores computer instructions. When the computer instructions are called by a computer, they are used to execute the drone swarm scheduling method based on brain-like computing.

[0133] In a fourth aspect of an embodiment of the present invention, an information data processing terminal is disclosed, which is used to implement the drone swarm scheduling method based on brain-like computing.

[0134] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A drone swarm scheduling method based on brain-inspired computing, characterized in that: This is achieved using a central computing server and computing nodes, where the computing nodes are located on each drone. The method includes: S1, each drone in the drone swarm receives a set of information about the task to be completed; S2, the computing node of each UAV performs a fusion evaluation process on the pending task information set and its own technical status information set to obtain a corresponding task participation information set and an updated technical status information set; the task participation information set includes a set of sequence numbers of the UAV's participating tasks and a set of sequence numbers of the UAV's non-participating tasks; S3, each UAV sends its corresponding mission participation information set and updated technical status information set to the central computing server; S4, the central computing server performs resource allocation calculation processing on all received task participation information sets and updated technical status information sets to obtain drone resource allocation information; the drone resource allocation information includes the task information in which each drone participates.

2. The method for dispatching drone swarms based on brain-inspired computing according to claim 1, characterized in that: The computing nodes are implemented using a computing platform based on a brain-like chip; The set of information about pending tasks includes a sequence number and a set of required index values for each pending task; the set of required index values includes a communication bandwidth requirement value, a communication channel number requirement value, a communication remaining duration requirement value, a communication frequency band requirement value, a flight time requirement value, a flight altitude requirement value, a positioning accuracy requirement value, a positioning distance requirement value, and a remaining computing resource requirement value; The computing node of each UAV performs a fusion evaluation process on the to-be-completed task information set and its own technical status information set to obtain a corresponding task participation information set and an updated technical status information set, including: S21, the computing node of each UAV collects a set of technical status information of the UAV; the set of technical status information includes a channel bandwidth information sequence, a communication channel number information sequence, a communication remaining duration information sequence, a communication frequency band information sequence, a flight endurance information sequence, a flight altitude information sequence, a positioning accuracy information sequence, a positioning distance information sequence, and a remaining computing resource information sequence; S22, performing usability evaluation processing on the technical status information set to obtain a usability evaluation value; S23, determining whether the availability evaluation value is greater than a set evaluation threshold, and obtaining a first determination result; if the first determination result is greater than, executing S24; if the first determination result is not greater than, determining that the participating task sequence number set in the task participation information set is empty, determining that the non-participating task sequence number set in the task participation information set is the sequence number of all tasks, and not updating the technical status information set; S24, obtaining a set of upper limit values for drone resource capabilities; constructing a single-machine resource allocation model using the set of upper limit values for drone resource capabilities, the set of technical status information of the drone, and the set of information about pending tasks; the set of upper limit values for drone resource capabilities includes an upper limit value for channel bandwidth, an upper limit value for the number of communication channels, an upper limit value for communication duration, an upper limit value for communication frequency band, an upper limit value for flight time, an upper limit value for flight altitude, an upper limit value for positioning accuracy, an upper limit value for positioning distance, and an upper limit value for computing resources; S25, solving the single-machine resource allocation model to obtain a resource allocation vector; S26: Based on the resource allocation vector, construct a mission participation information set and an updated technical status information set of the UAV.

3. The method for dispatching drone swarms based on brain-inspired computing according to claim 2, wherein: The performing usability evaluation processing on the technical status information set to obtain a usability evaluation value includes: S221, obtaining a standard value for each type of technical status information; S222, subtracting each information sequence of the technical status information set from the standard value of the corresponding technical status information to obtain a corresponding difference information sequence; S223, constructing a difference matrix using all difference information sequences; S224, performing singular value decomposition on the difference matrix to obtain a left decomposition matrix, a singular matrix, and a right decomposition matrix; S225, respectively extracting the diagonal vectors of the left decomposition matrix, the singular matrix, and the right decomposition matrix, and concatenating and fusing the diagonal vectors of the three matrices to obtain a fused diagonal vector; S226, performing Gram angular field transform on the fused diagonal vector to obtain a corresponding two-dimensional image matrix; S227, performing cross-correlation calculation on the two-dimensional image matrix to obtain a corresponding cross-correlation matrix; S228, performing transformation feature calculation on the cross-correlation matrix and the two-dimensional image matrix to obtain a transformation matrix; S229: Calculate the fluctuation characteristics of the transformation matrix to obtain an availability evaluation value.

4. The method for dispatching drone swarms based on brain-inspired computing according to claim 3, wherein: The expression for calculating the transformation feature is: Z=(R T R) -1 R T E, Where E is the two-dimensional image matrix, R is the cross-correlation matrix, and Z is the transformation matrix; The expression for calculating the fluctuation characteristics is: u ij =|(FFT(Z i ×conj(Z j )))| 2 , Among them, conj() means finding the conjugate, Z i and Z j They represent the i-th row vector and the j-th row vector of the transformation matrix, respectively, || represents the modulus value, ρ is the rank value of the transformation matrix, FFT represents the fast Fourier transform, z ij Represents the element in the i-th row and j-th column of the transformation matrix, u ij represents the complex cross-correlation value between the i-th row vector and the j-th row vector of the transformation matrix, M and N are the column dimension and row dimension of the transformation matrix respectively, and ky is the availability evaluation value.

5. The method for dispatching drone swarms based on brain-inspired computing according to claim 2, wherein: The expression of the single-machine resource allocation model is: Where CV(l) is the objective function, l is the resource allocation vector to be solved, I represents the total number of technical status information, J represents the total number of tasks to be completed, K represents the total number of elements contained in the information sequence of the technical status information set, and zy ik represents the kth element of the i-th information sequence of the technical status information set, l j represents the jth element of the resource allocation vector to be solved, l j When the value is 1, it means that the drone participates in the task to be completed with sequence number j. j When the value is 0, it means that the drone does not participate in the task to be completed with sequence number j. ji represents the i-th demand indicator value of the j-th task to be completed, z i0 Represents the i-th upper limit value of the drone resource capability upper limit value set.

6. The method for dispatching drone swarms based on brain-inspired computing according to claim 2, wherein: The step of constructing a task participation information set and an updated technical status information set of the UAV based on the resource allocation vector includes: S261, using the sequence number of each element in the resource allocation vector with a value of 1 as the sequence number of the drone's participating mission; using the sequence number of each element in the resource allocation vector with a value of 0 as the sequence number of the drone's non-participating mission; S262: Using all participating task serial numbers, construct a participating task serial number set in the task participating information set of the UAV; using all non-participating task serial numbers, construct a non-participating task serial number set in the task participating information set of the UAV; S263, performing update processing on each information sequence of the technical status information set to obtain an updated information sequence; S264: Utilize all updated information sequences to construct an updated technical status information set.

7. The method for dispatching drone swarms based on brain-inspired computing according to claim 1, wherein: The central computing server performs resource allocation calculation processing on all received task participation information sets and updated technical status information sets to obtain UAV resource allocation information, including: S41, finding the union of the non-participating task sequence number sets in all received task participation information sets to obtain a set of task sequence numbers to be assigned; S42, constructing a group resource allocation model based on the set of sequence numbers of the tasks to be assigned and the updated technical status information set of all drones; S43, solving the group resource allocation model to obtain drone resource allocation information.

8. A drone swarm scheduling device based on brain-like computing, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the drone swarm scheduling method based on brain-like computing as described in any one of claims 1 to 7.

9. A computer storable medium, characterized in that The computer storable medium stores computer instructions, which, when called by a computer, are used to execute the drone swarm scheduling method based on brain-like computing as described in any one of claims 1 to 7.

10. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the drone swarm scheduling method based on brain-like computing as described in any one of claims 1 to 7.

Images