CN-122018561-A - Unmanned aerial vehicle flight control method, equipment, medium and product based on multi-alliance game complete distributed Nash equilibrium search

CN122018561ACN 122018561 ACN122018561 ACN 122018561ACN-122018561-A

Abstract

The application discloses an unmanned aerial vehicle flight control method, equipment, medium and product based on multi-alliance game complete distributed Nash equilibrium search, relating to the field of multi-alliance game Nash equilibrium search, wherein the method comprises the steps of acquiring observation state data in real time; the method comprises the steps of establishing an unmanned aerial vehicle power model, obtaining a communication topological graph of a alliance where a current unmanned aerial vehicle is located, adopting a multi-alliance game fully-distributed Nash equilibrium search algorithm, iterating state data in the current unmanned aerial vehicle flight process based on observation state data, the communication topological graph and the unmanned aerial vehicle power model until the state data of the current unmanned aerial vehicle reaches a state stability requirement, enabling a multi-alliance game system to reach Nash equilibrium, and controlling the unmanned aerial vehicle to fly according to control input corresponding to the state data meeting the state stability requirement. The application can be applied to a multi-league gaming system containing non-cooperative leagues and a scene that related global information is difficult to acquire.

Inventors

YAN RUI
CHEN RUIJIE
HU CHENXI
HUA YONGCHAO
DONG XIWANG
LI XIAODUO
FENG ZHI
PAN CHENGWEI

Assignees

北京航空航天大学

Dates

Publication Date: 20260512
Application Date: 20260203

Claims (10)

1. The unmanned aerial vehicle flight control method based on the multi-alliance game full-distributed Nash equilibrium search is characterized by being applied to each unmanned aerial vehicle in a multi-alliance game system, wherein the multi-alliance game system comprises a plurality of alliances, each alliance comprises a plurality of unmanned aerial vehicles, and the method comprises the following steps: acquiring state data of other unmanned aerial vehicles except the current unmanned aerial vehicle in the multi-alliance game system observed by the current unmanned aerial vehicle in real time, and taking the state data as observed state data; constructing an unmanned aerial vehicle dynamic model; acquiring a communication topological graph of a alliance where the current unmanned aerial vehicle is located; Adopting a multi-alliance game complete distributed Nash equilibrium search algorithm, and iterating the state data in the current unmanned aerial vehicle flight process based on the observed state data, the communication topological graph and the unmanned aerial vehicle dynamic model until the state data of the current unmanned aerial vehicle reaches the state stability requirement; and controlling the unmanned aerial vehicle to fly according to the control input corresponding to the state data meeting the state stability requirement.
2. The unmanned aerial vehicle flight control method based on the multi-league game fully-distributed nash equilibrium search according to claim 1, wherein the iterative method for the state data in the current unmanned aerial vehicle flight process based on the observed state data, the communication topological graph and the unmanned aerial vehicle dynamic model is adopted by adopting a multi-league game fully-distributed nash equilibrium search algorithm and comprises the following steps: acquiring state data of the current unmanned aerial vehicle in real time, wherein the state data comprises position and speed; Initializing state parameters of the current unmanned aerial vehicle, wherein the state parameters comprise distributed observer parameters, bias parameters, actuator deviation fault compensation item parameters and gradient item parameters; determining a topology matrix based on the communication topology map; Adopting a multi-alliance game fully-distributed Nash equilibrium search algorithm, and iterating the state data of the current unmanned aerial vehicle based on the state parameters, the observed state data, the topology matrix and the unmanned aerial vehicle dynamic model of the current unmanned aerial vehicle; The process of iterating the state data of the current unmanned aerial vehicle by adopting the multi-alliance game fully-distributed Nash equilibrium search algorithm comprises the following steps: Determining a state observation value based on the topology matrix, the observation state data and distributed observer parameters by adopting a distributed observer; determining an observation position vector based on the state observations; Determining a first auxiliary variable based on the observation position vector, the topology matrix and the bias parameters; Determining an actuator deviation fault compensation term parameter and a gradient term parameter based on the first auxiliary variable and the current unmanned aerial vehicle speed; determining new state data based on a current unmanned aerial vehicle speed, a first auxiliary variable, an actuator deviation fault compensation term parameter, a gradient term parameter and the unmanned aerial vehicle dynamic model; Determining whether the state stability requirement is met or not based on state data obtained by two times of iteration of the current unmanned aerial vehicle; if the state stability requirement is met, stopping iteration; If the state stability requirement is not met, determining new distributed observer parameters based on the state observation values, determining new bias guide parameters based on the first auxiliary variables, and returning to execute the step of adopting the distributed observer to determine the state observation values based on the topology matrix, the observation state data and the distributed observer parameters.
3. The unmanned aerial vehicle flight control method based on the full-distributed nash equilibrium search of the multi-league game according to claim 1, wherein the unmanned aerial vehicle dynamics model is expressed as: ; In the formula, And Representation of Time alliance Middle (f) The position and speed of the individual unmanned aerial vehicle, Representation of Time alliance Middle (f) The actuator saturation function of the individual unmanned aerial vehicle, Is that Time alliance Middle (f) An ideal control input for the individual unmanned aerial vehicle, Representative of Time alliance Middle (f) The actuator deviation of the individual unmanned aerial vehicle fails, Representing alliances Middle (f) The efficiency coefficient of the actuator of the individual unmanned aerial vehicle, Representing alliances Is a number of unmanned aerial vehicles.
4. The unmanned aerial vehicle flight control method of claim 2, wherein determining a state observation based on the topology matrix, the observation state data, and distributed observer parameters using a distributed observer comprises: Using the formula Determining a derivative of the state observation; Integrating the derivative to obtain a state observation value; Wherein, the Representing alliances Middle (f) The derivative of the state observations of the individual drones, The representation is composed of Pressing the button A diagonal array of tensors in which And the diagonal elements are respectively from top left to bottom right ; Is the first In the individual alliance Third party in personal unmanned aerial vehicle and multi-alliance game system Inter-group observation item adaptive parameters between unmanned aerial vehicles, The representation is composed of Pressing the button A diagonal array of the sheets is formed, Representing alliances Middle (f) Observed by unmanned aerial vehicle The speed of the individual unmanned aerial vehicle, ; Representing alliances Middle (f) A state observation value observed by the unmanned aerial vehicle, and is determined according to the state observation data, Representing column vectors formed by the positions and speed states of all unmanned aerial vehicles in the multi-league gaming system; Represent the first In the individual alliance Unmanned aerial vehicle and the first The self-adaptive parameters of the interaction items in the group among the unmanned aerial vehicles; represented by the first Matrix of personal unmanned aerial vehicle Pressing the button A diagonal block matrix formed by stretching, and ; Represented by the first Input matrix of personal unmanned aerial vehicle Pressing the button A diagonal block matrix formed by stretching, and ; Representing alliances In the topology matrix Line 1 The elements of the column are arranged such that, Representing a second order identity matrix, Representing the kronecker product of the two, As a standard sign function.
5. The unmanned aerial vehicle flight control method based on the full distributed nash equilibrium search of the multi-league games according to claim 2, wherein the distributed observer parameters include inter-group observation item adaptation parameters and intra-group interaction item adaptation parameters; determining new distributed observer parameters based on the state observations, comprising: Using the formula Determining the derivative of the adaptive parameter of the new inter-group observation term; Using the formula Determining a derivative of the new intra-group interaction term adaptation parameter; integrating the derivatives respectively to obtain new inter-group observation item self-adaptive parameters and new intra-group interaction item self-adaptive parameters; Wherein, the Represent the first In the individual alliance Third party in personal unmanned aerial vehicle and multi-alliance game system The derivatives of the inter-group observation term adaptation parameters between the individual unmanned aerial vehicles, Representing alliances Middle (f) Observed by unmanned aerial vehicle The speed of the individual unmanned aerial vehicle, Represent the first In the individual alliance Unmanned plane pair The state estimation vector of the individual drone, Represent the first The state vector of the individual drone, Represent the first An input matrix of the individual unmanned aerial vehicle, In order to take the modulus operator, Is 1 norm; 、 Setting parameters; Represent the first In the individual alliance Unmanned aerial vehicle and the first The derivatives of the adaptive parameters of the intra-group interaction term between the individual unmanned aerial vehicles, Representing alliances In the topology matrix Line 1 Elements of a column; Representing alliances Middle (f) The state observation value observed by the unmanned aerial vehicle is determined according to the state observation data; represented by the first Input matrix of personal unmanned aerial vehicle Pressing the button A matrix of diagonal blocks.
6. The unmanned aerial vehicle flight control method based on the multi-alliance game fully-distributed Nash equilibrium search of claim 2, wherein the requirement for state stability is met when the difference value of state data obtained by two iterations of the unmanned aerial vehicle is smaller than a set threshold value.
7. The unmanned aerial vehicle flight control method based on the multi-league game fully-distributed nash equilibrium search according to claim 1, wherein the communication topological graph is a UNICOM undirected graph.
8. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the unmanned aerial vehicle flight control method of any one of claims 1-7 based on a multi-league gaming fully distributed nash equilibrium search.
9. A computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements the unmanned aerial vehicle flight control method of any one of claims 1-7 based on a multi-league gaming fully distributed nash equilibrium search.
10. A computer program product comprising a computer program which, when executed by a processor, implements the unmanned aerial vehicle flight control method of any one of claims 1-7 based on a multi-league gambling fully distributed nash equilibrium search.

Description

Unmanned aerial vehicle flight control method, equipment, medium and product based on multi-alliance game complete distributed Nash equilibrium search Technical Field The application relates to the field of multi-alliance game Nash equilibrium search, in particular to an unmanned aerial vehicle flight control method, equipment, medium and product based on multi-alliance game complete distributed Nash equilibrium search. Background Aiming at the characteristics of high efficiency, flexibility, difficult defense and the like of cluster combat, counter measures for clusters become important factors for guaranteeing regional safety. However, the countermeasures based on laser and microwave weapons depend on the use scene and have defects in terms of killing precision, so that cluster countermeasures are utilized to form an emerging combat means for perfecting a cluster defense system, and corresponding cluster countermeasure technologies become a current research hotspot. In cluster antagonism, different clusters attempt to evolve to their respective optimal antagonism, a process that is highly consistent with the theoretical framework of multi-league gaming. That is, all individuals are divided into a plurality of alliances, the alliances are in competition relationship, and the individuals in the alliances cooperate with each other to minimize the overall cost of the alliances in which they are located. Currently, in the research of Nash equilibrium search problem oriented to multi-league games, most researches default to the knowledge of individual state information in gradient items or obtain the information through distributed estimation, but such research is only applicable to the case that leagues are cooperative relations. In addition, the existing Nash equilibrium search algorithm for multi-alliance game contains global information such as the number of individuals, the characteristic root of the communication topology flap matrix and the like, and cannot be deployed in a completely distributed mode. On one hand, in a strong-countermeasure battlefield environment, communication cannot be directly established between hostile alliances, individuals cannot acquire the states or decision information of all individuals through distributed estimation, and on the other hand, in order to improve the adaptability of a cluster system in a battlefield, the introduction of a central node should be avoided as much as possible, and a decision algorithm is deployed in a completely distributed mode. Therefore, research on a fully distributed Nash equilibrium search method in the multi-alliance game problem is one of the problems to be solved in the current engineering field. In recent years, a Nash equilibrium search algorithm for multi-league games attracts attention of many expert scholars, but until now, related research results are only applicable to the situation of cooperative leagues, and state information acquisition and algorithm design under non-cooperative leagues are considered to be very interesting. The multi-league gaming Nash equilibrium search methods involved in the related studies can be divided into three major categories. The first class of methods regards the status information of all individuals as known information, i.e. global information is introduced into the controller protocol. However, in a practical application scenario, global information is often difficult to obtain by a single individual. In the second class of methods, one individual is selected as a central node within each federation, and the remaining individuals act as subordinate nodes to the central node. In addition, communication is established between the center nodes of different alliances, and interaction behaviors between the alliances are represented by interaction behaviors between the center nodes. Therefore, the method has the characteristic of semi-centralization, and the overall evolution condition of the alliance depends on the central node of the alliance. When a central node is disturbed or damaged in a task, the federation in which the central node is located is greatly affected. In order to estimate the state information of all individuals, a third type of method adopts a distributed consistency estimation strategy, namely, each individual transmits the state information of the individual to a neighbor individual and transmits the state estimation information of the individual to other individuals to the neighbor individual. It is known that each individual's state estimation information for other individuals can only be communicated in a communication link, which can only exist between cooperative leagues, and thus the third class of methods is applicable to multi-league gaming problems constituted by cooperative leagues. In addition, the feedback gains of the three methods contain global information such as Li Puxi z constants, individual numbers, characteristic roots of a communication topology flap matrix a