Unmanned cluster cooperative navigation method and system based on distributed cluster self-organization model

Abstract

The present application relates to the technical field of unmanned cluster, and designs a kind of unmanned cluster cooperative navigation method and system based on distributed cluster self-organizing model.Step 1: obtain task environment, and pretreat it;Step 2: the unmanned cluster platform is expressed using characteristics;Step 3: the unmanned cluster platform of step 2 characteristic expression is carried out based on the interaction frame criterion design of inside and outside cluster of Boyde ring, so that each unmanned platform adjusts its motion state according to the surrounding platform motion state and external environment information perceived by the sensor carried;Step 4: based on the interaction frame criterion designed in step 3, each unmanned vehicle uses its carried model predictive controller to solve the control input of next moment online and updates flight state using four-order Runge-Kutta method;Each unmanned vehicle uses the sensor carried by itself to update the flight state of surrounding unmanned platform online and returns to execute step 3 until reaching target point.The problem of realizing cooperative navigation behavior of cluster overall level under the condition of reducing communication dependence is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned cluster, and discloses an unmanned cluster cooperative navigation method based on a distributed cluster self-organizing model and a system thereof. BACKGROUND

[0002] In recent years, with the rapid development of unmanned aerial vehicle technology, unmanned cluster has become an important technical field. In particular, under the condition of information-based war, unmanned cluster plays an increasingly important role in combat. Unmanned cluster can be used to perform various tasks, such as monitoring, searching, search and rescue, etc. In order to meet the requirements of diversified tasks, unmanned cluster must be able to effectively adapt to complex and dynamic external environment. Existing solutions usually choose large and medium-sized unmanned aerial platforms to perform tasks, however, with increasingly stringent air defense measures, it is difficult for them to penetrate into enemy territory. In contrast, small unmanned platforms are becoming a sharp weapon for penetrating into enemy territory in a large-scale form on modern battlefields due to their low cost and low detectability. However, how to realize safe and autonomous cooperative flight from the task starting point to the predetermined target point in a complex environment is still a problem to be solved.

[0003] For the above problems, according to the form of information interaction and processing with the ground station during flight, the existing methods can be divided into centralized, distributed and centralized and distributed. Considering that it is difficult to guarantee communication with the rear area after the cluster is released into the enemy area, even if a communication relay is established through an aerial platform, it is difficult to guarantee the real-time of the instruction transmission. Therefore, the local coordination method based on the distributed structure is becoming the mainstream. It can be seen that the operating environment of unmanned cluster can be regarded as a communication limited environment, and the existing group navigation method mainly relies on frequent information exchange through the communication network to realize stable navigation behavior. However, this dependence brings challenges to realizing coordinated cooperative behavior in a communication limited and obstacle rich environment. How to realize the cooperative navigation behavior of the whole cluster under the condition of reducing communication dependence is a technical problem urgently to be solved in the field. SUMMARY

[0004] The present application provides an unmanned cluster cooperative navigation method based on a distributed cluster self-organizing model, which solves the problem of realizing the cooperative navigation behavior of the whole cluster under the condition of reducing communication dependence.

[0005] The present application provides an unmanned cluster cooperative navigation system based on a distributed cluster self-organizing model, which realizes the unmanned cluster cooperative navigation method based on the distributed cluster self-organizing model.

[0006] The present application is realized by the following technical scheme:

[0007] An unmanned cluster cooperative navigation method based on a distributed cluster self-organizing model, the method comprising the following steps:

[0008] Step 1: Obtain the task environment and preprocess it;

[0009] Step 2: Express the characteristics of the unmanned cluster platform;

[0010] Step 3: Design the interaction framework criterion inside and outside the cluster based on the Boyde ring for the unmanned cluster platform expressed in Step 2, so that each unmanned platform adjusts its motion state according to the surrounding platform motion state and external environment information perceived by the sensor carried;

[0011] Step 4: Based on the interaction framework criterion designed in Step 3, each unmanned vehicle uses its carried model predictive controller to solve the control input at the next moment online and updates the flight state using the fourth-order Runge-Kutta method; each unmanned vehicle uses its carried sensor to update the flight state of the surrounding unmanned platform online and returns to Step 3 until it reaches the target point.

[0012] Further, Step 1 specifically includes the following steps,

[0013] Step 1.1: According to the existing elevation DEM data processing or through digital computer simulation, for the shape of the obstacle, use its circumscribed circle to approximate, and its radius is represented as T r ; A cylindrical model with random position distribution is used to simulate the obstacles that the unmanned cluster may encounter in low-altitude flight;

[0014] Step 1.2: For non-convex obstacles that may appear in the environment, the unmanned cluster converts them into convex obstacles by adjusting the flight height and further approximates them using the circumscribed cylinder in Step 1.1.

[0015] Further, Step 2 specifically includes the following steps:

[0016] Step 2.1: Simplify the platform kinematic model;

[0017] Step 2.1 specifically assumes that the unmanned cluster participating in the task is marked as N = {1,...,N}, and for each unmanned platform i, the following simplified motion model is adopted, represented as formula (1):

[0018]

[0019] Where x i , y i , h i are the three-dimensional coordinates of the unmanned platform i, and are the control inputs of the platform i's position holding, heading holding and height holding autopilot respectively; τ v τ ψ, τ h and τ λ are the time constants of the three autopilots;

[0020] After determining the time constants, the model is simplified to a second-order integrator model as shown in equation (2):

[0021]

[0022] where p i = (x i , y i , h i ) T represents the position vector, represents the velocity vector,

[0023] represents the control input of each platform; at each time step k, set v min ≤ v i (k) ≤ v max , u min ≤ u i (k) ≤ u max ; based on the control input, the fourth-order Runge-Kutta method is used to update the motion state of each unmanned platform at each time;

[0024] Step 2.2: Based on the simplified platform kinematics model in step 2.1, divide the platform task role;

[0025] Said step 2.2 specifically, according to whether the unmanned platform has global target information, it can be divided into informed members or uninformed members;

[0026] Where the informed unmanned platform has global navigation information and can guide the uninformed unmanned platform to fly to the target;

[0027] The uninformed unmanned platform can only perceive the local state change of its neighbors, and use this information to determine its state at the next time step;

[0028] During initialization, informed individuals are randomly assigned to the group in a specified proportion Δ.

[0029] Further, said step 3 specifically includes the following steps:

[0030] Step 3.1: Each unmanned platform observes the motion state of the unmanned platform in the adjacent airspace according to the sensor carried by itself and obtains the predicted motion state at the current time and the future time by means of the corresponding filtering prediction algorithm;

[0031] Step 3.2: Each platform makes local decision based on the state observation of neighboring platforms and environmental observation information to achieve dynamic self-adaptive adjustment of key parameters;

[0032] Step 3.3: After obtaining the optimized parameter set , the motion model in step 2.1 is used to predict the future state of T p time step as the reference trajectory Decoupled MPC controllers are used for horizontal and vertical directions, where is the reference trajectory before optimization based on the current state; the final control input is provided, which is then used to update the state of the next time step.

[0033] Further, step 3.1 is specifically step 3.1.1: with the current speed v i (k) and A reference coordinate system is established for the reference direction, with counterclockwise rotation defined as positive; the field of view is defined as where α max is the maximum viewing angle; then use to represent the detectable individuals within the local perception range R sen of the unmanned platform i; divide the field of view into Q equal intervals; then the individuals located in the qth angular interval can be represented by formula (3), where ∠p ij (k) represents the relative angle between unmanned platform i and unmanned platform j, and ||p ij (k)||2 represents the relative distance between unmanned platform i and unmanned platform j.

[0034]

[0035] Use formula (4) to calculate the nearest perceptible individual in each sub-interval and form the neighbor set Specifically,

[0036]

[0037] Since the relative positions between unmanned platforms will not change abruptly, when platform i detects platform j, it will create and store the state information structure of platform j locally;

[0038] Step 3.1.2: If the current UAV platform calculates that it is at the front of the swarm flight according to the relative position and velocity information, if the current UAV platform is a non-informed role, it applies to the informed UAV platform in the team through the communication network to obtain the target information to lead the swarm to fly towards the target, and once it is no longer dominant, it will automatically convert to a follower role; the information flow only includes the target and the corresponding marker information, and does not involve the exchange of real-time state information; if the current platform is an informed role, this step is skipped;

[0039] Step 3.1.3: If the current UAV platform detects external obstacles and needs to take obstacle avoidance behavior, define the key neighbor set under the condition of detecting obstacles in formula (5) Where <·> represents the angle between two vectors, and λ represents the critical neighbor angle range;

[0040]

[0041] Key neighbor set is a subset of the perceived neighbor set When the key neighbor set of the individual is empty, it obtains the global target information from the informed platform of the local group through the communication network and assumes the identity of an "informed person" to continue flying towards the target and restore to the original identity after returning to the swarm;

[0042] If the current UAV platform does not detect obstacles, continue to use the perceived neighbor set to perform subsequent state parameter calculation and update.

[0043] Further, the step 3.2 includes the following steps:

[0044] Step 3.2.1: If the current flight state triggers the optimization flag and reaches the optimization interval T0, go to step 3.2.2;

[0045] If the current flight state is safe, keep the current parameter set and continue to fly, and when the fixed optimization interval T1 is reached, go to step 3.2.2;

[0046] At other times, update the flight state according to step 3.3;

[0047] Step 3.2.2: Interaction criteria design among members in the swarm;

[0048] Step 3.2.3: Based on the interaction criteria of step 3.2.2, further define the information interaction items of the UAV platform and the environment;

[0049] Step 3.2.4: Aggregate various control items to obtain the overall control input, as shown in formula (20);

[0050] Step 3.2.5: After obtaining the virtual control input, predict its state in the next T p steps where denotes the predicted state of each time step; based on the predicted state of itself and its neighbors, a local performance evaluation function is defined for each UAV;

[0051] Step 3.2.6: Based on the above local optimization index, the local optimization goal of each UAV is defined as formula (26):

[0052]

[0053] where γ ∈ (0, 1) is a state decay factor representing the importance between different time.

[0054] Further, the step 3.2.2 is specifically,

[0055] Step 3.2.2.1: Calculate the relative speed alignment term to prevent the dispersion phenomenon caused by the local speed inconsistency during the flight of the cluster, which can be expressed as formula (6):

[0056]

[0057] First, calculate the relative distance between the neighboring members and the UAV i as Get the normalized vector Take it as the local weight to calculate the total alignment term, as shown in formula (7):

[0058]

[0059] where, is a constant to be determined, reflects the influence degree of the neighbors of the UAV i on its speed at the current time k;

[0060] Step 3.2.2.2: Calculate the local cohesion term of the members to ensure that the members within the perception range are kept close, which is expressed as formula (8),

[0061]

[0062] where d coh is the minimum distance between UAVs that produces cohesion, is a unit vector from UAV i to UAV j; the local total cohesion term is calculated using the method of weighting the relative distance, as shown in formula (9),

[0063]

[0064] where, is a cohesion term coefficient, which needs to be determined dynamically;

[0065] Step 3.2.2.3: Calculate the local repulsion term to ensure that members are always kept at a minimum safety distance; the repulsion effect is defined as formula (10)

[0066]

[0067] where, d rep is the maximum interaction range at which the UAV platforms start to repel each other; is the avoidance direction of UAV j to UAV i; the total repulsion term of UAV i and its neighboring individuals is represented as formula (11):

[0068]

[0069] where, is the weight parameter of the repulsion term;

[0070] Step 3.2.2.4: In order to maintain stability during flight, a speed maintenance term is added, as shown in formula (12)

[0071]

[0072] where, is the reference speed; is a coefficient, is the average flight direction of the aircraft, as shown in formula (13):

[0073]

[0074] where, is the unit vector of UAV j;

[0075] Definition to represent all interactions within the swarm, as shown in formula (14):

[0076]

[0077] Further, the step 3.2.3 includes the following steps:

[0078] Step 3.2.3.1: A target-oriented interaction term is designed for the informed individuals of the global navigation information, as shown in formula (15);

[0079]

[0080] where, p g is the target coordinates, C tarFor attracting item gain, Ω is the set of informed UAVs, for uninformed individuals, this step is skipped;

[0081] Step 3.2.3.2: Assuming the field of view of UAV i is α i (k); after the obstacle enters the detection range, its boundary intersection point P l = (x l , y l ) can be calculated, P r = (x r , y r ); the angle θ i,o (k) is calculated using the cosine theorem, which represents the angle occupied by the obstacle The feasible angle of UAV i is represented as formula (16):

[0082]

[0083] Divide the feasible angle A i (k) into W subintervals A i,w (k); evaluate the angle difference between the center line of each subinterval and the heading of the current local neighbor, and select the subinterval with the smallest relative angle difference as the local obstacle avoidance direction; the calculation is formula (17):

[0084]

[0085] In calculating the obstacle avoidance direction, the key neighbor set The external obstacle avoidance behavior term is defined as formula (18):

[0086]

[0087] wherein, and represent the boundary angle of each subinterval; by analyzing the navigation direction, the expected navigation term of UAV i is defined, wherein represents the gain factor; this factor allows to adapt its navigation gain by adapting its heading to different requirements of the task in different situations; similarly, define to represent the sum of the interaction between the UAV and the external environment, as shown in formula (19):

[0088]

[0089] Further, the step 3.2.6 includes the following steps,

[0090] Step 3.2.6.1: initialize the meta-heuristic optimization algorithm, including the feasible range of each parameter, the number of populations, the maximum number of iterations, and the set of factors to be optimized represented as representing the weights in the virtual control input;

[0091] Step 3.2.6.2: bring the set of variables to be optimized into formula (26) for calculating the current individual motion state function value in the future period of time;

[0092] Step 3.2.6.3: perform multiple iteration optimizations to obtain the control parameter set at the future time

[0093] An unmanned cluster cooperative navigation system based on a distributed cluster self-organizing model, which uses the unmanned cluster cooperative navigation method based on the distributed cluster self-organizing model as described above, and the system comprises:

[0094] A preprocessing module: acquires a task environment and pre-processes it;

[0095] A characteristic expression module: expresses the unmanned cluster platform using characteristics;

[0096] An interaction framework criterion design module: based on the characteristic expression of the unmanned cluster platform, performs the interaction framework criterion design of the cluster based on the Boyde ring, so that each unmanned platform adjusts its own motion state according to the surrounding platform motion state and external environment information perceived by the sensor carried by the platform;

[0097] A flight state updating module: based on the interaction framework criterion, each unmanned vehicle uses its carried model predictive controller to solve the control input at the next time online and updates the flight state using the fourth-order Runge-Kutta method; each unmanned vehicle uses the sensor carried by itself to update the flight state of the surrounding unmanned platform online and returns to the interaction framework criterion design module until the target point is reached.

[0098] The beneficial effects of the present application are:

[0099] The present application can realize the cooperative autonomous navigation behavior of the cluster in the static obstacle and dynamic obstacle environment.

[0100] The self-organizing model in the present application has good scalability and can reduce the dependence on the communication network by means of local perception and event-triggered communication.

[0101] The asynchronous model parameter online adjustment mechanism in the present application has high calculation real-time performance and can give the individuals in the cluster higher autonomy to adapt to the changeable external environment. BRIEF DESCRIPTION OF DRAWINGS

[0102] Figure 1 is the method flowchart of the present application.

[0103] Figure 2Fig. 1 is a schematic diagram of the determination of the local aware neighbor set (left) and the key neighbor set (right) of the present application.

[0104] Figure 3 Fig. 2 is a schematic diagram of the cooperative navigation of the present application in three dimensions.

[0105] Figure 4 Fig. 3 is a schematic diagram of the cooperative navigation of the traditional Flocking model in three dimensions.

[0106] Figure 5 Fig. 4 is a schematic diagram of the cooperative navigation of the MPC in three dimensions.

[0107] Figure 6 Fig. 5 is a schematic diagram of the cooperative navigation of the parameter adjustment in three dimensions.

[0108] Figure 7 Fig. 6 is a schematic diagram of the relative distance curve between the cluster members of the present application and different methods.

[0109] Figure 8 Fig. 7 is a schematic diagram of the relative distance curve between the cluster members of the present application and different methods and environmental obstacles.

[0110] Figure 9 Fig. 8 is a schematic diagram of the speed change curve of the cluster members of the present application and different methods.

[0111] Figure 10 Fig. 9 is a schematic diagram of the angular velocity change curve of the cluster members of the present application and different methods.

[0112] Figure 11 Fig. 10 is a screenshot of the motion state of the present application at t = 30s in a dynamic environment.

[0113] Figure 12 Fig. 11 is a screenshot of the motion state of the present application at t = 70s in a dynamic environment.

[0114] Figure 13 Fig. 12 is a screenshot of the motion state of the present application at t = 110s in a dynamic environment.

[0115] Figure 14 Fig. 13 is a screenshot of the motion state of the present application at t = 150s in a dynamic environment.

[0116] Figure 15 Fig. 14 is a screenshot of the motion state of the present application at t = 160s in a dynamic environment.

[0117] Figure 16 Fig. 15 is a screenshot of the motion state of the present application at t = 200s in a dynamic environment.

[0118] Figure 17 Fig. 16 is a schematic diagram of the relative distance change curve within the cluster of the present application in a dynamic environment.

[0119] Figure 18is a schematic diagram of a relative distance change curve of a cluster outside in a dynamic environment. DETAILED DESCRIPTION

[0120] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, techniques, etc. in order to provide a thorough understanding of the embodiments of the application. However, it will be apparent to those skilled in the art that the application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, and circuits are omitted so as not to obscure the description of the application with unnecessary detail.

[0121] It is to be understood that the terminology "including", "comprising", "consisting" and "consisting essentially of" used in the specification and the appended claims, are used in the sense of open ended inclusion, that is, to include not only the recited features or steps but also others that are not recited.

[0122] It is also to be understood that the terminology used in the description is for the purpose of describing specific embodiments only and is not intended to be limiting. As used in this specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0123] The above description is only a summary of the application and the following Figures 1-18 The technical solutions in the embodiments of the application are described clearly and completely. Obviously, the described embodiments are only some of the embodiments of the application, but not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the application.

[0124] In the following description, many specific details are set forth in order to provide a thorough understanding of the application. However, the application can be practiced without the specific details. In other instances, well-known methods, procedures, components, and circuits have not been described in detail so as not to obscure the description of the application.

[0125] Embodiment one

[0126] The embodiment provides an unmanned cluster cooperative navigation method based on a distributed cluster self-organizing model. The method first establishes an interactive decision process between a single unmanned platform and its external environment based on a Boyde OODA cycle (observation, orientation, decision, action) and a locally limited perception model. Further, roles of different unmanned platforms in the cluster are classified, and the cooperative flight efficiency is improved through the guiding behavior of key nodes. Each unmanned platform uses a dynamic adjustment mechanism of control parameters to allow adaptive adjustment based on local flight states. In addition, each unmanned platform is equipped with a model predictive control (MPC) controller which provides feasible control inputs to ensure robust and reliable operation in complex and dynamic scenarios. Through integration of the above modules, a distributed cluster autonomous cooperative navigation method in the embodiment is finally formed, and a method flowchart is shown in Figure 1 The method comprises the following steps:

[0127] Step 1: Obtain a task environment and pre-process it;

[0128] Further, the step 1 specifically comprises the following steps,

[0129] Step 1.1: Consider an unknown three-dimensional task environment which can be obtained according to existing height DEM data processing or through digital computer simulation. For the shape of the obstacle therein, an inscribed circle is used for approximation, and the radius can be represented as T r ; therefore, the obstacles and no-fly zones in the environment can be represented by a set. In the present application, a cylindrical model with random position distribution is used to simulate the obstacles that the unmanned cluster may encounter in low-altitude flight;

[0130] Step 1.2: For the non-convex obstacles that may appear in the environment, the unmanned cluster can convert them into convex obstacles by adjusting the flight height and further use the inscribed cylinder in step 1.1 for approximation;

[0131] Step 2: Express the characteristics of the unmanned cluster platform;

[0132] Further, the step 2 specifically comprises the following steps:

[0133] Step 2.1: Simplify the kinematic model of the platform;

[0134] The step 2.1 is specifically that, in the task scenario involved in the present application, the unmanned cluster participating in the task is marked as N={1,...,N}. For the sake of convenience, a simplified kinematic model of each unmanned platform i is used, which can be represented as formula (1):

[0135]

[0136] where x i , y i , h i are the three-dimensional coordinates of the unmanned platform i, and are the control inputs of the position-keeping, heading-keeping and height-keeping autopilots of platform i, respectively; τ v τ ψ , τ h and τ λ are the time constants of the three autopilots;

[0137] After determining the time constants, the model is further simplified to a second-order integrator model, as shown in equation (2):

[0138]

[0139] where p i = (x i , y i , h i ) T represents the position vector, represents the velocity vector, represents the control input of each platform; further restrictions are imposed on the motion state and acceleration input to ensure that they can be processed by the subsequent actuators; at each time step k, set v min ≤ v i (k) ≤ v max , u min ≤ u i (k) ≤ u max ; based on the obtained control input, the fourth-order Runge-Kutta method is used to update the motion state of each unmanned platform at each time;

[0140] Step 2.2: Based on the simplified platform kinematics model in step 2.1, divide the platform task roles;

[0141] Specifically, according to whether the unmanned platform has global target information, it can be divided into informed members or uninformed members;

[0142] Among them, the informed unmanned platform has global navigation information and can guide the uninformed unmanned platform to fly to the target;

[0143] In contrast, the uninformed unmanned platform can only perceive the local state changes of its neighbors and use this information to determine its state at the next time step;

[0144] During initialization, informed individuals are randomly assigned to the group in a specified proportion Δ.

[0145] Step 3: Design the interaction framework criterion inside and outside the cluster based on the Boyce ring for the unmanned cluster platform described in step 2, in which each unmanned platform adjusts its own motion state according to the surrounding platform motion state and external environment information perceived by the sensor carried by the platform, so as to realize safe cooperative navigation from the initial position to the target;

[0146] Further, step 3 specifically includes the following steps:

[0147] Step 3.1: Each unmanned platform observes the motion state of the unmanned platform in the adjacent airspace according to the sensor carried by itself and obtains the predicted motion state at the current time and the future time by means of the corresponding filtering prediction algorithm;

[0148] Further, step 3.1 specifically includes step 3.1.1: taking the current speed v i (k) and A reference coordinate system is established with the reference direction, and counterclockwise rotation is defined as positive; on this basis, the field of view (FOV) is defined as Where α max is the maximum viewing angle; then use to represent the detectable individual in the local perception range R sen of the unmanned platform i; in order to describe the unmanned platform at different positions, the field of view is divided into Q equal parts; then the individual located in the qth angular interval Can be expressed as formula (3), where ∠p ij (k) represents the relative angle between the unmanned platform i and the unmanned platform j, and ||p ij (k)||2 represents the relative distance between the unmanned platform i and the unmanned platform j;

[0149]

[0150] In order to reduce the complexity of perception, formula (4) is used to calculate the nearest perceptible individual in each sub-interval, and the neighbor set N Specifically,

[0151]

[0152] Since the relative position between the unmanned platforms will not change abruptly, when platform i detects platform j, it will create and store the state information structure of platform j locally; in actual application, means such as Kalman filtering can be used to estimate the motion state of platform j;

[0153] Step 3.1.2: If the current UAV platform calculates that it is at the front of the swarm flight according to the relative position and velocity information, if the current UAV platform is a non-informed role, it applies to the informed UAV platform in the team through the communication network to obtain the target information to lead the swarm to fly towards the target, and once it is no longer dominant, it will automatically convert to a follower role; this mechanism ensures that at any time, at least one leader guides the swarm to move towards the target; it can be seen that the information flow only includes the target and the corresponding marker information, and does not involve the exchange of real-time state information; this event-triggered mechanism effectively reduces the frequency of information exchange, thereby reducing the communication burden; if the current platform is an informed role, skip this step;

[0154] Step 3.1.3: If the current UAV platform detects external obstacles and needs to take obstacle avoidance behavior, define the key neighbor set under the condition of detecting obstacles in formula (5) Where <·> represents the angle between two vectors, and λ represents the critical neighbor angle range; by narrowing the key neighbor set, the platform can more purposefully guide the neighbors in the forward direction;

[0155]

[0156] It can be seen that the key neighbor set is a subset of the perceived neighbor set This relationship is established without the need for a communication link; it should be noted that when the individual's key neighbor set is empty, it obtains the global target information from the informed platform in the local group through the communication network and assumes the identity of an "informed" person to continue flying towards the target and restore to the original identity after returning to the swarm;

[0157] If the current UAV platform does not detect obstacles, continue to use the perceived neighbor set to perform subsequent state parameter calculation and update.

[0158] Step 3.2: Each UAV platform makes a local decision based on the observed state values of neighboring platforms and environmental observation information to achieve dynamic adaptive adjustment of key parameters;

[0159] Further, the step 3.2 includes the following steps:

[0160] Step 3.2.1: If the current flight state triggers the optimization flag (internal safety distance is too small or external encounters obstacles) and reaches the optimization interval T0, go to step 3.2.2;

[0161] If the current flight state is safe, keep the current parameter set and continue to fly, and when the fixed optimization interval T1 is reached, go to step 3.2.2 as well;

[0162] The flight states are updated according to step 3.3 at the rest of the time; it is worth noting that the present application does not assume that the optimization processes between platforms are synchronous; the described method does not need to explicitly exchange the motion states through the communication network, thereby eliminating the need for strict clock synchronization between platforms in the parameter optimization process; this asynchronous method can be more flexible and real-time to adapt to the external dynamic environment;

[0163] Step 3.2.2: Interaction criterion design between intra-cluster members;

[0164] Further, the step 3.2.2 is specifically,

[0165] Step 3.2.2.1: Calculate the relative speed alignment term to prevent the dispersion phenomenon of the cluster in the flight process due to the inconsistency of the local speed, which can be expressed as formula (6);

[0166]

[0167] Considering that the cluster members tend to align their own states with nearby individuals, first calculate the relative distance between the neighboring members and the unmanned platform i as Considering that the individuals with closer distances have greater influence on the speed consistency, obtain the normalized vector Take it as a local weight to calculate the total alignment term, as shown in formula (7):

[0168]

[0169] Wherein, is a constant to be determined, reflects the influence degree of the neighbors of the unmanned platform i on its speed at the current time k;

[0170] Step 3.2.2.2: Calculate the member local cohesion term to ensure that the members within the perception range maintain close to prevent the decomposition of the unmanned cluster, which can be expressed as formula (8),

[0171]

[0172] Wherein, d coh is the minimum distance between unmanned platforms to produce cohesion, is a unit vector from the unmanned platform i to the unmanned platform j; considering that the cluster members tend to prefer to realize the cohesion effect with the individuals with the farthest distance within the sensing range, the relative distance weighting method is used here to calculate the local total cohesion term as shown in formula (9),

[0173]

[0174] Wherein, is a cohesion term coefficient, which needs to be determined dynamically;

[0175] Step 3.2.2.3: Calculate the local repulsion term to avoid collision between members, ensure that the members always maintain a minimum safety distance; the repulsion effect is defined as formula (10)

[0176]

[0177] where d rep is the maximum interaction range at which the unmanned platforms start to repel each other; is the avoidance direction of unmanned platform j to unmanned platform i; considering that individuals prefer to avoid collision with neighbors, the relative distance is used as the weighted weight of the avoidance term; the total repulsion term of unmanned platform i and its neighboring individuals can be expressed as formula (11):

[0178]

[0179] where, is the weight parameter of the repulsion term;

[0180] Step 3.2.2.4: In order to maintain stability during flight, the change of unmanned platform speed will not exceed the performance limit; therefore, a speed maintenance term is added, as formula (12)

[0181]

[0182] where, is the reference speed, usually represented by the average performance of the unmanned platform; is the coefficient, is the average flight direction of the aircraft, as shown in formula (13), and the relative distance weighting method is also used to ensure that the unmanned platform aligns with the individuals far away first, thereby reducing the probability of splitting:

[0183]

[0184] where, is the unit vector of unmanned platform j;

[0185] Through the above analysis, define to represent all interactions within the swarm, as shown in formula (14):

[0186]

[0187] Step 3.2.3: Based on the interaction criteria of step 3.2.2, further define the information interaction term of the unmanned platform and the environment;

[0188] Further, the step 3.2.3 includes the following steps:

[0189] Step 3.2.3.1: Target-oriented interaction items are designed for informed individuals in the cluster who have global navigation information, as shown in equation (15);

[0190]

[0191] where p g is the target coordinate, C tar is the attraction item gain, and Ω is the set of informed UAVs. For uninformed individuals, this step is skipped.

[0192] Step 3.2.3.2: Further according to the relative interaction relationship between the platform and the environmental obstacles, the obstacle avoidance interaction item is defined, assuming that the field of view of UAV i is α i (k); after the obstacle enters the detection range, its boundary intersection P l = (x l , y l ) and P r = (x r , y r ) can be calculated; the angle θ i,o (k) occupied by it is calculated using the cosine theorem, assuming that UAV i detects obstacle at time k. The feasible angle of UAV i can be expressed as equation (16):

[0193]

[0194] To prevent the obstacle avoidance behavior based on local perception from causing the UAV to split, the feasible angle A i (k) is divided into W subintervals A i,w (k); then, the angle difference between the center line of each subinterval and the heading of the current local neighbor is evaluated, and the subinterval with the smallest relative angle difference is selected as the local obstacle avoidance direction; which can be calculated as equation (17):

[0195]

[0196] It is worth noting that in the calculation of the obstacle avoidance direction, the key neighbor set is used, which helps the UAV to select individuals that can guide it forward and allow a more targeted response to external obstacles; on this basis, the external obstacle avoidance behavior item is defined as shown in equation (18):

[0197]

[0198] where and represent the boundary angles of each subinterval; by analyzing the navigation direction, the expected navigation item of UAV i is defined, where represents the gain factor; this factor allows to adapt to different requirements of the mission in different situations by adaptively modifying its navigation gain to adapt to its heading; similarly, we define

[0199]

[0200] Step 3.2.4: The various control terms are aggregated to obtain the overall control input, as shown in equation (20); this integrated input integrates the effects of neighboring velocity, obstacle avoidance, and target attraction, enabling the swarm of UAVs to coordinate safe navigation in complex environments;

[0201] Step 3.2.5: After obtaining the virtual control input, it is combined with the motion equation to predict its state in the next T p

[0202] Each term in the function is normalized to facilitate optimization.

[0203]

[0204] represents the concentration of local neighbor positions of the UAV i. When the set of perceived neighbors is non-empty, is calculated according to the distance between their predicted positions and the current position, where is the farthest neighbor, δ is a parameter used to adjust the rate of change of the index

[0205]

[0206] represents the degree of directional consistency between the local neighbors of the UAV i; represents the collision cost of the UAV with its neighbors within the perception range, where d rep represents the predetermined safety radius of the UAV i. represents the collision cost of the UAV with external obstacles, where r o is the preset safety distance for each obstacle; by limiting the local objective function, each UAV can independently optimize its parameters and flexibly adjust its local behavior.

[0207] Step 3.2.6: Based on the above local optimization index, the local optimization goal of each UAV is defined as equation (26):​​​​​​

[0208]

[0209] where γ ∈ (0, 1) is a state decay factor that characterizes the importance between different time steps. In essence, this is achieved by adjusting the weight strategy set to maximize the cumulative reward on the time stamp p To facilitate deployment, we employ the same particle swarm solver for each UAV to achieve real-time online optimization.

[0210] Further, the step 3.2.6 comprises the following steps,

[0211] Step 3.2.6.1: initialize meta-heuristic optimization algorithm, including the feasible range of each parameter, the number of population, the maximum number of iterations, the set of factors to be optimized can be represented as represent the weight of each item in the virtual control input;

[0212] Step 3.2.6.2: bring the set of variables to be optimized into formula (26) to calculate the current individual motion state function value in the future period of time;

[0213] Step 3.2.6.3: perform multiple iterations of optimization to obtain the control parameter set

[0214] Step 3.3: after obtaining the optimized parameter set , use the motion model in step 2.1 to predict the future state of T p time step as the reference trajectory To ensure the stability of the input, a decoupled MPC controller is used in formula (27) for the horizontal and vertical directions, where is the reference trajectory before optimization based on the current state; this method provides the final control input, which is then used to update the state at the next time step;

[0215]

[0216] Step 4: based on the interaction framework criteria designed in step 3, each UAV uses its onboard model predictive controller to solve the control input at the next time step online and updates the flight state using the fourth-order Runge-Kutta method; each UAV uses its own onboard sensors to update the flight state of surrounding UAVs online and returns to step 3 until it reaches the target point;

[0217] Specifically, assume that there are M obstacles randomly distributed in a 2x2x2 km task area, where the radius of each obstacle is randomly distributed within 3%-6% of the task area size. The initial positions of the unmanned cluster are randomly generated within the specified area and the relative distance between them is guaranteed to be safe, and the initial heading is randomly generated within ±30° on both sides of the line connecting the start and end points, where the informed individuals are randomly selected for initialization according to Δ=20%. Our goal is to achieve autonomous cooperative navigation of the cluster from the start point to the end point in the obstacle environment. The following table gives the relevant parameter ranges.

[0218] Table 1 Key simulation parameters

[0219]

[0220]

[0221] As can be seen from Figures 3-6 , the proposed method can achieve collision-free cooperative navigation behavior from the start point to the end point, meeting the task requirements.

[0222] Figure 7 and Figure 8 reflect the relative distance change curves between the inside and outside of the cluster during flight, where figure 0 represents the proposed method, and 1, 2, and 3 represent the remaining three methods. It can be seen that it can maintain a safe flight state to achieve the established task goal.

[0223] Figure 9 and Figure 10 reflect the speed and angular velocity change curves of the cluster members during flight, and it can be seen that different methods can obtain control inputs within the specified range. In addition, it can be seen that the state quantity change of method 2 is relatively flat compared to the other methods, because it uses the rolling control strategy of MPC and adds energy consumption to the optimization model, but it can be seen from the final result that it takes the longest time to reach the end point compared to other methods.

[0224] Figures 11-16 shows the cooperative navigation behavior verification results of the proposed method in a dynamic obstacle environment. The red obstacles represent obstacles that are in a stationary state at that time, and the green obstacles represent obstacles that are in a moving state at that time. It can be seen that our method avoids external collisions and dispersion of the cluster while maintaining cooperative navigation behavior, which can be verified from Figure 17 and Figure 18 the distance change curves of the cluster relative to the inside and outside, proving the effectiveness of the proposed method.

[0225] Embodiment two

[0226] The embodiment provides an unmanned cluster cooperative navigation system based on a distributed cluster self-organizing model, the system uses an unmanned cluster cooperative navigation method based on a distributed cluster self-organizing model as described in embodiment one, and the system comprises the following modules:

[0227] A preprocessing module: obtaining a task environment and preprocessing the task environment;

[0228] A characteristic expression module: expressing the unmanned cluster platform by using characteristics;

[0229] An interaction framework criterion design module: based on the characteristic expression of the unmanned cluster platform, the interaction framework criterion design module is based on a Boyde ring, in the framework, each unmanned platform adjusts its own motion state according to the surrounding platform motion state and external environment information perceived by the sensor carried by the platform, thereby realizing safe cooperative navigation from an initial position to a target;

[0230] A flight state updating module: based on the interaction framework criterion, each unmanned aerial vehicle uses a model predictive controller carried by the unmanned aerial vehicle to solve the control input at the next moment online and uses a fourth-order Runge-Kutta method to update the flight state; each unmanned aerial vehicle uses a sensor carried by the unmanned aerial vehicle to update the flight state of the surrounding unmanned platform online and returns to the interaction framework criterion design module until the target point is reached.

Claims

1. An unmanned swarm cooperative navigation method based on a distributed swarm self-organization model, characterized in that, The method comprises the following steps: Step 1: Obtain the task environment and preprocess it; Step 2: Express the unmanned cluster platform using characteristics; Step 3: Design the interaction framework criterion inside and outside the cluster based on the Boyde ring for the unmanned cluster platform expressed in step 2, so that each unmanned platform adjusts its own motion state according to the surrounding platform motion state and external environment information perceived by the carried sensor; Step 4: Based on the interaction framework criterion designed in step 3, each unmanned aircraft uses its carried model predictive controller to solve the control input at the next moment online and updates the flight state using the fourth-order Runge-Kutta method; each unmanned aircraft updates the flight state of the surrounding unmanned platform online using its own carried sensor and returns to step 3 until it reaches the target point; Step 3 specifically comprises the following steps: Step 3.1: Each unmanned platform observes the motion state of the unmanned platform in the adjacent airspace according to the sensor carried by itself and obtains the predicted motion state at the current time and in the future by means of the corresponding filtering prediction algorithm; Step 3.2: Each unmanned platform makes a local decision according to the state observation value and environmental observation information of the adjacent platform to realize dynamic adaptive adjustment of the key parameters; Step 3.3: Obtain optimized parameter set After that, the motion model in step 2.1 is used to predict T p the future state of the time step as the reference trajectory Decoupled MPC controllers are employed for horizontal and vertical directions, where is the reference trajectory before optimization based on the current state; provides the final control input, which is then used to update the state for the next time step; Specifically, step 3.1 is as follows: Step 3.1.1: At the current speed v i (k) and Establish a reference coordinate system for the reference direction, defining counterclockwise rotation as positive; define the field of view as... Where α max Set the maximum viewing angle; then use R represents the local perception range of unmanned platform i. sen Detectable individuals within; dividing the field of view into Q equal parts; then individuals located within the q-th angular interval. This can be expressed as formula (3), where ∠p ij (k) represents the relative angle between unmanned platform i and unmanned platform j, ||p ij (k)||2 represents the relative distance between unmanned platform i and unmanned platform j; Using equation (4) to calculate the nearest perceptible individual in each sub-interval and form the neighbor set Specifically, Since the relative position between the unmanned platforms will not change abruptly, when platform i detects platform j, it will create and store the state information structure of platform j locally; Step 3.1.2: If the current unmanned platform calculates that it is located at the front end of the cluster flight according to the relative position and relative speed information, if the unmanned platform is currently a non-informed role, it will apply to the informed unmanned platform in the group through the communication network to obtain the target information to lead the cluster to fly towards the target, and once it is no longer dominant, it will automatically convert to a follower role; the information flow only includes the target and the corresponding marker information, and does not involve the exchange of real-time state information; if the current platform is an informed role, skip this step; Step 3.1.3: If the current UAV detects external obstacles and needs to take obstacle avoidance behavior, define the key neighbor set under the condition of detecting obstacles in formula (5) where <·> represents the angle between two vectors, and λ represents the critical neighbor angle range; Key neighbor set is a subset of the aware neighbor set When the individual's key neighbor set is empty, it acquires global target information from the informed platform of the local group by means of the communication network and assumes the identity of the informed person to continue flying towards the target and resumes the original identity after returning to the group. If no obstacles are detected by the current unmanned platform, continue to use the perceived neighbor set Perform a subsequent state parameter computation update; Step 3.2 comprises the following steps: Step 3.2.1: If the current flight state triggers the optimization flag and reaches the optimization interval T0, go to step 3.2.2; If the current flight state is safe, keep the current parameter set to continue flying and reach the fixed optimization interval T1, also go to step 3.2.2; At other times, update the flight state according to step 3.3; Step 3.2.2: Design of interaction criterion among members in the cluster; Step 3.2.3: Based on the interaction criterion of step 3.2.2, further define the information interaction items of the unmanned platform and the environment; Step 3.2.4: Sum up various control items to obtain the overall control input; Step 3.2.5: After obtaining the virtual control input, predict its state in the next T p steps using the motion equation where denotes the predicted state of each time step; a local performance evaluation function is defined for each unmanned platform based on its own and its neighbors' predicted states; Step 3.2.6: Based on the local optimization index, define the local optimization goal of each unmanned platform as formula (26): Where γ ∈ (0, 1) is a state decay factor representing the importance between different times.

2. The method of claim 1, wherein, Step 1 specifically comprises the following steps, Step 1.1: Obtain or by digital computer simulation, according to existing elevation DEM data processing, for the shape of the obstacles in it using its circumscribed circle approximation, whose radius is expressed as T r ; Adopt a cylindrical model with random distribution of positions to simulate the obstacles that the unmanned cluster may encounter in low-altitude flight; Step 1.2: For non-convex obstacles that may appear in the environment, the unmanned cluster converts them into convex obstacles by adjusting the flight height and further approximates them using the circumscribed cylinder in step 1.

1.

3. The method of claim 1, wherein, The step 2 specifically comprises the following steps: Step 2.1: Simplifying the platform kinematic model; The step 2.1 is specifically, assuming that the unmanned cluster participating in the task is marked as N={1,...,N}, the following simplified motion model is adopted for each unmanned platform i, denoted as formula (1): where x i , y i , h i are the three-dimensional coordinates of the unmanned platform i, and are the control inputs of the platform i's position-keeping, heading-keeping and height-keeping autopilots respectively; τ v τ ψ , τ h and τ λ are the time constants of the three autopilots; After determining the time constant, the model is simplified to a second-order integrator model, as shown in formula (2): where p i = (x i , y i , h i ) T represents the position vector, represents the velocity vector, represents the control input of each platform; at each time step k, set v min ≤ v i (k) ≤ v max , u min ≤ u i (k) ≤ u max ; using the fourth-order Runge-Kutta method to update the motion state of each unmanned platform at each time based on the obtained control input; Step 2.2: Based on the simplified platform kinematic model in step 2.1, the platform task role is divided; The step 2.2 is specifically, according to whether the unmanned platform has global target information, it can be divided into informed members or uninformed members; Where the informed unmanned platform has global navigation information and can guide the uninformed unmanned platform to fly to the target; The uninformed unmanned platform can only perceive the local state change of its neighbors and determine its next time step state using this information; In the initialization process, informed individuals are randomly allocated to the group in a specified proportion Δ.

4. The method of claim 1, wherein, The step 3.2.2 specifically comprises the following steps: Step 3.2.2.1: Calculate the relative speed alignment term to prevent the cluster from dispersing due to inconsistent local speed during flight, which can be expressed as formula (6); First, the relative distance between the neighboring member and the unmanned platform i is calculated as p ij (k) = p j (k) - p i (k), The normalized vector is obtained It is used as a local weight to calculate the total alignment item, as shown in equation (7): wherein, is a constant to be determined, reflects the degree of influence of the neighbors of the unmanned platform i on its speed at the current time k; Step 3.2.2.2: Calculate member local cohesion terms to ensure members within the perception range remain close is represented as equation (8), where d coh is the minimum distance between unmanned platforms that produces cohesion, is the unit vector from unmanned platform i to unmanned platform j; the local global cohesion term is calculated using a relative distance weighting method as shown in equation (9), wherein is a cohesion term coefficient, which needs to be determined dynamically; Step 3.2.2.3: Calculate the local repulsive force term to ensure that members always maintain a minimum safety distance; the repulsive force effect is defined as formula (10) where d rep is the maximum interaction range at which the UAV platforms start to repel each other; is the avoidance direction of UAV platform j to UAV platform i; the total repulsion term of UAV platform i and its neighboring individuals is expressed as equation (11): wherein is a weight parameter for the repulsion term; Step 3.2.2.4: In order to maintain stability during flight, a speed maintenance term is added, as shown in formula (12) wherein is the reference speed; is a coefficient, is the average flight direction of the aircraft, as shown in equation (13): wherein, is a unit vector of the unmanned platform j; Definitions All interactions within a colony are characterized by the equation (14) as follows:

5. The method of claim 1, wherein, The step 3.2.3 comprises the following steps: Step 3.2.3.1: A target-oriented interaction term is designed for informed individuals with global navigation information, as shown in formula (15); where p g is the target coordinate, C tar is the attraction term gain, and Ω is the set of informed UAVs. For uninformed individuals, this step is skipped. Step 3.2.3.2: Assume the field of view of UAV i is a i (k); after the obstacle enters the detection range, its boundary intersection point P l = (x l , y l ) can be calculated; the angle θ r = (x r , y r ) occupied by it can be calculated by using the cosine theorem i,o (k) Assume that the obstacle is detected by UAV i at time k The feasible angle of UAV i is obtained as formula (16): A feasible angle A i (k) is divided into W sub-intervals A i,w (k); the angle difference between the center line of each sub-interval and the heading of the current local neighbor is evaluated, and the sub-interval with the smallest relative angle difference is selected as the local obstacle avoidance direction; the calculation is formula (17); The key neighbor set is used when computing the obstacle avoidance direction The external obstacle avoidance behavior term is defined as shown in equation (18): wherein, and represent the boundary angles of each sub-interval; by analyzing the navigation directions, the expected navigation items of the unmanned platform i are defined, wherein represents the gain factor; this factor allows to adapt its navigation gain by adapting to different situations, adapting to different requirements of the mission; likewise, it is defined to represent the sum of the interactions between the unmanned platform and the external environment, as shown in equation (19):

6. The method of claim 1, wherein, The step 3.2.6 comprises the following steps, Step 3.2.6.1: initialize meta-heuristic optimization algorithm, including the feasible range of each parameter, the number of population, the maximum number of iterations, the set of factors to be optimized is represented as represent the weight of each item in the virtual control input; Step 3.2.6.2: The set of variables to be optimized is brought into formula (26) to calculate the current individual motion state function value in the future period of time; Step 3.2.6.3: Perform multiple iterations of optimization to obtain a set of control parameters for a future time 7. An unmanned swarm cooperative navigation system based on a distributed swarm self-organization model, characterized in that, The system uses the unmanned cluster cooperative navigation method based on the distributed cluster self-organizing model according to any one of claims 1-6, and the system comprises: A preprocessing module: obtaining the task environment and preprocessing it; A characteristic expression module: expressing the unmanned cluster platform with characteristics; An interaction framework criterion design module inside and outside the cluster: based on the characteristic expression of the unmanned cluster platform, the interaction framework criterion design module inside and outside the cluster based on the Boyer ring is designed, so that each unmanned platform adjusts its own motion state according to the surrounding platform motion state and external environment information perceived by the sensor; A flight state updating module: based on the interaction framework criterion, each unmanned vehicle uses its onboard model predictive controller to solve the control input at the next time and updates the flight state using the fourth-order Runge-Kutta method; each unmanned vehicle uses its own onboard sensor to update the flight state of the surrounding unmanned platform online and returns to the interaction framework criterion design module inside and outside the cluster until reaching the target point.

Images