A method for unmanned aerial vehicle assisted public safety network connectivity maintenance

Abstract

The application relates to a method for maintaining the connectivity of a public safety network by using a UAV, which comprises the following steps: establishing a data transmission network model of a ground user to an undamaged base station based on a public safety network, and establishing a UAV movement model; maintaining the communication between a UAV group and the ground user by combining the network model and the UAV movement model with an OFDMA technology. The application maximizes the utility function while maintaining the network connectivity, uses algebraic connectivity, that is, by introducing the Fiedler eigenvalue of the network topology, updates the selection probability, ensures the reliability of the UAV data transmission from the macro and long-term perspectives, and finally makes the UAV network achieve a better balance among coverage, mobility and connectivity.

Description

Technical Field

[0001] This invention relates to the field of dynamic network communication technology, and in particular to a method for maintaining connectivity in a drone-assisted public safety network. Background Technology

[0002] With the advent and development of the information age, massive amounts of data exchange occur in my country every moment. Natural disasters (such as earthquakes, typhoons, and tsunamis) can damage ground-based communication infrastructure. In situations where infrastructure is lacking or inadequate, such as base stations damaged by unforeseen events, it is crucial to restore the Public Safety Network (PSN) within a very short time to ensure normal communication. Ensuring effective coverage and reliable information transmission within such a PSN is key to restoring the operation of the public safety communication system.

[0003] Unmanned aerial vehicles (UAVs), with their flexible deployment, unmanned operation, and ability to fly in hazardous spaces, provide real-time communication services, significantly mitigating communication disruptions. On one hand, UAVs act as aerial base stations, providing real-time and flexible communication services to ground users; on the other hand, they can serve as auxiliary relays, improving the transmission performance and coverage of ground communication facilities. Therefore, UAVs play a crucial role in communication services for public safety rescue. In particular, due to the complexity of missions, UAV swarms often operate collaboratively. Maintaining connectivity between multiple UAVs is essential for enabling them to work together effectively. To reduce excessive routing overhead while ensuring multi-hop link connectivity, existing technologies propose using graph theory-based d-hop connected dominance sets (d-CDS) to construct a Virtual Backbone Network (VBN) for UAV swarms. One existing multi-relay UAV selection scheme based on fuzzy optimal selection addresses the trade-off between surveillance tasks and connectivity maintenance by utilizing historical detection information and surveillance benefit estimation. However, in large-scale UAV-assisted public safety network scenarios, the distinct roles or specific contributions of each UAV to network connectivity are rarely discussed. For example, a network may disconnect due to the high mobility and limited coverage of a drone, thus compromising the reliability of its communication.

[0004] To address the above issues, this patent solves a typical problem in public safety networks: drones aim to cover as many ground users (GUs) as possible while ensuring reliable data transmission. Summary of the Invention

[0005] To ensure the connectivity of local and overall networks, this invention proposes a method for maintaining the connectivity of UAV-assisted public safety networks. A learning-based role selection mechanism is designed based on the location and coverage of the UAVs, and the selection probability is updated by introducing Fiedler features of the network topology, thus ensuring the reliability of UAV swarm data transmission from a macro and long-term perspective.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A method for maintaining connectivity in a drone-assisted public safety network includes:

[0008] A data transmission network model from ground users to undamaged base stations is established based on the public safety network, and a UAV mobility model is also established.

[0009] By combining the network model and the UAV mobility model with orthogonal frequency division multiple access technology, communication between the UAV swarm and ground users can be maintained.

[0010] Preferably, establishing the data transmission network model includes:

[0011] Determine the collection of drones and the collection of ground users And use u to represent the undamaged base stations in the public safety network. N+1 express;

[0012] Based on the data transmission network formed by the set of drones, the set of ground users, and the undamaged base stations, a graph model G = (V, E) is constructed; where V = {u1, u2, ..., u...} N ,u N+1} represents a set of nodes containing N drones and one undamaged base station. Denotes the set of edges if and only if the unmanned drone u i and drones j When a communication link is established between them, it makes (u i ,u j )∈E;

[0013] Let Y1 represent the (N+1)×(N+1) adjacency matrix of the graph model G, when (u i ,u j When )∈E, the value of the (i,j)th element Y1(i,j) of Y1 is 1, otherwise the value is 0.

[0014] Preferably, in the data transmission network model, the UAV autonomously selects to act as a role for information transmission recovery; wherein, the role for information transmission recovery includes relay UAV, airborne base station UAV, and backup UAV.

[0015] Preferably, the relay drone is used to perform relay tasks and maintain connectivity between the base station and other drones; the airborne base station drone is used to perform coverage and service tasks, enabling ground users to temporarily restore communication.

[0016] Preferably, establishing the UAV movement model includes:

[0017] The positions of the drone and the ground user are kept constant within each time slot t, and the drone is positioned at a fixed altitude h. u Starting from a certain point, the time for the drone to hover and move horizontally at a constant speed is recorded to construct the drone's movement model.

[0018] Preferably, the UAV movement model is as follows:

[0019]

[0020] Where, β i (t) is based on Center R mov A uniform random distribution on a circle with radius R; mov This represents the maximum distance traveled during each change of spatial location.

[0021] Preferably, the method further includes: using strict potential game theory. Maintaining connectivity means that there exists a global function F: make Satisfy the following formula:

[0022]

[0023] in, It is a collection of players. It is the action set of player i, and Each player i in each time step Choose one action a i To maximize its expected utility function U i That is, the local objective function U i : a -i This indicates the actions of all drones except drone i; (a i ,a -i F(a) represents the combination of actions of all drones; i ,a -i ) represents the potential function of a strict potential game.

[0024] Preferably, the drone i, i.e., u i Select its action from a i Unilaterally changing into another action a′i The difference in their potential functions F is equal to the difference in their utility functions U:

[0025]

[0026] Where 'a' represents the action combination; in the utility function U i In the middle, C i Indicate u i The number of ground users served; This represents the neighbor set of drone i, i.e., the set of drones that can communicate with u. i Other drones that successfully transmitted information; C k The number of ground users served by drone k that is not in that neighborhood cluster; x i,j It is a binary indicator; if the number of served users exceeds a certain value N within the overlapping coverage area between drone i and drone j in its neighbor set. c Then x i,j The value is 0 if it is 0, otherwise it is 1.

[0027] The beneficial effects of this invention are as follows:

[0028] This invention maximizes the utility function (the number of ground users served by the drone) while maintaining network connectivity. It utilizes algebraic connectivity, that is, by introducing the Fiedler eigenvalue of the network topology to update the selection probability, to ensure the reliability of drone data transmission from a macroscopic and long-term perspective, and ultimately enables the drone network to achieve a better balance between coverage, mobility and connectivity. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram of the public safety network structure in an embodiment of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] like Figure 1 As shown, this embodiment considers an emergency rescue scenario. In this scenario, multiple base stations in a public safety network are damaged due to a natural disaster, and a swarm of drones is used to achieve data transmission from ground users (GUs) to undamaged base stations. Let... Let N be the set of drones, where N is the number of drones. Let M be the set of GUs, and M represent the number of GUs.

[0034] To simplify the problem, first use u to represent an undamaged BS in PSN. N+1 Secondly, data transmission in the PSN is modeled using a graph G = (V, E), where V = {u1, u2, ..., u...}. N ,u N+1}and Let u represent the set of nodes and the set of edges respectively, if and only if u i and u j When a communication link is established between them, it makes (u i ,u j Let Y1 denote the (N+1)×(N+1) adjacency matrix of graph G, when (u i ,u j When ∈ E, the (i,j)th element Y1(i,j) of Y1 has a value of 1; otherwise, it has a value of 0. The goal of the UAV swarm is to cover as many ground users as possible. The distribution of ground users follows an independent homogeneous Poisson point process (PPP)Φ with density λ. The UAVs are initially uniformly deployed at an altitude of h in the target area. u The location of the drone. Since the drone moves much faster than the ground user, it can be assumed that the ground user is stationary relative to the drone. Considering role management, the drone will autonomously choose to act as one of three roles for information transmission recovery: Relay UAV (RU), Air BS UAV (BU), and Standby UAV (SU). The functional definitions of the three roles are as follows:

[0035] Relay UAV (RU): Performs relay tasks and maintains connectivity between the base station and other UAVs.

[0036] Air Base Station UAV (BU): Performs coverage and service tasks to temporarily restore communication for ground users.

[0037] Standby UAVs (SUs): Due to their disconnected network topology, they make no contribution to the rescue mission.

[0038] In summary, apart from the backup drone, the other two roles can communicate with each other; only the aerial base station drone has the capability to cover ground users. The sets of the three roles will be represented below as follows: and And their corresponding quantities are represented as N. RU N BU and N SU .

[0039] Moving Model

[0040] According to PKSharma's description, in this embodiment, the drone's motion can be decomposed into vertical motion and horizontal motion (spatial offset). Furthermore, when the average altitude remains constant, vertical motion has a negligible impact on the drone's coverage. Therefore, the drone's mobility model can be simplified to horizontal motion with a certain probability.

[0041] Assume the positions of the drone and the ground user remain constant within each time slot t. Specifically, the drone initially starts at a fixed altitude h. u They start from there. Next, they face a choice with probability p. T The hovering probability is 1-p T It moves horizontally at a constant speed. The dwell time is T. h follow The uniform distribution. If the drone decides to move horizontally, its new spatial position will become...

[0042]

[0043] Where, β i (t) is based on Center R mov It is a uniform random distribution on a circle with radius R. mov This represents the maximum distance traveled during each change of spatial location.

[0044] Channel Model

[0045] To ensure the Quality of Experience (QoE) for UAV services to ground users, Orthogonal Frequency Division Multiple Access (OFDMA) technology is employed, thus ignoring interference between different links. Furthermore, this patent considers free-space path loss models in air-to-air (A2A) and air-to-ground (A2G) wireless channels dominated by line-of-sight (LoS) links. Given that this application scenario focuses on emergency communication restoration in densely populated areas, the UMa-AVLOS channel model is adopted in this embodiment.

[0046] u i and u j The path loss between them can be expressed as:

[0047]

[0048] in, f represents the relative distance between drone i and drone j. c Indicates the carrier frequency. Given from u j to u i Transmission power p i , then u i The signal-to-noise ratio (SNR) at that location is:

[0049]

[0050] Where, σ 2 Indicates noise power.

[0051] From u j to u i The condition for successful data transmission is u i The signal-to-noise ratio at point γ is not less than the threshold γ th .

[0052] Therefore, u i The maximum communication range is:

[0053]

[0054] if That is u i and u j If there is a connection between them, then u i to u j The A2A link was successfully established. Therefore, the maximum coverage area of ​​drone i is:

[0055]

[0056] Therefore, if a ground user is within the coverage area of ​​the drone i, which acts as an aerial base station, then it is considered that the user is accessible to u. i Provide emergency rescue services, i.e., ground users of drone services (UAV-served GU).

[0057] Optimization problem

[0058] Assuming the ground emergency equipment vehicle is very close to the drone, the charging process can be ignored. Therefore, in this scenario, the focus of this patent is to maximize the number of ground users served by the drone swarm over a period of time. Although drone swarms are characterized by flexible deployment and wide coverage, the large distances between them may lead to network topology disconnection, resulting in unreliable information transmission. Therefore, the optimization problem of this patent is to maximize the number of ground users served by the drone while maintaining network connectivity during emergency rescue. A constraint is also imposed such that each aerial base station drone u... i The number of hops in the transmission link to the undamaged BS is less than or equal to h. In this way, the aerial base station drone u i It connects with undamaged BS (Base Station). Therefore, the drone swarm can successfully provide communication services to ground users. Definition:

[0059]

[0060] Among them, matrix The (i,j)th element Indicates from u i to u j The number of hops is less than or equal to the number of paths, h.

[0061] Furthermore, the optimal solution to the optimization problem P is to determine each drone. The role r i , where r i ∈{RU,BU,SU}.

[0062]

[0063] Where, N GU This indicates the number of ground users who were able to restore communication via a maximum of h hops using a drone swarm.

[0064] Drone-assisted coverage and connectivity maintenance based on strict game theory

[0065] The optimization problem, objective P, is essentially a combinatorial problem and an NP-hard problem. This patent aims to achieve a suboptimal solution for the drone swarm, namely, to serve as many ground users as possible during real-time deployment. Therefore, this patent proposes a mechanism in which the positions of the drone swarm and ground users are fixed in each snapshot of the network. The effectiveness of the proposed method will be verified across hundreds of network snapshots. Details of the proposed algorithm are given below.

[0066] First, by introducing strict potential game theory, the roles forming the discrete action space are considered as the strategies of the drone (player), including r(1)(RU), r(2)(BU), and r(3)(SU). The potential function is similar to the concept of potential in vector field analysis, and it conforms to the following definition in strict potential game theory:

[0067] Definition 1 (Strict Potential Game): A game is called a game. For a game to be strictly potential, the condition is that there exists a global function F: make and

[0068]

[0069] in, It is a collection of players. It is a set of actions for player i, and Each player i selects an action at each time step. It maximizes its expected utility, i.e., the local objective function U. i :

[0070] Definition 2 (Nash Equilibrium (NE) in a strict potential game): Action set The condition for a condition known as Nash equilibrium is: if and only if,

[0071]

[0072] Definition 2 states that when the game reaches Nash equilibrium, no player can gain an advantage by changing their actions if all other players keep their actions unchanged. In this patent, P represents the set of drones. a i and a′ i Belongs to the action space This indicates the role selection for drone i. -i This indicates the actions of all drones except drone i. (a i ,a- i This represents the combination of actions for all drones. i Let F(a) represent the utility function of drone i.i ,a -i ) represents the potential function of a strict potential game.

[0073] Considering that the primary optimization objective is to maximize the number of ground users served by the drone, the utility function U i (a) is defined as:

[0074]

[0075] Among them, C i Indicates drone u i The number of ground users served. Indicates drone u i The neighbor set, that is, the set of neighbors that can be sent to u i Other drones that successfully transmitted information. i,j It is a binary indicator. If the number of users served in the overlapping coverage area between drone i and drone j exceeds a certain value N. c Then x i,j The value is 0 otherwise, meaning this patent rewards smaller overlap coverage. Therefore, drones are more likely to choose roles that maximize the number of ground users served and within acceptable overlap coverage. Furthermore, this mechanism can adjust the deployment density of drone swarms; for example, aerial base station drones can be deployed as dispersedly as possible to achieve greater coverage across the overall topology, which helps promote a better balance between coverage and connectivity.

[0076] This patent considers the following potential function F:

[0077]

[0078] Next, by substituting equations (10) and (11) into equation (8), the proof is as follows.

[0079] The drone i selects its action from a i Unilateral change to a i The difference in potential functions between ′ is:

[0080]

[0081] Here, 'a' represents the action combination. When drone k is not a neighbor of drone i, its actions remain unchanged, therefore the fourth line of equation (12) equals 0. When the number of users served by the drone exceeds N... c At that time, x i,j The value is 0, the second row of equation (12) is equal to 0, and the action switching pair of drone i is set. The drones in the middle were unaffected. Otherwise, if x i,jIf the value is 1, then the third row equals 0, and the second row shows the positive impact of the drone i's action switching. Therefore, the function design conforms to the EPG model.

[0082] Regarding rescue missions, although some drones choosing the relay drone role may sacrifice some coverage, they contribute to network connectivity maintenance and energy consumption. After selecting an action in time slot t-1, the drone will select its action in time slot t according to the following strategy:

[0083]

[0084] Where ζ≥0 are the exploration parameters for the UAV to select suboptimal actions. If ζ=0, u i Select actions randomly with equal probability. Otherwise, if ζ→∞, it will choose the best response action with a probability of 1. However, due to the presence of a backup drone role, role selection may cause network disconnection. Therefore, ζ in equation (13) is replaced with Where λ2 originates from That is, the algebraic connectivity of the topology of a multi-UAV network.

[0085] To more intuitively represent the topological structure resulting from role selection, we first construct a subgraph G in graph G = (V, E). sub =(V g E g ), where V g =V\SU and E g =E∩V g ×V g Let n be V g The momentum, then g1 < g2 < ... < g n =N+1. Let Y2 be G sub Given an n×n adjacency matrix, the (i,j)th element y of Y2 is... i,j It is given by the following formula:

[0086] y i,j =Y1(g i ,g j (14)

[0087] G sub The Laplace matrix is:

[0088]

[0089] The second smallest eigenvalue of the Laplace matrix (Feidler eigenvalue) Representing the topological graph G sub The algebraic connectivity of PSN. Since the necessary and sufficient condition for PSN connectivity is... Therefore, in order to satisfy the constraints in equation (7), algebraic connectivity is required. Transform into λ2:

[0090]

[0091] After the above transformation operation, the rational factor ζ of the drone is replaced with On the one hand, if This indicates that the topological connectivity of time slot t-1 (excluding the backup drone) remains good. Therefore, the drone tends to choose the same action again with a higher probability. The higher the connectivity, the greater the probability of the action combination continuing. On the other hand, if If the value is 0 or close to 0, the drone will reselect its role with equal probability to escape the current disconnected topology. Therefore, the designed strategy and utility function enable all relay drones and airborne base station drones to connect with undamaged BSs via multiple hops, not only more strictly satisfying the connectivity constraints of P, but also avoiding drones blindly pursuing coverage. The pseudocode for Algorithm 1, the Drone-Assisted Rescue Maximized Coverage and Connectivity Maintenance Algorithm (EPGCAC) based on strict potential game theory, is shown in Table 1 below:

[0092] Table 1

[0093]

[0094] This invention addresses a typical problem in public safety networks: unmanned aerial vehicles (UAVs) must cover as many ground users (GUs) as possible while ensuring reliable data transmission. To guarantee both local and overall network connectivity, this invention proposes an Exact Potential Game (EPG) based Coverage Maximization Algorithm with Connectivity Preservation for UAV-assisted Rescue (EPGCAC). Specifically, a learning-based role selection mechanism is designed based on the UAV's location and coverage area. Furthermore, the selection probability is updated by incorporating Fiedler features of the network topology, ensuring the reliability of UAV swarm data transmission from both a macroscopic and long-term perspective.

[0095] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for maintaining connectivity in a drone-assisted public safety network, characterized in that, include: A data transmission network model from ground users to undamaged base stations is established based on the public safety network, and a UAV mobility model is also established. By combining the network model and the UAV mobility model with orthogonal frequency division multiple access technology, communication between the UAV swarm and ground users can be maintained. Establishing the data transmission network model includes: Determine the collection of drones and the collection of ground users And the undamaged base stations in the public safety network are used express; A graph model is constructed based on the data transmission network formed by the collection of drones, the collection of ground users, and the undamaged base stations. ;in Indicates inclusion A collection of nodes consisting of a drone and an undamaged base station. Denotes the set of edge sets if and only if there are no drones and drones When a communication link is established between them, it enables ; set up Let G be an (N+1)×(N+1) adjacency matrix of the graph model G, when At that time, the The element The value is 1 if it is not 1, otherwise the value is 0. In the data transmission network model, the UAV autonomously selects to act as a role for information transmission recovery; wherein, the role for information transmission recovery includes relay UAV, airborne base station UAV, and backup UAV; The method also includes: using strict potential game theory. Maintaining connectivity means that a global function exists. ,make Satisfy the following formula: in, It is a collection of players. It is the action set of player i, and Each player i in each time step Choose one action To maximize its expected utility function That is, the local objective function ; This indicates the actions of all drones except drone i. This represents the combination of actions of all drones; The potential function representing a strict potential game; drone i, i.e. Select its action from Unilateral change to another action potential function The difference is equal to its own utility function Difference: Where 'a' represents a combination of actions; in the utility function middle, express The number of ground users served; Indicates drone The neighbor set, that is, the set of neighbors that can be accessed by... Other drones that successfully transmitted information; Drones not located in that neighborhood cluster The number of ground users served; It is a binary indicator; if in a drone drones clustered with its neighbors Within the overlapping coverage area, the number of users being served exceeds a certain value. ,but The value is 0 if it is 0, otherwise it is 1.

2. The method for maintaining connectivity of a drone-assisted public safety network according to claim 1, characterized in that, The relay drone is used to perform relay tasks and maintain connectivity between the base station and other drones; the airborne base station drone is used to perform coverage and service tasks, enabling ground users to temporarily restore communication.

3. The method for maintaining connectivity of a drone-assisted public safety network according to claim 1, characterized in that, Establishing the UAV movement model includes: The positions of the drone and the ground user are kept constant within each time slot t, and the drone is positioned at a fixed altitude. Starting from a certain point, the time for the drone to hover and move horizontally at a constant speed is recorded to construct the drone's movement model.

4. The method for maintaining connectivity of a drone-assisted public safety network according to claim 3, characterized in that, The drone's movement model is as follows: in, Therefore Center R mov A uniform random distribution on a circle with radius R; mov This represents the maximum distance traveled during each change of spatial location.

Images