Method for heterogeneous network communication based on hierarchical opportunistic routing protocol

Abstract

The application is a heterogeneous network communication method based on a layered opportunistic routing protocol, and belongs to the technical field of wireless communication.The application converts the data forwarding problem of a large-scale dynamic cluster scene into an air-ground layered routing forwarding, ground node clustering and unmanned aerial vehicle three-dimensional deployment optimization problem, proposes a layered opportunistic routing protocol, adds an air node for relay transmission, an upper unmanned aerial vehicle uses a geographic routing, and a lower unmanned vehicle uses an opportunistic routing, so as to cope with the routing hole problem in the data transmission process between large-scale ground nodes;in view of a high dynamic environment, the application proposes a fast clustering algorithm and an unmanned aerial vehicle deployment method based on polygon cluster coverage, and can deploy unmanned aerial vehicles in real time to forward data between clusters.The application can guarantee the communication connection of ground nodes in a dynamic large-scale environment, improve energy efficiency and transmission success rate, and solve the problem of routing holes.

Description

Technical Field

[0001] This invention relates to wireless communication technology, and more particularly to a heterogeneous network communication method based on a hierarchical opportunistic routing protocol. Background Technology

[0002] With the continuous development of communication technology and people's ever-increasing demand for the Internet, mobile Internet has permeated all aspects of life. Dynamic unmanned vehicle swarm scenarios, as a special form of mobile self-organizing network, consist of a group of mobile or predefined wireless nodes. They provide a communication method for vehicles operating at high speeds and irregularly due to their independence from any infrastructure, the need for pre-deployed network topology, and the highly dynamic nature of the network topology. However, in dynamic unmanned vehicle swarm scenarios, link interruptions frequently occur due to the mobility of onboard nodes and the highly dynamic changes in network topology, leading to communication disruptions and data transmission failures. Furthermore, due to network bandwidth limitations, transmission performance deteriorates further in dynamic network environments. Therefore, designing routing protocols with high transmission success rates and low latency is a crucial aspect of ensuring the efficient operation of dynamic unmanned vehicle swarms.

[0003] Furthermore, dynamic unmanned vehicle swarms are characterized by their distributed, self-organizing, and highly dynamic nature, classifying them as a type of self-organizing network. Currently, the main routing protocols used in self-organizing networks are the Active Link State Protocol (OLSR) for proactive routing and the On-Demand Distance Vector Protocol (AODV) for on-demand routing. OLSR establishes a global routing table in the early stages of network construction by sending a large amount of signaling information to serve subsequent message forwarding. AODV explores the optimal transmission path for the message to be sent by sending signaling packets before sending the message. Both of these routing protocols are table-based, and due to the support of routing tables, they exhibit low transmission latency and high transmission reliability after the routing tables are established. However, they send a large amount of signaling information during routing table establishment, consuming valuable communication resources and energy. Maintaining the routing table requires periodic signaling, resulting in poor scalability. The size of the stored routing table is quadratically proportional to the number of nodes, placing a significant storage burden on large-scale networks. Geographic routing algorithms have been proposed to utilize geographic information for data communication. These methods only require low-frequency broadcast signaling information carrying their own coordinates, storing a neighbor information table, and transmitting data using a greedy approach. Compared to table-based routing protocols, they have advantages in scalability, energy efficiency, and low cost. However, geographic routing protocols are easily affected by routing holes, leading to increased data transmission latency and reduced success rate. Furthermore, because data transmission is based on a greedy approach, link congestion can easily occur, rapidly consuming the energy of a particular node. Summary of the Invention

[0004] To address the aforementioned problems, the present invention aims to provide a heterogeneous network communication method based on a hierarchical opportunistic routing protocol for use in large-scale, highly dynamic unmanned vehicle cluster scenarios. By deploying unmanned aerial vehicle nodes as relays for data forwarding, it solves problems such as communication link obstruction and routing holes caused by topology changes in ground unmanned vehicle nodes, thereby improving the energy efficiency of ground nodes and the data transmission success rate.

[0005] This invention provides a heterogeneous network communication method based on a hierarchical opportunistic routing protocol, specifically for autonomous vehicle cluster scenarios, comprising the following steps:

[0006] (1) First, add aerial drone nodes to the ground unmanned vehicle nodes to build a two-layer network; the upper layer is a network composed of aerial drone nodes, which move in three-dimensional space and forward data between nodes through geographic routing; the lower layer is a network composed of ground unmanned vehicle nodes, which move in two-dimensional ground plane and forward data between nodes through opportunistic routing.

[0007] (2) Next, the unmanned vehicle nodes are clustered in real time, and the clustering results are broadcast to each unmanned vehicle; a representative node is selected for each cluster, and the representative node collects the location and perception data of each unmanned vehicle in the cluster and uploads it to the drone.

[0008] (3) Obtain the location of all unmanned vehicle representative nodes and use the polygon overlay method to deploy the location of aerial UAV nodes;

[0009] (4) The UAV node obtains representative nodes and UAV nodes within the communication coverage area and establishes a neighbor list;

[0010] (5) Based on the cluster structure, the communication of unmanned vehicle nodes is divided into inter-cluster communication and intra-cluster communication; for communication between nodes within the same cluster, only opportunistic routing is used for transmission; for communication between nodes across clusters, aerial UAV nodes are called for relay transmission.

[0011] When an autonomous vehicle node sends a data packet, it obtains the clusters of the source and destination nodes from the packet header. If the source and destination nodes are in the same cluster, the autonomous vehicle node directly broadcasts the data packet. If the source and destination nodes are not in the same cluster, the autonomous vehicle node finds the nearest connected drone node, records the drone node as the forwarding address in the packet header, and broadcasts the data packet.

[0012] (6) After receiving the data packet broadcast by the unmanned vehicle node, the UAV node checks whether the forwarding address in the data packet header is itself. If not, the data packet is discarded. If it is, it determines whether the target node is within the communication coverage of the UAV. If so, the UAV node forwards the data packet directly to the target node. Otherwise, the UAV node searches its own neighbor list for the neighbor node closest to the target node as the next hop node and then forwards the data packet.

[0013] In this invention, the DBSCAN clustering algorithm based on signal-to-noise ratio and obstacle information is used to cluster unmanned vehicle nodes in real time, including: first, setting the core node and the nodes. - Method for determining the neighborhood: For the set of autonomous vehicle nodes In the node i, the node i's - A subset of the neighborhood The signal-to-noise ratio (SNR) of node i in the set of autonomous vehicle nodes S is not less than a threshold. The set of nodes that are not blocked by obstacles from node i; if the subset of the β-neighborhood of node i At least include If there are 1 node, then node i is a core node; then, traverse the nodes in the autonomous vehicle node set S and use the DBSCAN clustering algorithm to divide them into clusters; select a representative node for each cluster, which is the core node with the most β-neighborhood subset nodes in the cluster.

[0014] In the method of this invention, the deployment of aerial UAV nodes using a polygon coverage method includes: selecting the polygon with the smallest area that covers all representative nodes, then traversing each position within the polygon to obtain representative nodes covered by a circle with the deployment position as the center and the UAV communication range as the radius, maintaining a linked list that stores the representative nodes covered by each position within the polygon; sorting the positions according to the number of covered representative nodes, and selecting the H positions that cover all representative nodes and have the largest number of covered representative nodes as the deployment positions of the H aerial UAV nodes.

[0015] In the method of this invention, when the target node is outside the communication coverage of the UAV and none of the UAV node's neighboring nodes can connect to the target node, the UAV node will forward data packets to the ground unmanned vehicle node within its communication coverage that is closest to the target node.

[0016] The advantages and positive effects of this invention are as follows:

[0017] (1) The method of the present invention designs a hierarchical routing protocol based on the characteristics of heterogeneous network nodes. This protocol combines the high energy efficiency of opportunistic routing with the low latency of geographical routing.

[0018] (2) The method of the present invention restricts empty nodes to cross-cluster communication only, thus avoiding the problem of congestion of empty nodes.

[0019] (3) The method of the present invention designs a clustering algorithm for ground nodes based on signal-to-noise ratio and terrain information. By deploying air nodes to assist in the forwarding of ground node information, the problem of routing gaps is avoided to the greatest extent. Attached Figure Description

[0020] Figure 1 This is a flowchart of the heterogeneous network communication method based on the hierarchical opportunistic routing protocol of the present invention;

[0021] Figure 2 This is a schematic diagram of the structure of the IP packet of the air node in an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of a physical and communication simulation scenario according to an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0024] The heterogeneous network communication method based on hierarchical opportunistic routing protocol of the present invention first establishes a dynamic unmanned vehicle cluster model assisted by multiple UAVs. The established network includes two layers: the upper layer is the UAV platform, which moves in three-dimensional space and forwards data through geographical routing; the lower layer is the unmanned vehicle nodes, which move in two-dimensional ground plane and forward data through opportunistic routing.

[0025] The method of this invention uses an opportunistic routing algorithm to improve energy efficiency and increase transmission success rate in the data transmission of ground nodes. In order to deal with the routing hole problem, an aerial UAV node is added to the ground nodes for data relay transmission. For this purpose, a hierarchical opportunistic routing protocol is used for communication between air and ground heterogeneous networks. The lower-level unmanned vehicles use opportunistic routing, and the upper-level UAVs use geographic routing.

[0026] When dealing with upper-layer air nodes, due to their powerful communication capabilities and favorable line-of-sight communication environment, this invention employs geographic routing. Each air node encapsulates its location, identifier, and other information in the packet header and obtains the location information of ground nodes by periodically exchanging control messages. This allows for the establishment of neighbor lists and routes between nodes, thereby enabling data packet transmission between the air and ground. The geographic routing protocol, when used with the lower-layer autonomous vehicle cluster as an opportunistic routing protocol, does not require the establishment of routing tables and neighbor tables, nor does it need to send signaling information. It utilizes wireless network broadcasting and modifies the IP packet structure to add multiple destination addresses as candidate lists, improving node energy efficiency. Furthermore, due to the opportunistic routing's broadcast-then-competition characteristic, idle nodes are more likely to win the competition, thus reducing the likelihood of node congestion.

[0027] This invention presents a heterogeneous network communication method based on a hierarchical opportunistic routing protocol. To effectively deploy unmanned aerial vehicles (UAVs) and determine when to use airborne relay nodes for communication, the underlying ground-based UAV nodes are divided into multiple clusters. Based on the cluster structure, communication between ground nodes is divided into inter-cluster communication and intra-cluster communication. Representative nodes from each cluster are extracted to collect UAV data within the cluster. For communication between nodes within the same cluster, only lower-level opportunistic routing is used for transmission; for communication between nodes across clusters, airborne nodes are invoked for relaying. Furthermore, this invention proposes a DBSCAN (Density-Based Clustering with Noise) clustering algorithm based on signal-to-noise ratio and obstacle information, which can perform real-time clustering of ground nodes and dynamically allocate cluster information. Simultaneously, addressing the deployment problem of UAVs serving as relays for inter-cluster communication, this invention also proposes a polygon cluster coverage method to deploy UAV locations and provide communication services to ground nodes.

[0028] The heterogeneous network communication method based on a hierarchical opportunistic routing protocol according to embodiments of the present invention, such as... Figure 1 As shown, it mainly includes the following 6 steps.

[0029] Step 1: Initialize the node environment and obtain information such as the number of nodes, their locations, and obstacle information in the environment.

[0030] Initialize network nodes. The network nodes in this invention include ground nodes and air nodes. Obtain the location and number of ground unmanned vehicle nodes, and obtain the number of air nodes. Initialize environmental information and obtain obstacle information in the environment.

[0031] Step 2: The DBSCAN clustering algorithm based on signal-to-noise ratio and obstacle information clusters and groups ground nodes.

[0032] In the environment, due to factors such as obstruction by obstacles or excessive distance between nodes, connectivity between all nodes cannot be guaranteed. To enable aerial nodes to exert strong communication capabilities and good line-of-sight communication capabilities, this invention designs a DBSCAN clustering algorithm based on signal-to-noise ratio and obstacle information to divide the dynamic unmanned vehicle (UAV) cluster into inter-clusters. Multiple UAVs are deployed as relay nodes at network bottlenecks for routing and forwarding. UAV nodes within the same cluster transmit data through an opportunistic routing protocol, sending data to the representative node within the cluster. Subsequently, full data collection can be completed by exploring only the representative node of each cluster. When reaching the vicinity of a UAV node, data information of the current UAV and its multi-hop neighbor nodes can be collected simultaneously, which greatly shortens the path.

[0033] Let the set of ground-based autonomous vehicle nodes be considered in the current scenario. n is the number of autonomous vehicle nodes. This represents autonomous vehicle node i. Parameter settings are as follows. This indicates whether communication between node i and node j is blocked by an obstacle. The value is either 1 or 0, representing whether communication between nodes i and j is blocked by an obstacle or not, respectively. This embodiment of the invention sets a signal-to-noise ratio threshold. Define parameters This indicates the density of the distribution of nodes in the neighborhood. The signal-to-noise ratio (SNR) of the neighborhood of any node i is greater than or equal to... The number of nodes. It is defined as follows:

[0034] - Neighborhood: For ,That -The neighborhood contains a set of S pairs of autonomous vehicle nodes. The signal-to-noise ratio is not less than β, and there is no subset of samples blocked by obstacles. , where β takes the value . The number of nodes is .

[0035] Core nodes: For If the node subsets corresponding to the β-neighborhood At least include If there are 1 node, then It is a core node.

[0036] Representative node: For a cluster Select the core node with the most nodes in the β-neighborhood subset within the cluster, i.e., the node with the most nodes in the current cluster k. The core node serves as the representative node Where m represents the total number of clusters, and C represents the cluster set. This represents cluster k.

[0037] If a cluster contains multiple instances with the same The core node is then randomly selected as the representative node. In this embodiment of the invention, , All settings can be adjusted based on experience.

[0038] The DBSCAN clustering algorithm based on signal-to-noise ratio and obstacle information of this invention can acquire the set of unmanned vehicle nodes in real time. Given the node locations, traverse the set of autonomous vehicle nodes. For unmanned vehicle nodes, DBSCAN clustering is performed. Based on obstacle information, the set of unmanned vehicle nodes with the highest density of connected nodes is searched using density reachability relationships to form clusters. The DBSCAN clustering algorithm in this embodiment of the invention performs the following steps:

[0039] a) Traverse the set of autonomous vehicle nodes Find all core nodes and the core nodes' - Neighborhood. The set of nodes in the labeled set S that have not been assigned to a cluster is... .

[0040] b) Randomly select a core node i that has not yet been assigned to a cluster, and create a cluster with node i as the core. , will node i - A subset of the neighborhood And node i itself is assigned to a cluster In the middle. Update the collection. Nodes that will be assigned to clusters will be deleted from this list.

[0041] c) Examine node i - Neighborhood subset For each node j within the cluster, update the cluster. If node j is also a core node, then: remove node j from the cluster. Delete, and create a cluster with node j as the core. , and the set and node j - The intersection of the neighborhood subsets is assigned to the cluster. In the middle, traverse the clusters again. Middle node, delete the one that is related to the cluster. The connections between all nodes collide with nodes that are obstacles. Then, the physical distance between core node i and core node j is calculated. If the distance is less than a preset cluster distance threshold, the two clusters are considered sufficiently close, and the clusters are merged. Merge into cluster In the middle; otherwise, cluster Not merged into cluster Execute step d on core node j.

[0042] After traversing node i - After each node in the neighborhood subset, execute step e.

[0043] d) Examine the core node j according to the process in step c. - Update the cluster for each node in the neighborhood. Middle node.

[0044] e) Continue to randomly select a core node to be assigned to a cluster, and repeat steps bd above until all core nodes in the autonomous vehicle node set S have been traversed, ultimately forming multiple clusters. If all core nodes are assigned to clusters, the set... There are still nodes that act as outliers, and each node forms a cluster.

[0045] f) Select a representative node for each cluster. The representative node collects information perceived by all autonomous vehicle nodes in the cluster through opportunistic routing, including information such as the surrounding environment and current location, and broadcasts this information.

[0046] Each cluster collects information within the cluster and uploads it to the drone through a representative node. In this way, when deploying the drone, only the representative node needs to be considered, and the drone can directly receive the data packets from the representative node, reducing the computational complexity caused by the large number of unmanned vehicle nodes when deploying the drone.

[0047] Step 3: Deploy upper-level nodes based on geographical location.

[0048] Compared to ground nodes, air nodes have a wider communication range, greater communication bandwidth, are unaffected by terrain, and can reach the autonomous vehicle node more quickly to collect data. This invention designs an air node deployment algorithm to deploy air nodes to cover as many distant clusters as possible, enabling cross-cluster data transmission.

[0049] Assuming H aerial nodes are deployed, this invention uses a polygon coverage method to deploy their locations. In this method, the locations of representative nodes for each cluster of unmanned vehicles (UAVs) are first obtained. Then, the polygon with the smallest area is selected to cover the representative nodes of each cluster. Using the deployment location as the center and the UAV's communication range as the radius, each location within the polygon is traversed, maintaining a linked list containing the representative nodes covered at the current location. Finally, the linked list is sorted according to the number of covered nodes, and the H locations that cover all representative nodes and have the largest number of covered representative nodes are selected as the deployment locations for the H aerial nodes. This deployment of aerial nodes effectively enables cross-cluster communication and avoids routing gaps.

[0050] Step 4: The lower-level ground nodes read the clusters where the source node and the target node are located from the header, and forward the data according to the cluster ID information.

[0051] Data transmission between ground nodes uses signaling-free opportunistic routing. This protocol does not require the creation of routing tables, thus avoiding the consumption of node storage space and channel resources for storing and building routing tables. Airborne nodes maintain geographical routing records in their neighbor tables, enabling them to quickly find the optimal node for data forwarding. This invention combines the advantages of both protocols, improving network energy efficiency and reducing transmission latency.

[0052] Each aerial drone node encapsulates its location, identifier, and other information in the packet header. By periodically exchanging control messages, it obtains the location information of the ground-based drone representative node, establishes a neighbor list, and forms a geographical route between the aerial and ground nodes, thereby enabling data packet transmission between the air and ground. The drone's neighbor list records not only the drone representative node within its communication range but also other drone nodes within that range. The drone node obtains the geographical location and IP address information of various ground nodes within the cluster of its representative node.

[0053] When a drone node receives a data packet from a vehicle node that needs to be forwarded, it will use geographical routing to forward the data. For example... Figure 2 As shown, in the geographic routing protocol used by the drone, the drone stores the source and destination addresses, forwarding addresses, and current address in the IP header when transmitting data. In the IP header, the mode indicates the routing protocol's mode, divided into three types: drone, greedy, and backup, with a size of 2 bits. The source and destination clusters record the clusters of the source and destination nodes, respectively, with a size of 15 bits. Each type of address has a size of 4 bits (32 bits), and each type of location is stored as two-dimensional coordinates, with a size of 2 bits (16 bits).

[0054] Step 5: For communication within the same cluster or for ground nodes without air node connections, data packets are broadcast directly using an opportunistic routing protocol. For inter-cluster communication, ground nodes search for the nearest reachable air node, record the node's IP address in the header, and then broadcast the data packet.

[0055] To avoid congestion issues with air nodes, restrictions are placed on their use for message forwarding. Air nodes only function when the source and destination nodes are not in the same cluster. When a ground node sends a message packet, it first reads the clusters of the source and destination nodes from the header. For cross-cluster communication, the ground node searches for the nearest reachable drone node, records its IP address as the forwarding address in the header, and broadcasts the data packet. For intra-cluster communication or when no air node is available, the data packet is broadcast directly. After broadcasting, the sending node starts a timeout countdown. If no message forwarding is detected by the end of the countdown, the message forwarding is considered to have failed, and the node enters backup mode, using a temporary neighbor table to find the optimal node to forward the data packet.

[0056] In this protocol, to avoid hidden routing issues, the candidate forwarding domain is set as a sector with the sending node as the vertex, the destination node's direction as the positive direction, and an angle of 60 degrees. After a node within the candidate forwarding domain receives a data packet, it calculates a dynamic latency penalty. This setting ensures that, in unobstructed conditions, all nodes within the forwarding domain can monitor the sending activity of other nodes. The dynamic latency penalty (DDC) setting in this protocol is negatively correlated with the message progress, as follows:

[0057]

[0058] in, The maximum dynamic delay is represented by r, which is the communication radius of the sending node, and p is the length between the point on the line connecting the sending node and the target node where the received data packet is projected and the sending node.

[0059] Nodes within the forwarding domain calculate the dynamic latency and set a countdown timer equal to the sum of the queuing latency and the dynamic latency. When the countdown ends, they forward the data packet. When other nodes within the forwarding domain detect that a data packet has been forwarded, they stop their own countdown and delete the data packet. When a sending node detects that a data packet has been forwarded, it stops its own timeout countdown and broadcasts a stop acknowledgment character (ACK). When a node within the forwarding domain receives the stop ACK, nodes still in the countdown phase stop their countdown and delete the data packet.

[0060] Step 6: When an airborne node receives a data packet, it first checks if the forwarding address in the packet header matches its own. If not, the packet is discarded without further processing. If it does match, the node finds the nearest neighbor node in its neighbor list as the next hop and records it in the forwarding address of the packet header. If the node is the closest airborne node to the target node (i.e., the target node is a drone within the drone's communication coverage area), the node directly forwards the data packet to the target node.

[0061] When the target node is outside the drone's communication coverage area, the drone node searches its neighbor list to find the nearest neighbor node. If no neighbor node can directly connect to the target node, meaning the drone node cannot connect to the target node, the drone node forwards the data packet to the ground unmanned vehicle node within its communication coverage area that is closest to the target node.

[0062] To evaluate the performance of the method of the present invention, embodiments of the present invention construct a semi-simulation system that can simulate communication and physical environments, such as... Figure 3 As shown, the physical system simulates the clustering of ground nodes and the deployment of airborne nodes, while the communication system simulates channel occupancy and communication quality during information transmission. Experimental verification, considering both the energy efficiency and data acquisition success rate of the sensor network, demonstrates that when deploying multiple UAVs, the method of this invention achieves 81.2% of the work of OLSR and AODV using only 0.8% of the energy required by OLSR and 11.57% of the energy required by AODV. For data collection in large-scale dynamic ad hoc network scenarios, the routing algorithm proposed in this invention exhibits optimal performance. AODV represents the on-demand plane distance vector routing protocol for wireless ad hoc networks, and OLSR represents the optimized link-state routing protocol.

[0063] In summary, this embodiment addresses the task of providing communication relay services to ground users in a large-scale dynamic ad hoc network scenario within an inter-cluster UAV network. It utilizes the proposed DBSCAN algorithm based on signal-to-noise ratio and occlusion to perform real-time clustering of ground nodes and dynamically allocate cluster information. Next, multi-UAV deployment with polyhedral cluster coverage ensures coverage of communication services for ground users and improves system reachability. Finally, an air-to-ground hierarchical opportunistic routing protocol is used for communication between nodes in the dynamic UAV cluster, ensuring the continuity of communication links.

[0064] Except for the technical features described in the specification, all other technologies are known to those skilled in the art. Descriptions of well-known components and technologies are omitted in this invention to avoid redundancy and unnecessary limitation. The embodiments described above do not represent all embodiments consistent with this application. Various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this invention are still within the protection scope of this invention.

Claims

1. A heterogeneous network communication method based on a hierarchical opportunistic routing protocol, characterized in that, Includes the following steps: (1) Add aerial drone nodes to the ground unmanned vehicle nodes to build a hierarchical network. The upper layer is a network composed of aerial drone nodes, and the lower layer is a network composed of ground unmanned vehicle nodes. (2) Cluster the unmanned vehicle nodes in real time and broadcast the clustering results to each unmanned vehicle; For each cluster, a representative node is selected. The representative node collects the location and perception data of each unmanned vehicle within the cluster and uploads it to the drone. The communication method uses the DBSCAN clustering algorithm based on signal-to-noise ratio and obstacle information to cluster unmanned vehicle nodes in real time, including: first, setting the core node and nodes. - Method for determining the neighborhood: For the set of autonomous vehicle nodes In the node i, the node i's - A subset of the neighborhood The signal-to-noise ratio (SNR) of node i in the set of autonomous vehicle nodes S is not less than a threshold. The set of nodes that are not blocked by obstacles from node i; if the subset of the β-neighborhood of node i At least include If there are 1 node, then node i is a core node; then, traverse the nodes in the autonomous vehicle node set S and use the DBSCAN clustering algorithm to divide them into clusters; select a representative node for each cluster, which is the core node with the most β-neighborhood subset nodes in the cluster; (3) Obtain the location of all unmanned vehicle representative nodes and use the polygon overlay method to deploy the location of aerial UAV nodes; The deployment of aerial UAV nodes using the polygon coverage method includes: selecting the polygon with the smallest area that covers all representative nodes; then traversing each position within the polygon to obtain representative nodes covered by a circle with the deployment position as the center and the UAV communication range as the radius; maintaining a linked list that stores the representative nodes covered by each position within the polygon; sorting the positions according to the number of covered representative nodes; and selecting the H positions that cover all representative nodes and have the largest number of covered representative nodes as the deployment positions for the H aerial UAV nodes. (4) The UAV node obtains representative nodes and UAV nodes within the communication coverage area and establishes a neighbor list; (5) Based on the cluster structure, the communication of unmanned vehicle nodes is divided into inter-cluster communication and intra-cluster communication; unmanned vehicle nodes within the same cluster only use opportunistic routing for data forwarding; communication between unmanned vehicle nodes across clusters calls on drone nodes for relay transmission. When an autonomous vehicle node sends a data packet, it obtains the clusters of the source and destination nodes from the packet header. If the source and destination nodes are in the same cluster, the autonomous vehicle node directly broadcasts the data packet. If the source and destination nodes are not in the same cluster, the autonomous vehicle node finds the nearest connected drone node, records the drone node as the forwarding address in the packet header, and broadcasts the data packet. (6) After receiving the data packet broadcast by the unmanned vehicle node, the UAV node checks whether the forwarding address in the data packet header is itself. If not, the data packet is discarded. If it is, it determines whether the target node is within the communication coverage of the UAV. If so, the UAV node forwards the data packet directly to the target node. Otherwise, the UAV node searches its own neighbor list for the neighbor node closest to the target node as the next hop node and then forwards the data packet.

2. The method according to claim 1, characterized in that, The communication method described herein acquires in real time the location and number of unmanned vehicle nodes, obstacle information in the environment, and the number of drone nodes in the application scenario.

3. The method according to claim 1, characterized in that, The DBSCAN clustering algorithm based on signal-to-noise ratio and obstacle information includes: a) Traverse the set of autonomous vehicle nodes Find all core nodes and the core nodes' - Neighborhood; the set of nodes in the label set S that have not been assigned to a cluster is ; b) Randomly select a core node i that has not yet been assigned to a cluster, and create a cluster with node i as the core. , will node i - A subset of the neighborhood And node i itself is assigned to a cluster In the middle; update the set Nodes that will be assigned to clusters will be deleted from this list; c) Examine node i - Neighborhood subset For each node j within the cluster, update the cluster. If node j is also a core node, then: remove node j from the cluster. Delete, and create a cluster with node j as the core. , and the set and node j - The intersection of the neighborhood subsets is assigned to the cluster. In the middle, traverse the clusters again. Middle node, delete the one that is related to the cluster. The connections between all nodes collide with nodes that are obstacles; then the distance between the physical locations of core node i and core node j is calculated. If the distance is less than a preset cluster distance threshold, the two clusters are considered sufficiently close, and the clusters are merged. Merge into cluster In the middle; otherwise, cluster Not merged into cluster Execute step d on core node j; After traversing node i - After each node in the neighborhood subset, execute step e; d) Examine the core node j according to the process in step c. - Update the cluster for each node in the neighborhood. Middle node; e) Continue to randomly select a core node to be assigned to a cluster, and repeat steps bd above until all core nodes in the autonomous vehicle node set S have been traversed, ultimately forming multiple clusters; if all core nodes have been assigned to clusters, the set... There are still nodes in the cluster; each node is divided into a cluster. f) Select a representative node for each cluster. The representative node collects information perceived by all autonomous vehicle nodes within the cluster through opportunistic routing.

4. The method according to claim 1, characterized in that, In the method described, when the autonomous vehicle node generates and sends a data packet, it adds the cluster where the source node is located and the cluster where the target node is located to the header of the data packet.

5. The method according to claim 1, characterized in that, In the method described, when the target node is outside the communication coverage of the drone and none of the drone node's neighboring nodes can connect to the target node, the drone node will forward data packets to the ground unmanned vehicle node within its communication coverage that is closest to the target node.

Images