An adaptive UAV cooperative opportunistic routing method based on grey perception

Abstract

The present invention discloses an adaptive UAV collaborative opportunistic routing method based on gray perception. This method uses link stability, node mobility adaptability, and residual energy as metrics for selecting candidate relay nodes for opportunistic routing. In link stability calculations, a gray prediction model is used to model and predict the historical movement information of the UAV, addressing the uncertainty of the UAV's position and speed. The node's mobility adaptability considers not only the geographic location of the candidate node and the destination node, but also the candidate node's own subsequent forwarding capabilities. The UAV's residual energy factor can ensure overall network load balancing. In addition, when selecting relay nodes, they can be dynamically adjusted according to different mission requirements.

Description

Technical Field

[0001] The present invention belongs to the field of unmanned aerial vehicle (UAV) networking, and in particular relates to an adaptive UAV collaborative opportunistic routing method based on grey perception. Background Art

[0002] With the emergence of drones in both military and civilian applications, flying ad-hoc networks (FANETs), comprised of multiple drones, have attracted widespread attention. As a new type of airborne ad-hoc network, FANETs feature higher node mobility and longer inter-node distances compared to traditional mobile ad-hoc networks (MANETs). However, due to the time-varying network topology and dynamic links, maintaining continuous communication during mission execution is challenging. Traditional routing protocols, including processes such as route discovery and routing table maintenance, require a complete route to be found before data is sent. These protocols maintain high efficiency in environments with minimal network topology changes. However, due to the rapid topology changes in drone swarm networks, using traditional routing protocols incurs significant routing overhead, degrading overall network performance. Therefore, designing a routing protocol for FANETs that ensures data transmission quality and improves communication efficiency is a challenging task.

[0003] Opportunistic routing (OR), also known as any-path routing, is a routing technology that leverages the broadcast nature of wireless networks. By dynamically selecting multiple routes, opportunistic routing can improve link reliability and overall system throughput. Unlike traditional routing, which preselects a designated relay node for each transmission, OR broadcasts a data packet to a group of candidate relay nodes. The candidate relay nodes that successfully receive the packet then run a coordination strategy to select the optimal relay to forward the packet. Data packets are transmitted over any possible link, rather than a designated link. OR is equivalent to having an additional backup link, reducing the probability of transmission failure. OR considers all possible links in a single transmission, including high-quality short-range links and low-quality long-range links. Therefore, a single transmission can jump directly to the farthest relay node that successfully received the packet, improving routing efficiency. Summary of the Invention

[0004] The purpose of this invention is to propose a gray-aware adaptive UAV collaborative opportunistic routing method based on a gray prediction model to reduce link instability caused by the uncertainty of UAV movement. Relay nodes are selected for routing transmission based on the node's mobility adaptability and remaining energy. To achieve this goal, the steps employed by this invention are as follows:

[0005] Step 1: Comprehensively process the historical records of the UAV's motion speed and direction, and establish a gray prediction model for the UAV's motion state. Using the metabolic model GM (1, 1), the sliding window method is used to keep the modeling length unchanged, continuously add new values, and remove old values to predict the next motion speed and direction of the UAV. The specific method is as follows:

[0006] Let the original data series be x (0) ={x (0) (1), x (0) (2),…,x (0) (n)}, where x (0) (k) is the data at time k, let x (1) (1) = x (0) (1), then x (0) The cumulative generated sequence of (1) is x (1) ={x (1) (1), x (1) (2),…,x (1) (n)}, where For the metabolic GM (1, 1) prediction model, its grey differential equation is:

[0007] x (0) (k)+az (1) (k) = b (1)

[0008] Where: a and b are constants, x (1) The expression of (t) is:

[0009]

[0010] Discretize it and get:

[0011]

[0012] Finally, the predicted value is:

[0013]

[0014] The true value x at time l (0) (l) and predicted value The difference between the two becomes the residual at time l, denoted as e (0) (l);

[0015] Step 2: Use the grey prediction model to determine the speed and direction of the drone at the next moment. Calculate the link stability and mobility adaptability between drone nodes. The specific method is:

[0016] The link stability and node mobility adaptability are defined as follows:

[0017] Assume that the initial distance between node A and node B is When calculating, the two moving nodes are considered as one node being stationary and the other node being in relative motion. The position of node A relative to node B is (x AB ,y AB )=(x B -x A ,y B -y A ). After time t, the position of node A relative to B is:

[0018]

[0019] where θ AB is the velocity angle of node A relative to node B, v AB is the speed of node A relative to B; when the distance between nodes A and B is the maximum communication distance R of the drone, the link between A and B is disconnected, that is:

[0020]

[0021] Substituting formula (5) into (6) we can obtain:

[0022]

[0023] v AB =v A -v B =(|v A |cosθ A -|v B |cosθ B )i+(|v A |sinθ A -|v B |sinθ B )j (8)

[0024] Among them, θ A is the moving direction of node A, θ B is the moving direction of node B, v A and v B are the velocities of nodes A and B respectively. Since t cannot be negative, the value of t is:

[0025]

[0026] Therefore, the stability of the link LS is defined as AB for:

[0027]

[0028] In the UAV cluster, there are active nodes s and relay nodes N s (1) Relay node N s (2) and the destination node d, assuming that the source node s is stationary, the movement speeds of the other three nodes are the velocity vectors relative to node s, and the angle ω between the movement direction of the candidate relay node i and the movement direction of the destination node d i The expression is:

[0029]

[0030] ω i The value range is [0, π]. When the node moves to the opposite direction of the target node, that is, when ω i When it is greater than 0.5π, it will be deleted from the relay node candidate set;

[0031] Each node i maintains a set of neighbors Np(i), where Np(i) represents the nodes that are closer to the destination node than node i after screening. When a node forwards a data packet to its neighbors, it inserts its own position into the data packet. When a node receives a data packet, it compares the position in the data packet with its own and updates its better neighbor set. The mobility adaptability metric PM(i) of node i is:

[0032]

[0033] Among them, d i，d and d s，d are the Euclidean distances from nodes i and s to the target node d, respectively. N(i) is the set of candidate nodes maintained by node i before receiving the routing request broadcast by node s.

[0034] Assume that E0 and E r The initial energy value and the remaining energy value of each UAV node are,in an ad hoc network, a mobile node can be selected as a forwarder only under the following conditions:,(1) It has enough energy E,within the effective time of the link with the sender. min1 Send data packets; (2) After the link validity time, the node still has enough energy E min2 Move to the designated location, the minimum threshold E for drone data forwarding min It consists of two parts: E min =E min1 +E min2 The energy index value calculation method is shown in formula (13). If the energy value is 0, the node should be removed from the candidate relay node set:

[0035]

[0036] In summary, based on the link stability prediction value, node mobility adaptability and energy measurement value, a priority measurement equation is proposed. The priority measurement equation of node i in the candidate relay set of the drone source node s is:

[0037] Val(N s (i)) = αLS(N s (i))+βPM(N s (i))+γEne(N s (i)) (14)

[0038] Among them, α, β, and γ are the weight coefficients of the three metrics, reflecting their importance in the priority metric calculation. Candidate relay nodes calculate their own priority metrics according to the above formula to sort candidate nodes and filter candidate node sets;

[0039] Step 3: When the source node generates a data packet or a data packet arrives at a drone node, the node extracts the destination information from the data packet to check whether it is the destination node. If so, the data packet forwarding process ends after confirmation. If not, it enters the candidate relay set selection phase. Based on the periodically maintained neighbor node information, the node broadcasts a routing request to the nodes within the communication range. After receiving the routing request, the nodes in the candidate relay set recalculate the node's link stability, mobility adaptability, and residual energy value based on their own status information and the source node's status information, and eliminate candidate nodes that do not meet the minimum requirements. In the node mobility adaptability, the node moves in the opposite direction of the destination node, or its own residual energy is lower than the threshold value. The Val value is calculated using the priority metric equation, and a route reply is finally made. After receiving a route reply from a neighboring node, the forwarding node updates its candidate relay set. If the number of nodes in the candidate relay set is zero, the node enters "store-carry-forward" mode and attempts new route discovery in the next cycle. If the number of nodes in the candidate relay set is not zero after filtering, the candidate node sets a corresponding timer size based on Val. The larger the Val value, the higher the priority of the node, the greater its contribution to routing, and the smaller the timer value set. When the candidate relay node's own timer expires, it broadcasts a route reply. Other candidate relay nodes automatically discard the data packet after receiving this broadcast. After that, the node becomes a forwarding node again and begins a new round of route discovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a flow chart of the method proposed by the present invention;

[0041] Figure 2 It is a schematic diagram of the link stability between drone nodes;

[0042] Figure 3It is a schematic diagram of the movement direction of candidate nodes and destination nodes; DETAILED DESCRIPTION

[0043] The present invention is further described in detail below with reference to the accompanying drawings and examples.

[0044] The flowchart of the adaptive UAV cooperative opportunity routing method based on grey perception proposed in this invention is shown in the attached figure. Figure 1 The specific steps include:

[0045] Step 1: Comprehensively process the historical records of the UAV's motion speed and direction, and establish a gray prediction model for the UAV's motion state. Using the metabolic model GM (1, 1), the sliding window method is used to keep the modeling length unchanged, continuously add new values, and remove old values to predict the next motion speed and direction of the UAV. The specific method is as follows:

[0046] Let the original data series be x (0) ={x (0) (1), x (0) (2),…,x (0) (n)}, where x (0) (k) is the data at time k, let x (1) (1) = x (0) (1), then x (0) The cumulative generated sequence of (1) is x (1) ={x (1) (1), x (1) (2),…,x (1) (n)}, where For the metabolic GM (1, 1) prediction model, its grey differential equation is:

[0047] x (0) (k)+az (1) (k) = b (15)

[0048] Where: a and b are constants, x (1) The expression of (t) is:

[0049]

[0050] Discretize it and get:

[0051]

[0052] Finally, the predicted value is:

[0053]

[0054] The true value x at time l(0) (l) and predicted value The difference between the two becomes the residual at time l, denoted as e (0) (l);

[0055] Step 2: Use the grey prediction model to determine the speed and direction of the drone at the next moment. Calculate the link stability and mobility adaptability between drone nodes. The specific method is:

[0056] The link stability and node mobility adaptability are defined as follows:

[0057] Assume that the initial distance between node A and node B is When calculating, the two moving nodes are considered as one node being stationary and the other node being in relative motion. The position of node A relative to node B is (x AB ,y AB )=(x B -x A ,y B -y A ). After time t, the position of node A relative to B is:

[0058]

[0059] where θ AB is the velocity angle of node A relative to node B, v AB is the speed of node A relative to B; when the distance between nodes A and B is the maximum communication distance R of the drone, the link between A and B is disconnected, that is:

[0060]

[0061] Substituting formula (5) into (6) we can obtain:

[0062]

[0063] v AB =v A -v B =(|v A |cosθ A -|v B |cosθ B )i+(|v A |sinθ A -|v B |sinθ B )j (22)

[0064] Among them, θ A is the moving direction of node A, θ B is the moving direction of node B, v A and vB are the velocities of nodes A and B respectively. Since t cannot be negative, the value of t is:

[0065]

[0066] Therefore, the stability of the link LS is defined as AB for:

[0067]

[0068] In the UAV cluster, there are active nodes s and relay nodes N s (1) Relay node N s (2) and the destination node d, assuming that the source node s is stationary, the movement speeds of the other three nodes are the velocity vectors relative to node s, and the angle ω between the movement direction of the candidate relay node i and the movement direction of the destination node d i The expression is:

[0069]

[0070] ω i The value range is [0, π]. When the node moves to the opposite direction of the target node, that is, when ω i When it is greater than 0.5π, it will be deleted from the relay node candidate set;

[0071] Each node i maintains a set of neighbors Np(i), where Np(i) represents the nodes that are closer to the destination node than node i after screening. When a node forwards a data packet to its neighbors, it inserts its own position into the data packet. When a node receives a data packet, it compares the position in the data packet with its own and updates its better neighbor set. The mobility adaptability metric PM(i) of node i is:

[0072]

[0073] Among them, d i，d and d s，d are the Euclidean distances from nodes i and s to the target node d, respectively. N(i) is the set of candidate nodes maintained by node i before receiving the routing request broadcast by node s.

[0074] Assume that E0 and E r The initial energy value and the remaining energy value of each UAV node are,in an ad hoc network, a mobile node can be selected as a forwarder only under the following conditions:,(1) It has enough energy E,within the effective time of the link with the sender. min1 Send data packets; (2) After the link validity time, the node still has enough energy E min2 Move to the designated location, the minimum threshold E for drone data forwardingmin It consists of two parts: E min =E min1 +E min2 The energy index value calculation method is shown in formula (13). If the energy value is 0, the node should be removed from the candidate relay node set:

[0075]

[0076] In summary, based on the link stability prediction value, node mobility adaptability and energy measurement value, a priority measurement equation is proposed. The priority measurement equation of node i in the candidate relay set of the drone source node s is:

[0077] Val(N s (i)) = αLS(N s (i))+βPM(N s (i))+γEne(N s (i)) (28)

[0078] Among them, α, β, and γ are the weight coefficients of the three metrics, reflecting their importance in the priority metric calculation. Candidate relay nodes calculate their own priority metrics according to the above formula to sort candidate nodes and filter candidate node sets;

[0079] Step 3: When the source node generates a data packet or a data packet arrives at a drone node, the node will extract the destination information from the data packet to check whether it is the destination node. If so, the data packet forwarding process ends after confirmation. If not, it enters the candidate relay set selection phase. Based on the periodically maintained neighbor node information, the node broadcasts a routing request to the nodes within the communication range. After receiving the routing request, the nodes in the candidate relay set will recalculate the node's link stability, mobility adaptability, and residual energy value based on their own status information and the source node status information, and eliminate candidate nodes that do not meet the minimum requirements. In the node mobility adaptability, the node moves in the opposite direction of the destination node, or its own residual energy is lower than the threshold value. The priority metric equation is used to calculate the Val value. Finally, a route reply is made. After receiving a route reply from a neighboring node, the forwarding node updates its candidate relay set. If the number of nodes in the candidate relay set is zero, the node enters the "store-carry-forward" mode and attempts a new route discovery in the next cycle. If the number of nodes in the candidate relay set is not zero after filtering, the candidate node sets a corresponding timer size based on Val. The larger the value of Val, the higher the priority of the node, the greater its contribution to routing, and the smaller the value of its timer setting. When the candidate relay node's own timer expires, it broadcasts a route reply. Other candidate relay nodes automatically discard the data packet after receiving this broadcast. After that, the node becomes a forwarding node again and begins a new round of route discovery.

[0080] The contents not described in detail in the present application belong to the prior art known to professional and technical personnel in this field.

Claims

1. An adaptive UAV cooperative opportunistic routing method based on grey perception, specifically comprising the following steps: Step 1: Comprehensively process the historical records of the UAV's motion speed and direction, and establish a gray prediction model for the UAV's motion state. Using the metabolic model GM (1, 1), the sliding window method is used to keep the modeling length unchanged, continuously add new values, and remove old values to predict the next motion speed and direction of the UAV. The specific method is as follows: Let the original data series be x (0) ={x (0) (1), x (0) (2),…,x (0) (n)}, where x (0) (k) is the data at time k, let x (1) (1) = x (0) (1), then x (0) The cumulative generated sequence of (1) is x (1) ={x (1) (1), x (1) (2),…,x (1) (n)}, where For the metabolic GM (1, 1) prediction model, its grey differential equation is: x (0) (k)+az (1) (k)=b (1) Where: a and b are constants, x (1) The expression of (t) is: Discretize it and get: Finally, the predicted value is: The true value x at time l (0) (l) and predicted value The difference between the two becomes the residual at time l, denoted as e (0) (l); Step 2: Use the grey prediction model to determine the speed and direction of the drone at the next moment, and calculate the link stability and mobility adaptability between drone nodes. The specific method is as follows: The link stability and node mobility adaptability are defined as follows: The initial distance between node A and node B is When calculating, the two moving nodes are considered as one node being stationary and the other node being in relative motion. The position of node A relative to node B is (x AB ,y AB )=(x B -x A ,y B -y A ), after time t, the position of node A relative to B is: where θ AB is the velocity angle of node A relative to node B, v AB is the speed of node A relative to B; when the distance between nodes A and B is the maximum communication distance R of the drone, the link between A and B is disconnected, that is: Substituting formula (5) into (6) we can obtain: v AB =v A -v B =(|v A |cosθ A -|v B |cosθ B )i+(|v A |sinθ A -|v B |sinθ B )j (8) Among them, θ A is the moving direction of node A, θ B is the moving direction of node B, v A and v B are the velocities of nodes A and B respectively. Since t cannot be negative, the value of t is: Therefore, the stability of the link LS is defined as AB for: In the UAV cluster, there are active nodes s and relay nodes N s (1) Relay node N s (2) and the destination node d, the source node s is stationary, the movement speeds of the other three nodes are the velocity vectors relative to node s, and the angle ω between the movement direction of the candidate relay node i and the movement direction of the destination node d i The expression is: ω i The value range is [0, π]. When the node moves to the opposite direction of the target node, that is, when ω i When it is greater than 0.5π, it will be deleted from the relay node candidate set; Each node i maintains a set of neighbors Np(i), where Np(i) represents the nodes that are closer to the destination node than node i after screening. When a node forwards a data packet to its neighbors, it inserts its own position into the data packet. When a node receives a data packet, it compares the position in the data packet with its own and updates its better neighbor set. The mobility adaptability metric PM(i) of node i is: Among them, d i，d and d s，d are the Euclidean distances from nodes i and s to the target node d, respectively. N(i) is the set of candidate nodes maintained by node i before receiving the routing request broadcast by node s. E0 and E r The initial energy value and the remaining energy value of each UAV node are,in an ad hoc network, a mobile node can be selected as a forwarder only under the following conditions:,(1) It has enough energy E,within the effective time of the link with the sender. min1 Send data packets; (2) After the link validity time, the node still has enough energy E min2 Move to the designated location, the minimum threshold E for drone data forwarding min It consists of two parts: E min =E min1 +E min2 The energy index value calculation method is shown in formula (13). If the energy value is 0, the node should be removed from the candidate relay node set: In summary, based on the link stability prediction value, node mobility adaptability and energy measurement value, a priority measurement equation is proposed. The priority measurement equation of node i in the candidate relay set of the drone source node s is: Val(N s (i))=αLS(N s (i))+βPM(N s (i))+γEne(N s (i)) (14) Among them, α, β, and γ are the weight coefficients of the three metrics, reflecting their importance in the priority metric calculation. Candidate relay nodes calculate their own priority metrics according to the above formula to sort candidate nodes and filter candidate node sets; Step 3: When the source node generates a data packet or the data packet arrives at a certain drone node, the node will extract the destination information from the data packet to check whether it is the destination node. If so, the data packet forwarding process ends after confirmation. If not, it enters the candidate relay set selection phase. Based on the periodically maintained neighbor node information, the node broadcasts a routing request to the nodes within the communication range. After receiving the routing request, the nodes in the candidate relay set will recalculate the node's link stability, mobility adaptability and residual energy value based on their own status information and the source node status information, and eliminate candidate nodes that do not meet the minimum requirements. In the node mobility adaptability, the node moves in the opposite direction of the destination node, or its own residual energy is lower than the threshold value. The Val value is calculated using the priority metric equation, and finally a route reply is made. After receiving the route reply from the neighbor node, the forwarding node updates its candidate relay set. If the number of nodes in the candidate relay set is zero, the node will enter the "store-carry-forward" mode and try a new route discovery in the next cycle. If the number of nodes in the candidate relay set is not zero after filtering, the candidate node will set the corresponding timer size according to Val. The larger the Val value, the higher the priority of this node, the greater the contribution to the route, and the smaller the value of its timer setting. When the candidate relay node's own timer times out, it will broadcast a route reply. Other candidate relay nodes will automatically discard the data packet after receiving this broadcast. After that, the node will become a forwarding node again and start a new round of route discovery.

Images