A heterogeneous ad hoc network OLSR protocol implementation method based on topology control and MPR selection optimization

Abstract

The application discloses a kind of based on topological control and MPR selection optimization's heterogeneous self-organizing network OLSR protocol implementation method, this method is first by calculating link quality factor LF and is clustered, to form stable cluster structure.Subsequently, using AOW clustering algorithm, according to the node's various factors Selection suitable node assumes cluster head, reduces the frequency and related overhead of cluster head replacement.In addition, to ensure the connectivity between each cluster and form a complete network topology, based on the design concept of cross-layer routing, introduce the multi-frequency clustering inter-cluster connection algorithm based on TDMA protocol.Through continuous maintenance topological routing.Introduce a link stability factor based on laser transceiver device scanning angle, to optimize MPR selection mechanism.Has been realized in EXata network simulation environment, and is verified by experiment.The optimized protocol of the application is more excellent in the environment of high transmission rate requirement, high mobility and atmospheric attenuation, effectively improves the performance and reliability of network.

Description

Technical Field

[0001] This invention relates to a method for implementing the OLSR protocol in heterogeneous ad hoc networks based on topology control and MPR selection optimization, belonging to the field of computer communication technology. Background Technology

[0002] In mobile ad hoc network scenarios with frequent network topology changes and high real-time requirements, immediate route availability is crucial. Unlike passive routing protocols such as AODV, the OLSR protocol does not require searching for or establishing a route path only when data transmission needs arise. Therefore, the OLSR protocol is widely used for real-time applications requiring fast and reliable responses, such as drone ad hoc networks or emergency response networks. However, the OLSR protocol only uses hop count as the criterion for route selection, making the selection metric too simplistic and difficult to adapt to the highly dynamic and high-load mobile ad hoc network environment.

[0003] Currently, RF communication is susceptible to interference from environmental magnetic fields such as high-voltage lines and base stations in modern battlefields and commercial applications. Its broadcast nature may also lead to co-channel interference and security issues, limiting stealth and communication efficiency, especially in military communications that require high immediacy, high speed, large bandwidth and high security.

[0004] To address these challenges, RF / FSO hybrid technology offers an effective solution. Within this framework, network nodes can flexibly choose between RF or FSO communication based on actual network needs, enhancing network adaptability and reliability. However, when using existing OLSR routing protocols, node movement and changes in atmospheric conditions in FSO networks can cause signal strength fluctuations, affecting reception quality. Furthermore, OLSR requires frequent broadcast signaling to update network topology and maintain routing tables. This is inefficient in point-to-point FSO networks, requiring each node to install a large number of laser transceivers. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings and deficiencies of existing technologies by proposing a method for implementing the OLSR protocol in heterogeneous ad hoc networks based on topology control and MPR selection optimization. This method employs a dynamic clustering topology control approach, based on the AOW algorithm, to dynamically cluster the network according to node movement and atmospheric attenuation, adapting to rapidly changing network environments. Furthermore, considering the unique characteristic of using laser as a carrier in FSO communication, and to reduce the broadcast requirements of the entire network, this invention introduces a link stability factor based on the scanning angle of the laser transceiver to optimize the MPR selection mechanism. It has been implemented in the EXata network simulation environment and verified experimentally. Experimental results show that the optimized protocol performs better in environments with high transmission rate requirements, high mobility, and atmospheric attenuation, effectively improving network performance and reliability.

[0006] The technical solution adopted by this invention to solve its technical problem is: a method for implementing the OLSR protocol in heterogeneous ad hoc networks based on topology control and MPR selection optimization, which includes the following:

[0007] First, clustering is performed by calculating the Link Quality Factor (LF) to form a stable cluster structure. Then, the AOW clustering algorithm is used to select suitable nodes as cluster heads based on various factors, aiming to reduce the frequency of cluster head replacements and related overhead. Furthermore, to ensure connectivity between clusters and form a complete network topology, this invention introduces a multi-frequency clustering inter-cluster connection algorithm based on the TDMA protocol, based on the cross-layer routing design concept. By continuously maintaining the topology routing, this invention can adjust its strategy in a timely manner according to dynamic changes in the network environment, effectively implementing topology control in heterogeneous networks.

[0008] (1) Clustering structure formation

[0009] During the formation of the cluster structure, the motion characteristics of the nodes are considered, and the formula is used. The link quality factor (LF) calculated in the protocol is used as the basis for cluster allocation. The protocol will set a global cluster ID to identify which cluster a node belongs to. The specific algorithm steps and flowchart are as follows: Figure 2 As shown.

[0010] (a) Initialization phase: The initial cluster ID of the node is set to the node ID, and the link quality factor threshold LF is set according to the specific network conditions. th The set LF th A higher value indicates a more stringent requirement for link quality, which leads to a reduction in the number of nodes in the cluster structure.

[0011] (b) The current receiving node u receives a one-hop neighbor node N iWhen creating a HELLO group, it will retrieve the neighbor node ID, cluster ID, and neighbor location information contained in that group.

[0012] (c) Node u will compare its own cluster ID with that of its neighbor node N. i The cluster ID. If the two are equal, it indicates that the two nodes are already in the same cluster structure, so the subsequent process is skipped.

[0013] (d) According to the formula Calculate the link quality factor between the two nodes and compare it with the set quality threshold LF. th The comparison shows that if the link quality between the two points is less than a threshold, the link quality is considered low, and the two nodes are unlikely to form a stable cluster structure. If the link quality is greater than the threshold, the two nodes are considered to meet the clustering conditions. Here, R... d This represents the maximum communication distance of a node. In the RWP model, if the node's maximum speed v max Lower and dwell time t p If the distance is longer, the relative travel distance per unit time will decrease, and the link quality between the two nodes will improve accordingly.

[0014] (e) Make a unified modification to the cluster IDs of the two nodes and their fellow cluster members, and select the cluster ID value with the smaller cluster ID of the two nodes as the unified cluster ID. If the node also receives HELLO packets from other neighbors, it will return to stage (b) and repeat the entire process until the entire network forms a stable cluster structure.

[0015] (2) Cluster head selection

[0016] After the clustering algorithm is completed, multiple independent cluster structures will be formed, and each cluster needs a cluster head to effectively manage its internal nodes. Selecting a suitable cluster head is crucial. Therefore, the AOW algorithm, which is adapted to the characteristics of RF / FSO networks, is used to select the cluster head.

[0017] When selecting a cluster head, the following three evaluation metrics should be given priority, using node u within the cluster as an example:

[0018] (a) Since each node has a limited number of transceiver devices, its node degree becomes a critical factor. A node can only be considered as a candidate for cluster head if it has sufficient remaining node degree.

[0019] The number N of neighboring nodes of node u is counted and defined as the degree d of node u. u The remaining node degree Du of node u is the difference between its access degree and physical degree K, calculated using the following formula:

[0020] D u =|Kd u | (4.4)

[0021] Physical degree K refers to the maximum number of concurrent links a node can support. For an FSO node, it refers to the number of independent point-to-point optical links that the node can maintain simultaneously. This value is determined by the number of laser transceiver units equipped on the node. For example, a node equipped with 5 transmitting units and 5 receiving units can be considered to have a physical degree K of 5.

[0022] (b) To ensure the stability of the cluster head, nodes with lower mobility are preferentially selected as cluster heads. The relative movement speed of node u with respect to node v is V(u,v):

[0023] V(u,v)=|V u -V v |(4.5)

[0024] Where V u and V v Let M be the velocity of node u and node v, respectively. Then the average relative mobility M of node u within the cluster is... u for:

[0025]

[0026] (c) In the design of RF / FSO networks, the transmit power of a node should be determined based on the actual distance it needs to cover and atmospheric attenuation. The ideal transmit power setting should ensure that the node can maintain effective communication with its one-hop neighbors, and should also consider the effects of atmospheric attenuation. Therefore, estimating the ideal transmit power by calculating the sum of path losses from node u to all its one-hop neighbors is an effective method. Therefore, the ideal transmit power P of node u... u for:

[0027]

[0028] Where, d u,v Let be the distance between nodes u and v, N1 represent the set of one-hop neighbor nodes of node u, and ε be the atmospheric attenuation coefficient.

[0029] Based on the previously calculated node mobility, remaining node degree, and ideal transmission power, the combined weight W of node u is... u for:

[0030]

[0031] Where ω d ω m ω p The weighting factors for mobility, remaining node degree, and transmission power are respectively, satisfying ω d +ω m +ω p=1. Finally, based on the overall weight W of the nodes, the node with the smallest weight will be selected as the cluster head. If there are nodes with equal weights, the node with the smaller ID will be selected as the cluster head.

[0032] (3) Inter-cluster connections

[0033] To reduce channel collisions and improve channel utilization in relatively dense mobile ad hoc networks, an inter-cluster connection mechanism based on cross-layer routing design is proposed. This mechanism employs multi-frequency clustering communication at the MAC layer and utilizes the TDMA protocol to allocate dedicated communication frequency bands to different clusters. For each cluster, a communication frequency is configured for the node based on its cluster ID. The gateway node uses an inter-cluster time-division frequency hopping access method to solve the communication access problem between different operating frequency bands. A schematic diagram of the connection between two adjacent clusters is shown below. Figure 3 As shown.

[0034] Specifically, before the protocol runs, several available communication frequencies are configured according to the scenario requirements. During protocol operation, the communication frequency band of the cluster to which the current node belongs is updated in real time using the cluster ID. This process can be periodically triggered by configuring a timer. For gateway nodes at the boundary of two adjacent clusters, in order to ensure stable data exchange between clusters, a certain frequency hopping capability is required. The following scheme is adopted: during odd-numbered frame periods, the node switches to the frequency band of the neighboring cluster with the smaller ID; during even-numbered frame periods, it switches to the frequency band of the neighboring cluster with the larger ID. The gateway node alternately switches its operating frequency band between two neighboring clusters, ensuring the continuity and efficiency of data transmission.

[0035] 1.1.1 MPR Selection Optimization Based on Link Stability

[0036] In RF / FSO communication networks, various factors during laser transmission and reception significantly affect link stability. Quantifying these factors is crucial for accurately evaluating and optimizing network performance. A link stability calculation method based on laser transmission and reception parameters and inter-node distance is introduced. The calculation formula for LS is as follows:

[0037]

[0038] Where α is the scanning angle of the laser transceiver. The impact on link stability is relative; systems with larger scanning angles are more tolerant of node movement and alignment errors, thus enhancing link stability. Distance d represents the actual distance between two nodes, while D... max This refers to the maximum effective communication distance between two transceivers. ρ describes the ratio of the maximum effective communication distance to the optimal communication distance, and this ratio can be adjusted according to the performance of different transceivers.

[0039] Link stability (LS) ranges from [0,1] and reflects the stability of each node in the network in real time. A higher LS value indicates greater node stability, thus increasing the probability of the node being selected as an MPR (Most Probable Node). This metric helps optimize routing protocols, ensuring more stable and efficient network communication.

[0040] Figure 4 This describes the optimized MPR selection process. Similar to the standard MPR node set selection algorithm, it assumes the MPR set selected for the center node is set M; the one-hop neighbor set of each node is N1; and the two-hop neighbor set of each node is N2, excluding the center node and nodes in N1. Specifically, it includes:

[0041] Step 1) First, add the neighbor node with the willingness value ALWAYS to set M;

[0042] Step 2) Obtain the geographical location of node N1 from the neighbor table, calculate the distance between nodes, and apply the formula... Calculate the link stability LS and depth D(y);

[0043] Step 3) If a two-hop neighbor can only be connected to the point of the present invention through a one-hop neighbor, add this one-hop neighbor to set M and remove the two-hop neighbors it covers.

[0044] Step 4) If any node in N1 is already the MPR of other nodes, select the node with the largest LS as the MPR of the central node; otherwise, proceed to the next step.

[0045] Step 5) Calculate the reachability of node N1 and execute the following steps sequentially:

[0046] (a) Add the node with the highest willingness to the set M;

[0047] (b) Among multiple candidates, select the node with the largest LS to join M;

[0048] (c) Among multiple candidates, select the node with the highest reachability to join M;

[0049] Step 6: Check if set M covers all two-hop neighbors N2. If it does, M is the final MPR set; otherwise, return to step 4.

[0050] Beneficial effects:

[0051] 1. Experimental results show that the optimized protocol of this invention performs better in environments with high transmission rate requirements, high mobility, and atmospheric attenuation, effectively improving network performance and reliability.

[0052] 2. The optimized MPR selection algorithm of this invention maximizes the utility of each MPR node by reusing nodes already used as MPRs, reducing the total number of nodes that need to be broadcast in the network. Simultaneously, selecting nodes with higher link stability as MPRs reduces the risk of link failures and enhances network stability.

[0053] 3. This invention first analyzes the limitations of traditional RF communication and the challenges encountered by existing OLSR protocols in RF / FSO networks. Secondly, it introduces the channel model of FSO networks, particularly the impact of atmospheric attenuation on signal transmission, and presents the node model and adaptive on-demand weighted clustering algorithm for RF / FSO networks. This invention describes the protocol design concept, proposing a clustering topology control method based on link quality and an MPR optimization strategy based on link stability. Finally, experimental results from Exata simulation software show that the optimized protocol exhibits superior performance under high mobility and atmospheric attenuation conditions, especially in environments with high transmission rate requirements. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the clustering topology control process of the present invention.

[0055] Figure 2 This is a flowchart of the clustering algorithm based on link quality in this invention.

[0056] Figure 3 This is a schematic diagram illustrating the connection between two adjacent clusters in this invention.

[0057] Figure 4 This is a flowchart of the MPR selection process based on link stability according to the present invention.

[0058] Figure 5 This is a schematic diagram of the overall process of the TC-M-OLSR protocol of the present invention.

[0059] Figure 6 This is a schematic diagram of a simulation scenario for an air-to-ground information self-organizing network according to the present invention.

[0060] Figure 7 This is a schematic diagram illustrating how time delay varies with node speed under different weather conditions according to the present invention.

[0061] Figure 8 This is a schematic diagram illustrating how the group delivery rate changes with node speed under different weather conditions according to the present invention.

[0062] Figure 9 This is a schematic diagram illustrating how the broadcast control overhead of the present invention changes with the node's moving speed.

[0063] Figure 10 This is a schematic diagram illustrating how the broadcast control overhead of the present invention varies with the number of nodes. Detailed Implementation

[0064] The invention will now be described in further detail with reference to the accompanying drawings.

[0065] The RF / FSO communication network node model specifically includes:

[0066] Although RF and FSO communication networks differ fundamentally in their transmission media and methods, it is necessary to combine the advantages of both to meet a wider range of application needs when designing network nodes. Therefore, nodes in an RF / FSO communication network should be designed to meet the following conditions:

[0067] (1) RF / FSO nodes allow coverage of multiple directions during communication.

[0068] Similar to omnidirectional antennas used in RF communication, RF / FSO nodes require multi-directional communication capabilities. This design feature is crucial because it enables routing protocols to effectively implement neighbor discovery via broadcast control messages. However, laser transmission is inherently highly directional. Currently, to achieve an omnidirectional antenna-like effect during communication, the node's transceiver can employ a three-dimensional symmetrical antenna structure (such as a dodecahedron or icosahedron), with each face equipped with multiple transmitting and receiving units. This design optimizes the distribution of transmission and reception, enabling signal transmission in multiple directions and achieving long-distance laser communication while maintaining a small beam divergence angle.

[0069] (2) RF / FSO nodes are equipped with precise signal alignment technology

[0070] APT (Alignment Point) technology is crucial for FSO (Fiber Optic Array) systems because FSO communication heavily relies on precise beam alignment. The core function of APT is to ensure stable and accurate laser paths between communicating devices, thereby enabling efficient data transmission. Currently, precise APT technology primarily relies on high-speed sensors, beam stabilization techniques, and efficient PID control algorithms.

[0071] (3) RF / FSO nodes support multiple network access protocols

[0072] In terms of working mechanism and function, the multidirectional communication capability of RF / FSO network nodes is similar to diversity reception and MIMO technologies in RF communication, and also supports different network access protocols and routing protocols. It possesses the ability to forward information over multiple hops, thereby supporting extensive data distribution and robust link control within the network. Simultaneously, the node supports multiple-transmitter and multiple-receiver information processing capabilities, enabling efficient information transmission within a specified communication range.

[0073] The FSO communication network channel loss model specifically includes:

[0074] The FSO communication channel loss model mainly includes two aspects: atmospheric attenuation effect and atmospheric turbulence effect. Since the atmospheric turbulence model requires not only a large number of real-time environmental parameters but also frequent computational updates, this significantly increases the computational load and routing decision complexity. Therefore, to simplify the channel loss model and improve computational efficiency, this invention chooses to consider only the atmospheric attenuation effect. The atmospheric attenuation effect is mainly caused by weather factors (such as haze, precipitation, etc.), its model is relatively simple, its computational cost is low, and its impact on channel loss can be assessed through relatively direct parameter measurements.

[0075] Considering the impact of atmospheric attenuation on network channels, in an FSO communication network, the path loss between nodes u and v is defined as:

[0076] PL FSO =εd u,v (4.1)

[0077] Where d u,v Let u be the distance between nodes u and v, and ε be the atmospheric attenuation coefficient, which is calculated using formula (2.1). This invention mainly studies and simulates network performance under two weather conditions: sunny and foggy. The average value of the corresponding weather attenuation coefficients is taken: ε1 = 0.137 for sunny weather and ε2 = 5.865 for foggy weather.

[0078] The FSO communication network control overhead model of this invention specifically includes:

[0079] In wireless communication networks, the control overhead of routing protocols is a significant factor affecting network performance. Due to differences in transmission media and communication characteristics, the control overhead varies significantly between RF and FSO networks. In FSO networks, due to their high directionality and environmental sensitivity, the control overhead of the OLSR protocol can increase significantly, thus impacting network efficiency. Therefore, to better understand the control overhead of using the OLSR protocol in FSO networks, it is necessary to establish a quantitative evaluation model to compare the differences in control signaling overhead between FSO and RF networks.

[0080] For each node, broadcast control messages need to be repeatedly transmitted to ensure that all neighboring nodes receive the message. This introduces the success rate P of single-hop packet transmission. For FSO networks, due to high directionality and environmental influences, the success rate P of a single transmission is... FSO The signal is relatively low. However, signal attenuation can occur during transmission due to atmospheric conditions. Taking all these factors into account, the control overhead C of the FSO network is relatively low. FSO It can be represented as:

[0081]

[0082] Where N is the total number of nodes in the network, F is the transmission frequency of control signaling, and its value is the reciprocal of the transmission interval. S is the size of a single control signaling message, and P... FSO PL represents the success rate of a single-hop signaling transmission. FSO It is used to characterize the attenuation of the channel in the atmosphere, and the calculation method is shown in formula (4.1).

[0083] This invention significantly improves network transmission rate and anti-interference capability when applying RF / FSO hybrid technology to mobile ad hoc networks. However, when considering the node model, signal loss model, and dynamic topology model of FSO networks, directly applying the existing OLSR protocol to RF / FSO networks mainly faces the following two problems:

[0084] (1) The dynamics of nodes and the influence of atmospheric factors on network stability

[0085] In RF / FSO heterogeneous networks, FSO communication is characterized by high directionality and environmental sensitivity, making signal transmission highly susceptible to disruptions in dynamic environments. Rapid node movement can lead to transmission link breaks; atmospheric factors such as fog and haze can cause laser signal blockage or attenuation along the transmission path. Referring to formula (4.1), atmospheric channel loss is significantly higher in foggy weather than in clear weather. Existing OLSR protocols have not adequately considered the impact of dynamics and the environment on network performance during their design.

[0086] (2) The broadcast nature of the OLSR protocol introduces a large amount of control overhead in FSO networks.

[0087] The OLSR protocol relies on periodic control message broadcasts to maintain the network's routing table and topology information. This broadcast mechanism is quite effective in traditional RF networks, but it is inefficient in FSO networks that use point-to-point communication. Referring to formula (4.2), due to the need for precise alignment in FSO networks, the transmission success rate P... FSO Lower. Furthermore, the signal attenuation PL of the FSO network is [not specified]. FSO It also fluctuates significantly due to changes in the atmospheric environment. Therefore, when using the OLSR protocol, FSO networks introduce greater control signaling overhead compared to RF networks. Although modifying the broadcast mechanism of the OLSR protocol is difficult, it is feasible to reduce the broadcast control overhead in the network by improving the protocol.

[0088] The heterogeneous ad hoc network OLSR protocol based on topology control and MPR selection optimization of this invention specifically includes:

[0089] The TC-M-OLSR protocol of this invention improves the OLSR protocol, mainly in two aspects:

[0090] (1) A clustering topology control method based on the characteristics of RF / FSO heterogeneous networks and considering node mobility is introduced. This topology control method is based on an adaptive on-demand weighted clustering algorithm, which effectively adapts to the dynamic changes of the network and reduces network overhead.

[0091] (2) To address link stability in RF / FSO heterogeneous networks, the MPR selection strategy was optimized within the cluster. Link stability takes into account both the distance between nodes and the scanning angle of the laser transceiver, thereby further reducing broadcast control overhead.

[0092] 1.1.2 Adaptive On-Demand Weighted Clustering Algorithm

[0093] In mobile ad hoc networks, clustering strategies are commonly used to construct hierarchical communication topologies. Stable clustering algorithms can significantly reduce network computational and communication overhead. The Adaptive On-demand Weighting Clustering (AOW) algorithm is renowned for its high flexibility. Unlike single-factor-based clustering algorithms, AOW selects cluster heads by comprehensively considering multiple factors affecting network performance through weighting factors. The weighting coefficients can be adaptively adjusted according to network needs.

[0094] When selecting a cluster head node, the AOW algorithm comprehensively considers four core factors: node mobility M, node degree D, transmit power P, and remaining energy E. The final weight W of the node is calculated using a weighting formula, which conforms to formula (4.3). In practical applications, the node selects the node with the smallest total weight from its neighbors to serve as the cluster head. If there are nodes with the same weight, the node with the smaller ID is selected as the cluster head.

[0095] W = ω m M+ω d D+ω p P+ω e E(4.3)

[0096] Introducing the AOW clustering model into RF / FSO heterogeneous ad hoc networks brings significant advantages. First, by optimizing cluster head selection, this model effectively reduces network control overhead, thereby improving system efficiency. Second, the AOW model enhances the network's adaptability, enabling it to more flexibly respond to changing environmental conditions. This approach not only optimizes resource allocation and utilization but also improves network stability and reliability under different operating environments.

[64] .

[0097] 1.1.3 Clustering Topology Control Method

[0098] In RF / FSO networks, node movement can lead to instability in link quality. To address this issue, clustering is first performed by calculating the link quality factor LF to form a stable cluster structure. Subsequently, the AOW clustering algorithm is used to select a suitable node as the cluster head based on various factors of the node, aiming to reduce the frequency of cluster head replacement and related overhead. In addition, to ensure connectivity between clusters and form a complete network topology, this invention introduces a multi-frequency clustering inter-cluster connection algorithm based on the TDMA protocol, based on the design concept of cross-layer routing. By continuously maintaining the topology routing, this invention can adjust the strategy in a timely manner according to the dynamic changes in the network environment, effectively implementing topology control of heterogeneous networks. The framework diagram of the entire process is shown below: (1) Cluster structure formation

[0099] During the formation of the cluster structure, the motion characteristics of the nodes are considered, and the formula is used. The link quality factor LF calculated in the algorithm is used as the basis for cluster allocation, where R... d This represents the maximum communication distance of a node. In the RWP model, if the node's maximum speed v max Lower and dwell time t p If the distance is longer, the relative travel distance per unit time will decrease, and the link quality between the two nodes will improve accordingly. The protocol will set a global cluster ID to identify which cluster a node belongs to. The specific algorithm steps and flowchart are as follows: Figure 2 As shown:

[0100] (a) Initialization phase: The initial cluster ID of the node is set to the node ID, and the link quality factor threshold LF is set according to the specific network conditions. th The set LF th A higher value indicates a more stringent requirement for link quality, which leads to a reduction in the number of nodes in the cluster structure.

[0101] (b) The current receiving node u receives a one-hop neighbor node N i When creating a HELLO group, it will retrieve the neighbor node ID, cluster ID, and neighbor location information contained in that group.

[0102] (c) Node u will compare its own cluster ID with that of its neighbor node N. i The cluster ID. If the two are equal, it indicates that the two nodes are already in the same cluster structure, so the subsequent process is skipped.

[0103] (d) Calculate the link quality factor between the two nodes according to formula (3.3), and compare it with the set quality threshold LF. th In comparison, if the link quality is less than the threshold, the link quality between the two points is considered to be low, and the two nodes are unlikely to form a stable cluster structure. If the link quality is greater than the threshold, the two nodes are considered to meet the conditions for cluster building.

[0104] (e) Make a unified modification to the cluster IDs of the two nodes and their fellow cluster members, and select the cluster ID value with the smaller cluster ID of the two nodes as the unified cluster ID. If the node also receives HELLO packets from other neighbors, it will return to stage (b) and repeat the entire process until the entire network forms a stable cluster structure.

[0105] (2) Cluster head selection

[0106] After the clustering algorithm is completed, multiple independent cluster structures will be formed, and each cluster needs a cluster head to effectively manage its internal nodes. Selecting a suitable cluster head is crucial. Therefore, the AOW algorithm, which is adapted to the characteristics of RF / FSO networks, is used to select the cluster head.

[0107] When selecting a cluster head, the following three evaluation metrics should be given priority, using node u within the cluster as an example:

[0108] (a) Since each node has a limited number of transceiver devices, its node degree becomes a critical factor. A node can only be considered as a candidate for cluster head if it has sufficient remaining node degree.

[0109] The number N of neighboring nodes of node u is counted and defined as the degree d of node u. u The remaining node degree Du of node u is the difference between its access degree and physical degree K, calculated using the following formula:

[0110] D u =|Kd u | (4.4)

[0111] Physical degree K refers to the maximum number of concurrent links a node can support. For an FSO node, it refers to the number of independent point-to-point optical links that the node can maintain simultaneously. This value is determined by the number of laser transceiver units equipped on the node. For example, a node equipped with 5 transmitting units and 5 receiving units can be considered to have a physical degree K of 5.

[0112] (b) To ensure the stability of the cluster head, nodes with lower mobility are preferentially selected as cluster heads. The relative movement speed of node u with respect to node v is V(u,v):

[0113] V(u,v)=|V u -V v |(4.5)

[0114] Where V u and V v Let M be the velocity of node u and node v, respectively. Then the average relative mobility M of node u within the cluster is... u for:

[0115]

[0116] (c) In the design of RF / FSO networks, the transmit power of a node should be determined based on the actual distance it needs to cover and atmospheric attenuation. The ideal transmit power setting should ensure that the node can maintain effective communication with its one-hop neighbors, and should also consider the effects of atmospheric attenuation. Therefore, estimating the ideal transmit power by calculating the sum of path losses from node u to all its one-hop neighbors is an effective method. Therefore, the ideal transmit power P of node u... u for:

[0117]

[0118] Where, d u,v Let be the distance between nodes u and v, N1 represent the set of one-hop neighbor nodes of node u, and ε be the atmospheric attenuation coefficient.

[0119] Based on the previously calculated node mobility, remaining node degree, and ideal transmission power, the combined weight W of node u is... u for:

[0120]

[0121] Where ω d ω m ω p The weighting factors for mobility, remaining node degree, and transmission power are respectively, satisfying ω d +ω m +ω p =1. Finally, based on the overall weight W of the nodes, the node with the smallest weight will be selected as the cluster head. If there are nodes with equal weights, the node with the smaller ID will be selected as the cluster head.

[0122] (3) Inter-cluster connections

[0123] To reduce channel collisions and improve channel utilization in relatively dense mobile ad hoc networks, this invention proposes an inter-cluster connection mechanism based on a cross-layer routing design concept. This mechanism employs multi-frequency clustering communication at the MAC layer and utilizes the TDMA protocol to allocate dedicated communication frequency bands to different clusters. For each cluster, a communication frequency is configured for the node based on its cluster ID. The gateway node uses an inter-cluster time-division frequency hopping access method to solve the communication access problem between different operating frequency bands. A schematic diagram of the connection between two adjacent clusters is shown below. Figure 3 As shown.

[0124] Specifically, before the protocol runs, several available communication frequencies are configured according to the scenario requirements. During protocol operation, the communication frequency band of the cluster to which the current node belongs is updated in real time using the cluster ID. This process can be periodically triggered by configuring a timer. For gateway nodes at the boundary of two adjacent clusters, in order to ensure stable data exchange between clusters, a certain frequency hopping capability is required. The following scheme is adopted: during odd-numbered frame periods, the node switches to the frequency band of the neighboring cluster with the smaller ID; during even-numbered frame periods, it switches to the frequency band of the neighboring cluster with the larger ID. The gateway node alternately switches its operating frequency band between two neighboring clusters, ensuring the continuity and efficiency of data transmission.

[0125] 1.1.4 MPR Selection Optimization Based on Link Stability

[0126] In RF / FSO communication networks, various factors during laser transmission and reception significantly affect link stability. Quantifying these factors is crucial for accurately evaluating and optimizing network performance. A link stability calculation method based on laser transmission and reception parameters and inter-node distance is introduced. The calculation formula for LS is as follows:

[0127]

[0128] Where α is the scanning angle θ of the laser transceiver. scan The impact on link stability is relative; systems with larger scanning angles are more tolerant of node movement and alignment errors, thus enhancing link stability. Distance d represents the actual distance between two nodes, while D... max This refers to the maximum effective communication distance between two transceivers. ρ describes the ratio of the maximum effective communication distance to the optimal communication distance, and this ratio can be adjusted according to the performance of different transceivers.

[0129] Link stability (LS) ranges from [0,1] and reflects the stability of each node in the network in real time. A higher LS value indicates greater node stability, thus increasing the probability of the node being selected as an MPR (Most Probable Node). This metric helps optimize routing protocols, ensuring more stable and efficient network communication.

[0130] Figure 4 This is the optimized MPR selection process. Similar to the standard MPR node set selection algorithm, it is assumed that the MPR set selected by the center node is set M; the one-hop neighbor set of the node is N1; and the two-hop neighbor set of the node is N2, which does not include the center node and the nodes in N1.

[0131] (1) First add a neighbor node with a willingness value of ALWAYS to set M;

[0132] (2) Obtain the geographical location of node N1 from the neighbor table, calculate the distance between nodes, and calculate the link stability LS and depth D(y) according to formula (4.9);

[0133] (3) If a two-hop neighbor can only be connected to the point of the present invention through a one-hop neighbor, add this one-hop neighbor to set M and remove the two-hop neighbors it covers;

[0134] (4) If any node in N1 is already the MPR of other nodes, then select the node with the largest LS as the MPR of the central node. Otherwise, proceed to the next step.

[0135] (5) Calculate the reachability of node N1 and execute the following steps in sequence:

[0136] (a) Add the node with the highest willingness to the set M;

[0137] (b) Among multiple candidates, select the node with the largest LS to join M;

[0138] (c) Among multiple candidates, select the node with the highest reachability to join M;

[0139] (6) Check if set M covers all two-hop neighbors N2. If it does, M is the final MPR set; otherwise, return to step (4).

[0140] The optimized MPR selection algorithm maximizes the utility of each MPR node by reusing nodes already used as MPRs, reducing the overall number of nodes that need to be broadcast in the network. Simultaneously, selecting nodes with higher link stability as MPRs reduces the risk of link failures, enhancing network reliability and stability.

[0141] 1.1.1 Overall Flow of the TC-M-OLSR Protocol

[0142] Figure 5 A schematic diagram of the overall process of the TC-M-OLSR protocol is shown. Similar to the protocol in Chapter 3, this protocol inherits the key processes of the OLSR protocol and optimizes two key processes: First, in the link-aware phase, nodes share local metric information with their neighbors and form cluster structures based on this information; second, the MPR selection process within the cluster is optimized by estimating the link stability of the RF / FSO network. Details of the relevant processes can be found in [link to relevant documentation]. Figure 4 The improvements made by the TC-M-OLSR protocol compared to the OLSR protocol are marked in gray in the diagram. These improvements enhance the protocol's adaptability to RF / FSO networks and improve its responsiveness and efficiency in the face of dynamic changes in the network environment.

[0143] The simulation results analysis of this invention specifically includes the following:

[0144] 1.1.2 Simulation Model Design

[0145] (1) Routing Protocol Model

[0146] The routing protocol model is consistent with the routing model in Section 3.4.1, and will not be repeated here.

[0147] (2) Channel Model

[0148] Consistent with the description in Section 3.4.1, the channel model mainly consists of two parts: shadowing fading and path loss.

[0149] (a) Shadow decay

[0150] In FSO networks, although there are models such as double gamma or exponential Weiber that are better suited to describe the effects of atmospheric turbulence, the log-normal model can also approximate the signal fading under relatively stable atmospheric conditions in order to simplify calculations as much as possible.

[66] Therefore, this chapter continues to use the log-normal model from the previous chapter as the shadow fading model.

[0151] (b) Path loss

[0152] In FSO communication networks, the method for calculating path loss has been detailed in Section 4.2.2. The path loss model based on FSO communication networks needs to be designed as a user-added third-party model. To integrate this model into the Exata simulation platform, the path loss calculation and related calling mechanism need to be added to the propagation.cpp source code file. Specifically, this includes:

[0153] Define the path loss function: Write a new function in propagation.cpp to calculate the path loss based on the FSO communication model. This function takes into account parameters such as transmission distance and atmospheric attenuation coefficient, and returns the calculated path loss value.

[0154] Modify the calling mechanism: Locate the call point for path loss calculation in the core code of the Exata simulation platform and modify it to call the new path loss function. Ensure that the newly defined path loss calculation method can be correctly called and applied during simulation.

[0155] The simulation scenarios and parameters of this invention specifically include the following:

[0156] The simulator uses the Exata simulation platform. To simulate real-world RF / FSO communication network scenarios, the network architecture adopts a structure similar to air-to-ground information networks. Communication between the ground network and high-altitude UAVs uses RF networking with an RF-based two-path fading model; communication between airships and between air and ground uses FSO networking with a custom FSO atmospheric fading model. The routing protocol used is the TC-M-OLSR protocol proposed in this chapter. To demonstrate the protocol's performance, it is compared with the standard OLSR protocol in the network simulation, and the following simulation experiments are designed:

[0157] (1) Verify the impact of node movement speed on network latency and packet delivery rate under different atmospheric conditions;

[0158] (2) Verify the impact of the average moving speed of nodes and the number of network nodes on broadcast routing overhead under the same atmospheric conditions.

[0159] The simulation scenario was set to a size of 10km × 10km, with the ratio of airships, drones, and ground units being 1:4.5:4.5. Figure 6 Taking the simulation scenario as an example, the total number of nodes is 80, with 8 airships, 36 drones, and 36 ground units. Airships, operating at high altitudes, have a larger scanning angle on their laser transceivers compared to drones. In the simulation, the scanning angle for airships is set to θ_scan_a = π rad, and the scanning angle for drones is set to θ_scan_u = π / 3 rad. Specific simulation scenarios and other simulation parameters are as follows... Figure 6 As shown in Table 4.1:

[0160] Table 4.1 Simulation Parameter Configuration

[0161]

[0162] The simulation performance metrics specifically include the following:

[0163] This invention also uses the average end-to-end latency and packet delivery rate of CBR services as evaluation metrics for network performance, with the calculation method described in Section 3.4.2. Furthermore, to more comprehensively measure network efficiency, this invention introduces the concept of Broadcast Control Overhead (BCO). The BCO of a node is defined as the ratio of the average number of bytes of control information sent per unit time to the average number of bytes of total data sent (including control data and service data), calculated as follows:

[0164]

[0165] Among them: B c B represents the average number of bytes of control information sent in the network.d This represents the average number of bytes sent in any data over the network.

[0166] The simulation results analysis of this invention includes the following:

[0167] To comprehensively evaluate the impact of network topology changes under different weather conditions on protocol performance, and to ensure adaptability to network scenarios and diversity in performance evaluation, this study designed multiple CBR service flows, including ground-to-ground, air-to-ground, and air-to-air flows, and conducted multiple rounds of simulation experiments. Statistical data were obtained by averaging the performance results of multiple service flows.

[0168] The changes in end-to-end delay and packet delivery rate when the node's maximum moving speed is between 0 and 25 m / s are as follows: Figure 7 and Figure 8 As shown, under the condition of consistent node movement speed, light fog results in greater path loss compared to clear weather. Correspondingly, at a maximum node speed of 5 m / s, the average latency in light fog is 27.3% higher than in clear weather. However, because the TC-M-OLSR protocol uses atmospheric fading as an evaluation factor for cluster head selection during cluster topology control, its latency is reduced by 2.6% and the packet delivery rate is increased by 2.8% in light fog compared to the OLSR protocol. On the other hand, because node motion factors are considered during cluster structure formation and MPR node selection, at a maximum node speed of 15 m / s, the average latency is reduced by 4.2% and the packet delivery rate is increased by 14.0%.

[0169] The broadcast control overhead varies with node movement speed and the number of nodes as follows: Figure 9 and Figure 10 As shown, Figure 9 This shows the case when the number of nodes is fixed at 60, while Figure 10 The maximum movement speed of the node is set to 5m / s.

[0170] As can be observed from the figure, the TC-M-OLSR protocol, by introducing a clustering topology control method based on the AOW algorithm and a link stability calculation method for RF / FSO networks, and optimizing the MPR selection mechanism based on this stability, significantly reduces broadcast control overhead compared to the OLSR protocol. While the control overhead of both protocols increases with increasing node movement speed and number, the TC-M-OLSR protocol reduces broadcast overhead by an average of 67.7% compared to the OLSR protocol in topology movement scenarios. When the number of nodes reaches 80, its overhead is reduced by 65.2%. These data indicate that the optimized protocol exhibits higher efficiency and stability in dynamic and large-scale network environments.

[0171] In summary, this invention first analyzes the limitations of traditional RF communication and the challenges encountered by existing OLSR protocols in RF / FSO networks. Secondly, it introduces the channel model of FSO networks, particularly the impact of atmospheric attenuation on signal transmission, and presents the node model and adaptive on-demand weighted clustering algorithm for RF / FSO networks. This invention describes the protocol design concept, proposing a clustering topology control method based on link quality and an MPR optimization strategy based on link stability. Finally, experimental results from Exata simulation software show that the optimized protocol exhibits superior performance under high mobility and atmospheric attenuation conditions, especially in environments with high transmission rate requirements.

Claims

1. A method for implementing the OLSR protocol in heterogeneous ad hoc networks based on topology control and MPR selection optimization, characterized in that, The method includes the following steps: Step 1: Cluster structure formation; (a) Initialization phase: The initial cluster ID of the node is set to the node ID, and the link quality factor threshold LF is set according to the specific network conditions. th ; (b) The current receiving node u receives a one-hop neighbor node N i When creating a HELLO group, it will obtain the neighbor node ID, cluster ID, and neighbor location information contained in the group; (c) Node u will compare its own cluster ID with that of its neighbor node N. i If the cluster IDs are equal, it indicates that the two nodes are already in the same cluster structure, and the subsequent process is skipped. (d) Calculate the link quality factor between the two nodes and compare it with the set quality factor threshold LF. th In comparison, if the link quality is less than the threshold, the link quality between the two points is considered to be low, and the two nodes are unlikely to form a stable cluster structure. If the link quality is greater than the threshold, the two nodes are considered to meet the conditions for clustering. (e) Make unified modifications to the cluster IDs of the two nodes and their fellow cluster members, and select the cluster ID value with the smaller cluster ID between the two nodes as the unified cluster ID. If the node also receives HELLO packets from other neighbors, it will return to stage (b) and repeat the entire process until the entire network forms a stable cluster structure. Step 2: Cluster head selection; After the clustering algorithm is completed, the AOW algorithm, which is adapted to the characteristics of RF / FSO networks, is used to select the cluster head, including: (a) Node u counts the number N of its neighboring nodes, which is defined as the degree d of node u. u The remaining node degree Du of node u is the difference between its access degree and physical degree K, calculated using the following formula: D u =|K-d u | (4.4) The physical degree K refers to the maximum number of concurrent connections that a node can support; (b) The relative moving speed of node u with respect to node v is V(u,v): V(u,v)=|V u -V v | (4.5) Where V u and V v Let M be the velocity of node u and node v, respectively. Then the average relative mobility M of node u within the cluster is... u for: (c) Ideal transmit power P of node u u for: Where, d u,v Let be the distance between nodes u and v, N1 represent the set of one-hop neighbors of node u, and ε be the atmospheric attenuation coefficient. Based on the previously calculated node mobility, remaining node degree, and ideal transmission power, the combined weight W of node u is... u for: Where ω d ω m ω p The weighting factors for mobility, remaining node degree, and transmission power are respectively, satisfying ω d +ω m +ω p =1. Finally, based on the comprehensive weight W of the nodes, the node with the smallest weight will be selected as the cluster head. If there are equal weights, the node with the smaller ID will be selected as the cluster head. Step 3: Inter-cluster connections; Before the protocol runs, several available communication frequencies are configured according to the needs of the scenario. During the protocol runs, the communication frequency band of the cluster to which the node belongs is updated in real time through the cluster ID of the current node. This process is periodically triggered by configuring a timer. For the gateway node at the boundary of two adjacent clusters, the frame period switches to the frequency band of the neighboring cluster with the smaller ID when the frame period is odd, and switches to the frequency band of the neighboring cluster with the larger ID when the frame period is even. The gateway node alternately switches the working frequency band between the two neighboring clusters.

2. The method for implementing the OLSR protocol in heterogeneous ad hoc networks based on topology control and MPR selection optimization according to claim 1, characterized in that, The method includes the following formula for calculating LS: Where α is the scanning angle θ of the laser transceiver. scan The weight of the impact on link stability, where distance d represents the actual distance between two nodes, and D max The maximum effective communication distance between two transceivers is ρ, which describes the ratio of the maximum effective communication distance to the optimal communication distance. This ratio is adjusted according to the performance of different transceivers. The value range of the link stability LS is [0,1].

3. The method for implementing the OLSR protocol in heterogeneous ad hoc networks based on topology control and MPR selection optimization according to claim 2, characterized in that, The optimized MPR selection process includes, if the MPR set selected by the central node is set M; The one-hop neighbor set of a node is N1; the two-hop neighbor set of a node is N2, which does not include the central node and nodes in N1, but includes: Step 1) First, add the neighbor node with the willingness value ALWAYS to set M; Step 2) Obtain the geographical location of node N1 from the neighbor table, calculate the distance between nodes, and apply the formula... Calculate the link stability LS and depth D(y); Step 3) If a two-hop neighbor can only be connected to a node through a one-hop neighbor, add this one-hop neighbor to set M and remove the two-hop neighbors it covers. Step 4) If any node in N1 is already the MPR of other nodes, select the node with the largest LS as the MPR of the central node; otherwise, proceed to the next step. Step 5) Calculate the reachability of node N1 and execute the following steps sequentially: (a) Add the node with the highest willingness to the set M; (b) Among multiple candidates, select the node with the largest LS to join M; (c) Among multiple candidates, select the node with the highest reachability to join M; Step 6: Check if set M covers all two-hop neighbors N2. If it does, M is the final MPR set; otherwise, return to step 4.

Images