Self-adaptive broadcasting method for ad hoc network in high-dynamic environment

Abstract

The invention belongs to the technical field of wireless communication, and particularly relates to an adaptive broadcasting method for an ad hoc network in a high-dynamic environment. According to the scheme, the network system comprises a high-dynamic mobile ad hoc network composed of k mobile nodes, the nodes operate under a synchronous time slot structure, and each node achieves neighbor discovery and link maintenance by periodically broadcasting HELLO messages. According to the method, optimization of neighbor discovery efficiency and energy consumption balance is taken as a target, and a dual-stage adaptive mechanism is adopted: rapid topological convergence is realized through fast / slow start dual-mode probability broadcast in an initial discovery stage, and an HELLO interval is dynamically adjusted based on link duration prediction and node density in a link maintenance stage. Compared with the prior art, the scheme of the invention can significantly improve the neighbor discovery efficiency in a high dynamic environment and reduce the system energy consumption without the support of a GPS (Global Positioning System).

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wireless communication, and particularly relates to a self-adaptive broadcast method for a high-dynamic environment self-organizing network. BACKGROUND

[0002] Mobile Ad Hoc Network (MANET) is a popular network architecture. High-dynamic mobile Ad Hoc networks constructed by high-speed nodes such as aircrafts are gradually evolving into a key support form of the new generation of tactical communication. Such networks are usually composed of high-speed flight platforms such as unmanned aerial vehicles, fighter jets, supersonic missiles, and have the characteristics of high dynamic, self-organization and distributed collaboration, and are widely used in key scenarios such as tactical communication and cluster collaboration. Compared with traditional ground mobile Ad Hoc networks, high-dynamic mobile Ad Hoc networks have larger scale and faster topology change, and the network operation highly depends on the rapid connection establishment and link maintenance between nodes, so an efficient neighbor discovery and maintenance mechanism plays a fundamental supporting role in such networks. In such networks, the HELLO message broadcast as a basic communication means for neighbor discovery, its sending interval strategy directly determines the convergence speed and energy consumption of the network. At present, the adjustment strategies used in the field mainly include a fixed interval-based HELLO mechanism (such as the AODV protocol) and a motion feature-based adaptive interval mechanism (such as the EEhello algorithm). However, the existing technologies still have inherent defects: the fixed interval mechanism is difficult to adapt to the rapid change of topology, and faces the core contradiction that the discovery efficiency and energy consumption are difficult to balance in a high-dynamic environment; and the motion feature-based adaptive mechanism improves the performance to a certain extent, but it relies on external precise positioning devices such as GPS to obtain speed or position information, which not only introduces additional hardware cost and power consumption, but also fails in GPS denial or restricted operational environments, which seriously limits its practical application and deployment on high-dynamic mobile Ad Hoc nodes with limited resources. Therefore, developing a neighbor discovery algorithm that does not rely on external positioning support, can adapt to high-dynamic environments and intelligently balance the discovery efficiency and energy consumption, has become a key technical problem to be solved in the field. SUMMARY

[0003] The technical problem this invention aims to solve is to overcome the shortcomings of existing neighbor discovery methods in highly dynamic mobile ad hoc networks. Specifically, existing technical solutions are polarized: First, fixed-interval HELLO mechanisms (such as the AODV protocol) are simple in principle but have low performance, making it difficult to adapt to rapid topology changes. In highly dynamic environments, they face the core contradiction of balancing neighbor discovery efficiency and energy consumption. Second, to improve performance, adaptive mechanisms based on motion characteristics (such as the EEhello algorithm) proposed in recent years have improved performance under specific conditions by introducing external sensing information to dynamically adjust strategies, representing the technological development direction in this field. However, such improved solutions (including EEhello) still have inherent defects: their core relies on external precise positioning devices such as GPS and radar to obtain speed or location information. This not only introduces additional hardware costs, power consumption, and system complexity, but also completely fails in real-world combat environments where GPS is denied or restricted, greatly limiting their deployment and application in actual highly dynamic combat scenarios. Therefore, this invention aims to make deeper innovations in the direction of improving adaptive mechanisms such as EEhello, and provide an adaptive interval broadcast method (AIB algorithm) for highly dynamic ad hoc networks to solve the key technical problems left by them: that is, to achieve the best balance between high efficiency and low energy consumption in neighbor discovery without any external positioning support and while maintaining low computational complexity, thereby promoting the technology from laboratory simulation to practical application.

[0004] The technical solution adopted in this invention is as follows:

[0005] An adaptive interval broadcast method for highly dynamic ad hoc networks is disclosed. This method is applied to a highly dynamic mobile ad hoc network consisting of k mobile nodes, where nodes operate under a synchronous time slot structure. Each time slot allows only one transmit or receive operation. Communication employs a fixed-radius broadcast mechanism and considers only a single receive channel. The method includes the following steps:

[0006] S1. Initialize parameters and set the base maximum probability. , communication radius Total number of time slots in system operation ( ), duration of state stability determination The total number of nodes in the current network Node cutoff time slot Number of neighbors discovered ; Nodes are randomly assigned to either fast-start or slow-start operating modes; Initialization ;

[0007] S2, the node is in the initial discovery phase, and the probability of a fast-starting node sending data is high. for:

[0008] ,

[0009] in , For the cut-off time slot Total number of time slots already used for transmission The current neighbor discovery rate, Indicates in time slot The number of HELLO messages successfully received. This represents the maximum transmission probability.

[0010] Slow-start node transmission probability for:

[0011] ,

[0012] in For the cut-off time slot The cumulative number of listening time slots, ;

[0013] S3, the probability of the node calculated in step S2. Broadcast a HELLO message;

[0014] S4. Based on the received HELLO message, the node dynamically estimates the number of neighbors in the network using the Expectation-Maximization (EM) algorithm. And update the maximum sending probability:

[0015] ,

[0016] in The base maximum probability is set. The estimated number of neighbors for a node. If in a continuous Within a given time slot, no new neighbors are added to the neighbor set (i.e., the number of neighbors observed by the node itself). If no neighbor is added, proceed to the S5 link maintenance phase; otherwise, return to step S2 and continue with the initial neighbor discovery.

[0017] S5. During the link maintenance phase, the node... Receive neighbor nodes The HELLO message, based on the received signal power Calculate the approximate distance between nodes The calculation formula is:

[0018] ,

[0019] in For transmission power, For received power, It is the product of transmit and receive gains. It is the carrier wavelength;

[0020] S6, Node Based on two consecutive time slots and Received from neighbor Calculate the radial velocity between nodes from two consecutive HELLO messages:

[0021] ,

[0022] in It is a time slot node and nodes Approximate distance between them It is a time slot node and nodes Approximate distance between them;

[0023] S7, Predicted Link Survival Time (LST) is And considering node density The impact, among which , , These are the length, width, and height of the network area, which dynamically adjust the broadcast interval of HELLO messages. ,in It's the safety factor. For nodes To the neighbor The lifespan of this link, For nodes In the time slot The set of neighbors;

[0024] S8. State rollback judgment: If the number of new neighbors detected during the link maintenance phase exceeds a preset threshold. If the neighbor discovery fails, return to S2 and start a new round of neighbor discovery; otherwise, remain in the maintenance phase and return to S5.

[0025] Furthermore, in step S4, the number of neighbors is dynamically estimated using the Expectation-Maximization (EM) algorithm. The specific implementation process includes the following sub-steps:

[0026] S41. Define the HELLO message reception state matrix. ,in It is a binary variable, representing the time slot. Did the data from the node successfully receive the data? The HELLO message;

[0027] S42. Expected Step (E-Step): Calculate the latent variables. Expectations:

[0028] ,

[0029] This method is based on the aforementioned two-stage node model, namely fast-start and slow-start. A hybrid model is used to represent these two types of nodes, where the category index... Used to distinguish node types: This indicates a fast-start node. This indicates a slow-start node. Where: Indicates category The node at the HELLO message transmission probability distribution for each time slot Indicates category The proportion of nodes in the population (mixing coefficient) satisfies and . Represents a node Does it belong to a category? And for each ,satisfy That is, each node can only belong to one category.

[0030] S43. Maximize step M-Step and update the mixing coefficient. and transmission probability :

[0031] , ,

[0032] S44. Update Steps: Calculate the population size estimate .

[0033] Furthermore, the safety factor used in step S8 for dynamically adjusting the HELLO interval... This is a key adjustment parameter, and its value range is designed to be... The open interval. The determination of this range is based on the following engineering considerations and theoretical foundations: safety factor. Its core function is to adjust the HELLO message sending interval. Compared with the predicted link survival time The proportional relationship between them. Its value directly affects the balance between the conservatism of link maintenance and system energy consumption:

[0034] when When it approaches 1, the HELLO interval Approaching the expected lifespan of the link This strategy tends to reduce the frequency of sending HELLO messages to lower energy consumption, but correspondingly, it reduces the predictability of link failures and the timeliness of recovery, resulting in higher risks.

[0035] when When approaching 0, the HELLO interval Much smaller This strategy manifests as very conservative link maintenance, ensuring extremely high link state awareness timeliness by frequently sending HELLO messages, but it significantly increases network energy consumption and control overhead.

[0036] Therefore, Strictly limited to Within the open interval, theoretically preventing This resulted in a complete failure of link maintenance. (time), and also avoids the problem of too short an interval ( The irrational performance consumption caused by time (time). In practical system design and simulation, The specific value can be fine-tuned according to the network's specific requirements for link reliability and energy consumption. For example, a smaller value can be selected in scenarios requiring high reliability. A value (such as 0.2) can be selected, while a larger value can be chosen in energy-sensitive scenarios. Values ​​(e.g., 0.8). This invention provides a flexible adjustment dimension for system optimization by introducing this safety factor.

[0037] This invention is based on the analysis of the network characteristics of highly dynamic ad hoc networks, where the efficiency and energy consumption of neighbor discovery primarily depend on the matching degree between the broadcast interval of HELLO messages and dynamic network changes. Existing fixed-interval mechanisms (such as AODV), due to their static nature, are difficult to adapt to rapid topology changes. While advanced dynamic-interval mechanisms (such as EEhello) can sense changes, their core reliance on external positioning devices to obtain motion information not only increases system complexity and cost but also fails in scenarios without GPS. Therefore, the fundamental starting point of this invention is to transform the problem of balancing neighbor discovery efficiency and energy consumption in highly dynamic environments into an adaptive interval adjustment problem based on local network observations, independent of external information. Although reinforcement learning methods such as Q-learning can achieve adaptive adjustment, their high computational complexity makes them difficult to deploy on resource-constrained nodes. To address this, this invention proposes an algorithm strategy that combines dual-stage adaptive probabilistic control with link prediction. The proposed AIB algorithm achieves rapid convergence in the initial discovery phase through fast / slow start dual-mode probabilistic broadcasting. In the maintenance phase, it dynamically adjusts the interval through link survival time prediction and node density perception. This enables the algorithm to intelligently balance discovery speed and energy consumption without prior location information, thereby improving the communication performance of highly dynamic ad hoc networks under low complexity conditions.

[0038] The beneficial effects of this invention are: it achieves an optimal balance between neighbor discovery efficiency and energy consumption in highly dynamic mobile ad hoc networks. On the one hand, the AIB algorithm proposed in this invention effectively avoids broadcast storms in the initial stage of the network through an adaptive probability transmission mechanism in the initial discovery phase, enabling rapid discovery of new neighbors. This reduces the convergence time in low- to medium-density networks by 60.0%-80.0% compared to the traditional AODV algorithm and by 40.0%-60.0% compared to the EEhello algorithm, significantly improving the topology construction speed. On the other hand, because the algorithm introduces link liveness prediction and adjusts the interval based on node density during the maintenance phase, the sending frequency of HELLO messages is highly matched with the actual stability of the links and the complexity of the network environment, significantly reducing the system's energy consumption. Simulation results show that its energy consumption is superior to the comparative algorithms in different scenarios. Therefore, the AIB scheme proposed in this invention can not only effectively adapt to the high-speed movement of nodes and changes in network density, exhibiting excellent robustness, but also does not rely on external hardware such as GPS, possessing low complexity and high practicality, providing an efficient neighbor discovery solution for highly dynamic application scenarios such as high-speed UAV swarms and aerial formations. The figure below visually demonstrates the performance advantages of the AIB algorithm compared to traditional methods in different scenarios. Attached Figure Description

[0039] Figure 1 For the comparison of node motion trajectories under different speed scenarios, (a) low-speed scenario. (b) high-speed scenario;

[0040] Figure 2 A comparison of the NDCR of three algorithms in low-speed scenarios. (a) 20-node scenario. (b) 40-node scenario. (c) 80-node scenario;

[0041] Figure 3 A comparison of the NDCR of three algorithms in high-speed scenarios. (a) 20-node scenario. (b) 40-node scenario. (c) 80-node scenario;

[0042] Figure 4 A comparison of the energy consumption of three algorithms in low-speed scenarios. (a) 20-node scenario. (b) 40-node scenario. (c) 80-node scenario;

[0043] Figure 5 A comparison of the energy consumption of three algorithms in high-speed scenarios. (a) 20-node scenario. (b) 40-node scenario. (c) 80-node scenario;

[0044] Figure 6Comparison of average link survival time for three algorithms in low-speed scenarios. (a) 20-node scenario. (b) 40-node scenario. (c) 80-node scenario;

[0045] Figure 7 Comparison of average link survival time for three algorithms in high-speed scenarios. (a) 20-node scenario. (b) 40-node scenario. (c) 80-node scenario. Detailed Implementation

[0046] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0047] This embodiment uses a highly dynamic mobile ad hoc network constructed from a high-speed drone swarm as the application scenario, as shown in the table below:

[0048] Table 1. AIB Algorithm Performance Simulation Parameter Settings

[0049]

[0050] This embodiment aims to demonstrate the full implementation details of the AIB algorithm in the initial discovery phase and the link maintenance phase.

[0051] The specific steps are as follows:

[0052] (a) Parameter initialization

[0053] Node initialization time It randomly selects between fast-start and slow-start working modes. The node motion trajectory comparison under different speed scenarios is as follows: Figure 1 As shown. For example, if a node is determined to be a fast-start type, its initial transmission probability is dynamically calculated based on the neighbor discovery rate; if it is a slow-start type, it is adjusted based on the received message statistics.

[0054] (ii) Probability broadcasting during the initial discovery phase

[0055] Fast-start nodes in time slots The formula for calculating the transmission probability is:

[0056]

[0057] in Indicates the node in the time slot If a device needs to be discovered by all nodes within the communication range, the number of additional broadcast time slots required is dynamically calculated based on the current discovery rate and the number of time slots already used. , For the cut-off time slot Total number of time slots already used for transmission The current neighbor discovery rate, .in Indicates the node's cutoff time slot Number of neighbors discovered Indicates in time slot Number of successfully received HELLO messages.

[0058] The probability of a slow-start node sending data is calculated as follows:

[0059]

[0060] Its transmission probability increases with time, among which For the cut-off time slot The cumulative number of listening time slots, Indicates the node in the time slot In order to be discoverable by all nodes within the communication range, the total number of time slots required is calculated using the following relationship: The nodes broadcast HELLO messages with probability to achieve an adaptive sending frequency.

[0061] (III) Neighbor Number Estimation and Probability Update

[0062] Based on the received HELLO messages, the number of neighbors is dynamically estimated using the Expectation Maximization (EM) algorithm. HELLO message reception status matrix This records the history of nodes receiving HELLO messages. , Indicates time slot Was the node successfully received? The message states that the expected values ​​of the latent variables are calculated via E-Step, and the mixing coefficients are updated via M-Step. and transmission probability Final estimate .according to Adjust the maximum transmission probability to balance network load.

[0063] (iv) State transition judgment

[0064] If continuous If no new neighbors are found within a time slot, the node enters the link maintenance phase; otherwise, it returns to step S2 to continue the initial discovery.

[0065] (v) Distance and speed estimation during link maintenance phase

[0066] The foundation of the maintenance phase algorithm is accurate estimation of the distance between nodes. It assumes that in a highly dynamic mobile ad hoc network, the transmit power of nodes is constant, and the nodes satisfy the free-space path loss model. (dB) represents the power loss of the signal. For transmission power, Received power. Transmitted power. With received power The ratio reflects the degree of signal power attenuation. When electromagnetic waves propagate in free space, their power density increases with distance. It decays with the square of the sum of its parts. It is the product of transmit and receive gains. It is the carrier wavelength. As shown in the following formula:

[0067]

[0068] When a node receives a data packet, it uses the received power... The distance is calculated based on the magnitude of the received power. The node can be calculated based on the received power. and Approximate distance between :

[0069]

[0070] Assuming nodes At any moment and Received from neighbors respectively Two consecutive HELLO messages, based on the calculated distances, are respectively and Then the radial velocity between nodes can be calculated:

[0071]

[0072] Positive values ​​indicate that the distance between nodes increases (nodes move further apart), while negative values ​​indicate that the distance decreases (nodes move closer together).

[0073] (vi) Link Survival Time Prediction and Interval Adjustment

[0074] Based on the current distance between nodes and radial velocity This allows for the estimation of the link's expected lifetime (LET). For a communication radius of... In this case, the expected link lifetime is calculated as follows:

[0075]

[0076] Taking into account radial velocity and node density, the nodes dynamically adjust the sending interval of HELLO messages:

[0077]

[0078] in This is a safety factor used to adjust the ratio between the HELLO interval and the expected link lifespan. Its value ranges from (0,1), with smaller values ​​preferred. This allows for a more conservative link maintenance strategy; Based on formula The algorithm calculates the node density. Unlike existing methods, this invention's algorithm considers not only radial velocity but also node density as a key factor in adjusting the HELLO interval. In high-density areas, higher node density means more potential neighbors and more complex topology changes, requiring more frequent HELLO message exchanges to maintain link state accuracy; while in low-density areas, the relatively sparse node distribution reduces the frequency of link changes, allowing for a more extended HELLO interval to save energy.

[0079] (vii) State rollback judgment

[0080] If the number of new neighbors detected during the maintenance phase exceeds the threshold If the node returns to the initial discovery phase, it will remain in the maintenance phase.

[0081] (viii) Output Results

[0082] The algorithm ultimately outputs a list of neighbors and an adaptive HELLO interval. This is used for subsequent routing and collaborative decision-making.

[0083] Simulation tests were conducted using the method described in this invention. Simulation tests were conducted in both low-speed (220–280 m / s) and high-speed (500–600 m / s) scenarios, taking into account the effects of different network densities (sparse 20 nodes, medium 40 nodes, dense 80 nodes). Figures 2-7 The simulation results for the embodiments show that, compared with the traditional AODV and the improved EEhello, AIB demonstrates significant advantages in three key performance indicators: convergence speed is improved by 40%-80% in low / medium density networks; in high-density, high-speed scenarios, the neighbor discovery completion rate is improved by up to 10% compared to AODV and 5% compared to EEhello; energy consumption is reduced by 10.81%-67.69%; and average link lifetime is improved by 1.34%-31.76%. The experimental results fully verify the effectiveness of the AIB algorithm in highly dynamic mobile ad hoc network environments, providing an efficient neighbor discovery solution for applications such as high-speed UAV swarms and aerial formations.

Claims

1. An adaptive broadcasting method for ad hoc networks in highly dynamic environments, applied to a highly dynamic mobile ad hoc network consisting of k mobile nodes, where nodes operate under a synchronous time slot structure, each time slot can only perform one transmit or receive operation, communication adopts a fixed-radius broadcasting mechanism, and only a single receiving channel is considered; characterized in that, The method includes the following steps: S1. Initialize parameters and set the base maximum probability. , communication radius Total number of time slots in system operation ( ), duration of state stability determination The total number of nodes in the current network Node cutoff time slot Number of neighbors discovered ; Nodes are randomly assigned to either fast-start or slow-start operating modes; Initialization ; S2, the node is in the initial discovery phase, and the probability of a fast-starting node sending data is high. for: ， in , For the cut-off time slot Total number of time slots already used for transmission The current neighbor discovery rate, Indicates in time slot The number of HELLO messages successfully received. This represents the maximum transmission probability. Slow-start node transmission probability for: ， in For the cut-off time slot The cumulative number of listening time slots, ; S3, the probability of the node calculated in step S2. Broadcast a HELLO message; S4. Based on the received HELLO message, the node dynamically estimates the number of neighbors in the network using the Expectation-Maximization (EM) algorithm. And update the maximum transmission probability: ， in The base maximum probability is set. The estimated number of neighbors for a node, if in consecutive Within a given time slot, no new neighbors are added to the neighbor set, meaning the node itself observes the number of its neighbors. If no neighbor is added, proceed to the S5 link maintenance phase; otherwise, return to step S2 and continue with the initial neighbor discovery. S5. During the link maintenance phase, the node... Receive neighbor nodes The HELLO message, based on the received signal power Calculate the approximate distance between nodes The calculation formula is: ， in For transmission power, For received power, It is the product of transmit and receive gains. It is the carrier wavelength; S6, Node Based on two consecutive time slots and Received from neighbor Calculate the radial velocity between nodes from two consecutive HELLO messages: ， in It is a time slot node and nodes Approximate distance between them It is a time slot node and nodes Approximate distance between them; S7, Predicted Link Survival Time (LST) is And considering node density The impact, among which , , These are the length, width, and height of the network area, which dynamically adjust the broadcast interval of HELLO messages. ,in It's the safety factor. For nodes To the neighbor The lifespan of this link, For nodes In the time slot The set of neighbors; S8. State rollback judgment: If the number of new neighbors detected during the link maintenance phase exceeds a preset threshold. If the neighbor discovery fails, return to S2 and start a new round of neighbor discovery; otherwise, remain in the maintenance phase and return to S5.

2. The adaptive broadcasting method for ad hoc networks in highly dynamic environments according to claim 1, characterized in that, In step S4, the Expectation-Maximization (EM) algorithm is used to dynamically estimate the number of neighbors in the network. The specific method is as follows: S41. Define the HELLO message reception state matrix. ,in It is a binary variable, representing the time slot. Did the data from the node successfully receive the data? The HELLO message; S42. Expected Step (E-Step): Calculate the latent variables. Expectations: ， Based on a two-stage node model, namely fast-start and slow-start, a hybrid model is used to represent these two types of nodes, where the category index... Used to distinguish node types: This indicates a fast-start node. Indicates a slow-start node; Indicates category The node at the HELLO message transmission probability distribution for each time slot Indicates category The proportion of nodes in the total population satisfies and , Represents a node Does it belong to a category? And for each ,satisfy That is, each node can only belong to one category; S43. Maximize step M-Step and update the mixing coefficient. and transmission probability : ， ， S44. Update Steps: Calculate the population size estimate .

3. The adaptive broadcasting method for ad hoc networks in highly dynamic environments according to claim 1, characterized in that, The safety factor mentioned in step S8 The value range is (0,1).

Images