Shipborne Network Wireless Access Point Deployment Planning Method

Abstract

This invention proposes a deployment planning method for shipborne network wireless access points, belonging to the field of deployment planning. During the iterative optimization of access point locations, this method constructs a movement distance based on the attractiveness between the access point (AP) and hotspot areas, as well as the repulsive force between APs, and moves the access point accordingly. It uses c-degree coverage and average signal strength as optimization objectives to screen the best candidate solution for this round; this process is repeated to complete the access point deployment. Furthermore, this method flexibly sets the receiving sensitivity of receivers in different areas to meet the differentiated needs of users, thereby achieving differentiated deployment in WLAN scenarios and improving network performance in hotspot areas. Using this invention, with the same number of APs deployed, network coverage can be improved, and users in hotspot areas can enjoy a better network performance experience.

Description

Technical Field

[0001] This invention relates to the field of deployment planning technology, and specifically to a deployment planning method for a shipborne network wireless access point (AP). Background Technology

[0002] Ship network systems have evolved from analog signal transmission to digital signal transmission, distributed control, and finally bus control. However, with the rapid development of ship automation and the widespread application of wireless sensors, shipboard network systems need to be reconsidered. With the rapid development of wireless network technology, wireless networks have seen significant improvements in bandwidth, transmission rate, reliability, cost, and ease of installation and maintenance, and are poised to replace wired networks as the mainstream network system. Furthermore, with the continuous development of IoT technology, in-cabin wireless networks will become the primary communication method.

[0003] Using a wireless network system within ship cabins not only eliminates the hassle of cabling but also significantly reduces the cost of the system, making it only one-tenth the cost of a wired network. Wireless Local Area Networks (WLANs) are a key representative of wireless communication technology. Having undergone numerous standard upgrades and updates, WLANs have become one of the most widely used and efficient network access technologies. From a long-term cost perspective, wireless network systems facilitate equipment maintenance and replacement. When upgrading the networks of older ships, wireless networks solve the problems of drilling and cabling, offering significant advantages in reducing construction costs and shortening the construction cycle. For military vessels, wireless network equipment is flexible to deploy and resistant to damage, making it the best choice for backup network systems. Designing a stable and reliable wireless network system with full ship coverage, within the complex structure and environment of ships, is a key focus of shipboard wireless network research.

[0004] Then, the widespread application of shipboard wireless network systems and various wireless access devices has led to an increasingly scarce spectrum resource. At the same time, electromagnetic interference on wireless channels cannot be ignored. On the other hand, with the rapid development of network services such as gaming, video calls, and short videos, the communication performance and quality of shipboard wireless networks have also attracted much attention. Optimizing the coverage and performance of wireless networks within cabins has become paramount. In military vessels, wireless access devices have the important requirement of flexible deployment and indestructibility, and their communication quality is particularly important. Therefore, when planning shipboard wireless networks, it is essential to design a stable, high-quality, reliable, and ship-wide covered wireless network system. The internal space structure of military vessels is complex and highly enclosed. Various obstacles and bulkheads within cabins can cause varying degrees of attenuation of wireless signals. Furthermore, the confined or semi-enclosed spaces within cabins can cause reflection, transmission, and scattering phenomena. Therefore, when conducting deployment planning studies for cabins, the impact of non-line-of-sight transmission and multipath effects must be considered.

[0005] However, conducting on-site measurements of wireless network performance metrics requires significant financial and human resources, and is also time-consuming and labor-intensive. In contrast, using software to simulate wireless networks is convenient, reliable, cost-effective, and efficient. It can predict the propagation characteristics of wireless channels and allow for advance planning of wireless networks. Currently, extensive research is being conducted both domestically and internationally on the use of metaheuristic optimization algorithms for wireless network deployment planning.

[0006] In the research on coverage optimization of wireless networks, Zheng et al. designed a deployment optimization algorithm based on genetic algorithm to deploy as few APs as possible while meeting the indoor positioning requirements of network performance (see Zheng Y, Liu J, Sheng M, et al. Toward practical access point deployment for angle-of-arrivalbased localization[J]. IEEE Transactions on Communications, 2020, 69(3):2002-2014.). However, due to the use of a statistical model in the selection of signal model, the predicted signal loss has a certain deviation from the actual situation. Yang et al. designed a fine-grained AP deployment strategy (see Yang Y, Zhou A, Ma H. FineAP: fine-grained access point deployment strategy for 60GHz millimeter-wave wireless networks[J]. IEEE Communications Letters, 2022, 27(1):381-385.). This strategy first uses ray tracing in a two-dimensional scene to determine the signal distribution at each location in the environment, and then uses a hybrid fruit fly algorithm based on the principle of maximizing signal strength to determine the AP deployment location. The disadvantage of this method is that it only considers two-dimensional modeling.

[0007] Existing work mainly focuses on WLAN deployment planning in two-dimensional environments, and statistical models are often used for signal loss prediction, resulting in low accuracy and discrepancies between predictions and reality. However, a few papers have studied WLAN deployment planning in three-dimensional environments. For example, Mustapha et al. obtained approximate signal strengths for all deployment schemes through brute-force combination by actually measuring WLAN signal strength, and then selected the optimal deployment scheme (see Mustapha WNFW, Aziz MAA, Masrie M, et al. WLAN approximated strength measurement method with brute force algorithm for a minimum number of APs and maximum WLAN coverage[C] / / 2020IEEE 10th Symposium on Computer Applications & Industrial Electronics (ISCAIE).IEEE,2020:180-185.). However, the actual measurement method is labor-intensive and costly, and is only suitable for small deployment areas and limited deployment schemes, not for large-scale scenarios. Shareef et al. used simulation software based on a three-dimensional ray tracing model to calculate the signal strength at the receiving point of the alternative deployment schemes and selected the optimal scheme as the deployment result. However, the algorithm did not fully consider all situations and could only determine the better scheme among specific schemes (see Shareef OA, Abdulwahid MM, Mosleh MF, et al. The optimal location for access point deployment based on RSS for indoor communication[J]. International Journal of Simulation,2019,20(S1):2.1-2.6.).Chen et al. used the Golden Jackal optimization algorithm to jointly optimize the location and transmission power of each access point while ensuring effective global coverage in the three-dimensional environment, thereby minimizing the overall transmission power of all APs (see Chen L, Xu F, Jin K, et al. Energy-saving access point configurations in WLANs: a swarm intelligent approach[J]. The Journal of Supercomputing, 2023, 79(17): 19332-19364.). Du et al. considered the three-dimensional environment, introduced a ray tracing algorithm to calculate signal transmission loss, and proposed a genetic algorithm to optimize the signal strength of the network (see Du J, Xiong W, Wang J, et al. A heuristic ap deployment approach for industrial wireless networks[C] / / 2021 China Automation Congress (CAC). IEEE, 2021: 8035-8040.). However, Chen and Du's schemes did not consider the necessity of non-uniform deployment, and differences in user needs would lead to the deployment of redundant access points. Liu et al. considered the real 3D environment and fully considered the differentiated coverage requirements (see Liu P, Hu Q, Jin K, et al. Toward the energy-saving optimization of WLAN deployment in real 3-D environment: a hybrid swarm intelligent method[J]. IEEE Systems Journal, 2021, 16(2): 2425-2436.). They used a swarm intelligence algorithm combining the fruit fly algorithm and the particle swarm algorithm to jointly optimize the transmission power and location of each AP, minimizing the total AP transmission power while ensuring effective coverage. However, these methods do not consider the actual user performance requirements and are difficult to meet users' expectations for network performance, so there is still much room for improvement. Summary of the Invention

[0008] In view of this, the present invention provides a method for planning the deployment of wireless access points for shipborne networks. This method can utilize the attractive and repulsive forces acting on the access points to construct the movement distance, optimize the location of the access points, and obtain the final access point deployment scheme. This improves the overall performance of the shipborne network and solves the problem that current deployment planning methods cannot meet user performance requirements.

[0009] To solve the above-mentioned technical problems, the present invention is implemented as follows.

[0010] This invention provides a method for planning the deployment of wireless access points (APs) in a shipborne network, which is achieved through iterative optimization of candidate solutions composed of APs. Its key feature is that, during iterative optimization, a moving distance is constructed based on the attractive force between the APs and hotspot areas, as well as the repulsive force between APs, to move the access points; the optimal candidate solution is selected using C-degree coverage and average signal strength as optimization objectives; this process is repeated to complete the access point deployment.

[0011] The specific implementation steps of this shipborne network wireless access point deployment planning method include:

[0012] Step S1: Randomly move the access point AP of each candidate solution in the first candidate solution set to obtain the second candidate solution set; the candidate solution set includes multiple candidate solutions, and each candidate solution includes multiple access points AP;

[0013] Step S2: For the second candidate solution set, predict the signal strength of the access point (AP) at the receiving point in the candidate solution;

[0014] Step S3: Based on the positional relationship of the access points (APs) in the candidate solutions and the signal strength, calculate the attractive force between the access points (APs) and the hotspot area, as well as the repulsive force between the access points (APs);

[0015] Step S4: Decompose the resultant force of the repulsive and attractive forces on the access point AP into xy-axis components in the plane coordinate system, determine the moving distance of the access point AP according to the xy-axis components, and apply it to the second candidate solution set to obtain the third candidate solution set;

[0016] Step S5: For the third candidate solution set, calculate the target value for each candidate solution. The target value consists of c-degree coverage and average signal strength. Select the candidate solution with the largest target value that is greater than the largest target value in the previous round as the best candidate solution.

[0017] Step S6: Use the best candidate solution of this round as the initial position of the candidate solution of the next round of iteration, and iterate in the manner of steps S1-S5 to obtain the next best candidate. Repeat this process to complete the deployment of the access point.

[0018] In step S1, the access point AP of each candidate solution in the first candidate solution set is randomly moved to obtain the second candidate solution set:

[0019] Within the preset maximum travel distance d max Under the constraints, the movement distance of the x and y components in the current location of the access point (AP) is randomly determined and superimposed on the current location to obtain the new location of the access point (AP), forming the second candidate solution set.

[0020] In step S2, the signal strength at the access point AP in the predicted candidate solution is predicted by using a ray tracing algorithm based on SBR / IM to predict the signal strength at the receiving point.

[0021] In step S3, the attractive force and the repulsive force are calculated as follows:

[0022] Set up access point (AP) q The attractive force experienced at the receiving point H for:

[0023]

[0024] Among them, AP q Defined as the q-th access point after sorting the signal strength of each access point AP at the receiving point H; k1 is the attraction coefficient, used to adjust the strength of the attraction. Access Point (AP) q Signal strength at receiving point H; R th The set receiving sensitivity; Indicates access point (AP) q Distance between the receiving point H and the receiving point H; Indicates access point (AP) q The direction of the force between the receiving point H and the receiving point H; when Less than the set receiving sensitivity R th At that time, the access point AP will be generated. q The attractive force pulling towards the receiving point H;

[0025] Set up access point (AP) i With AP j Repulsive force between for:

[0026]

[0027] Where r represents the repulsive force coefficient; d th Represented as distance threshold; d ij Indicates access point (AP) i With AP j The Euclidean distance between them; α ij Indicates that it is from the access point AP i Pointing to AP j The direction, α ij +π represents the direction of the repulsive force; when the distance between two access points (APs) is less than the distance threshold d... th When this occurs, a repulsive force is generated.

[0028] In step S4, the resultant force of the repulsive and attractive forces acting on the access point AP is decomposed into xy-axis components in a planar coordinate system as follows:

[0029] For each access point (AP) in the second candidate solution set, use AP q This indicates that the access point (AP) will be... q Project all attractive and repulsive forces onto the x-axis and y-axis of the plane coordinate system, respectively, and calculate the magnitude of the resultant force in both directions:

[0030]

[0031] Among them, F q_x and F q_y Access Points (APs) q The magnitude of the resultant force acting on the x-axis and y-axis; and These represent the receiver point k and the access point AP, respectively. q The attraction is decomposed into magnitudes along the x and y axes; F qj_x and F qj_y These represent access points (APs). j AP q The repulsive force is decomposed into magnitudes along the x-axis and y-axis, and is negative.

[0032] Preferably, the movement distance is constructed by moving the access point as follows:

[0033] Let F be the x-axis component of the resultant force of the repulsive and attractive forces acting on the access point AP. q_x The y-axis component is F q_y Then, the distances Δx and Δy represented by exponential functions are:

[0034]

[0035] Among them, F q For F q_x and F q_y The resultant force, d max The maximum distance to move is set.

[0036] The access point is moved using the moving distances Δx and Δy.

[0037] Preferably, the method further includes: before applying the movement distance constructed based on attractive and repulsive forces to the access point, determining whether the xy-axis component of the resultant force of the repulsive and attractive forces is greater than a set limiting movement threshold F. th If so, update the location of the access point (AP) by moving the distance; otherwise, keep the AP's location unchanged.

[0038] The optimization objective used in this invention for iterative optimization is composed of c-degree coverage and average signal strength:

[0039]

[0040] Among them, S t Let F(c) be the objective value for optimizing the candidate solution, and F(c) be the c-degree coverage of the entire receiving area. α represents the normalized average signal strength; α and β are the importance weights.

[0041]

[0042] Where M and N are the maximum number of rows and columns of the receiving point grid within the receiving area, respectively; f(c,i,j) indicates whether the receiving point at coordinates (i,j) satisfies c-degree coverage, if the receiving point (i,j) can receive signals with a strength higher than the set receiving sensitivity R from at least c different access points (APs). min If the signal is c, then it is determined that the c-degree coverage is satisfied, and f(c,i,j)=1; otherwise, f(c,i,j)=0.

[0043]

[0044] Where, r ij r represents the signal strength value before normalization at (i,j); max r represents the maximum absolute value in the RSSI matrix; ij ′ represents the normalized signal strength value.

[0045] Preferably, the receiving sensitivity R min The acquisition method is as follows: based on the modulation and coding scheme (MCS) index of the access area where the access point (AP) is located, the pre-configured MCS index and the receiver sensitivity R are queried. min We obtain the corresponding relationship.

[0046] After obtaining the optimal solution in step S6, redundancy detection and optimization are further performed on the AP layout corresponding to the optimal solution, specifically as follows:

[0047] Calculate the average signal strength of each access point and sort the access points in ascending order according to the average value; then start from the access point with the poorest signal coverage. x Starting from this point, the Pearson correlation coefficient vectors with other access points are calculated sequentially; if the similarity is higher than a set value, further verification is performed to delete the access point AP. x Whether the network performance is affected afterward, if the impact is less than the set value, then it is confirmed that redundant deployment has occurred, and only then can the access point AP be removed. x delete.

[0048] This invention utilizes the attractive and repulsive forces acting on access points to construct the movement distance, optimizing access point locations and obtaining a final access point deployment scheme. Regarding the attractive force, as signal strength decreases and distance increases, the magnitude of the attractive force gradually increases to enhance the signal strength at that location, meeting the c-degree coverage requirement and optimizing the access point location towards the preferred direction. However, if only the attractive force generated by hotspot locations on APs is considered, multiple APs may become excessively close together. This not only leads to severe signal interference but may also affect coverage in other locations. Therefore, this invention further introduces the repulsive force between APs. The attractive and repulsive forces combine to determine the movement distance of the access point towards the preferred location. This scheme can improve network performance in hotspot areas, solving the problem that current deployment planning methods cannot meet user performance requirements. Through comparative tests of multiple algorithms, the results show that the overall performance of this invention's scheme is excellent.

[0049] This invention flexibly adjusts the receiving sensitivity of receivers in different areas based on the MCS index. By comparing the Pearson correlation coefficient matrix of the signal strength matrices of two AP nodes, it determines whether node redundancy occurs, effectively detecting redundant nodes. The system performance change after removing the node confirms whether removal is necessary, thus reducing deployment costs while meeting coverage requirements. A c-degree coverage measurement method is proposed, ensuring that each location in the wireless network receives a sufficiently strong signal from at least c access points, thereby improving network redundancy and reliability, and significantly enhancing network performance and user experience.

[0050] This invention designs an exponential form expression for the distance traveled. The distance traveled in this invention is proportional to the magnitude of the force; however, directly using the magnitude of the resultant force may lead to excessive travel, thus affecting the stability of the algorithm. Therefore, the inventors introduce an exponential function to avoid this phenomenon and set a threshold F to limit the travel. th Only when the component size is greater than the threshold will it have an effect on the AP's position, thus ensuring the rationality of the movement distance and avoiding the deterioration of the AP's position due to movement. Attached Figure Description

[0051] Figure 1 This is a flowchart of the shipborne network wireless access point deployment planning method of the present invention.

[0052] Figure 2 This is a schematic diagram of the force analysis.

[0053] Figure 3 This is a diagram showing the decomposition of forces.

[0054] Figure 4 It is a schematic diagram of a 3D scene.

[0055] Figure 5 This is a comparison of overall regional coverage.

[0056] Figure 6 This is a comparison of coverage rates in hotspot areas.

[0057] Figure 7 This is a comparison of the average signal strength across the entire region.

[0058] Figure 8 This is a comparison of the average signal strength in hotspot areas.

[0059] Figure 9 This is the result of comparing algorithm stability.

[0060] Figure 10 This is a heatmap of the proposed optimal deployment scheme. Detailed Implementation

[0061] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0062] This invention provides a method for planning the deployment of wireless access points for shipborne networks. The basic idea is to optimize the access point location by determining the attractive and repulsive forces acting on the access point based on the predicted signal strength, determining the movement distance based on the resultant force, and optimizing the access point location. The optimal solution is selected by using an index composed of c-degree coverage and average signal strength, and the optimal solution is obtained through multiple iterations to complete the access point deployment.

[0063] As can be seen, this invention optimizes access point locations based on attractive and repulsive forces. For attractive forces, the magnitude of the attractive force gradually increases as signal strength decreases and distance increases, thus enhancing the signal strength at that location and meeting the c-degree coverage requirement, optimizing the access point location towards the preferred direction. However, if only the attractive force generated by hotspot locations on the APs is considered, multiple APs may become excessively close together. This not only leads to severe signal interference but may also affect coverage in other locations. Therefore, this invention further introduces repulsive forces between APs. The attractive and repulsive forces combine to determine the distance the access point should move towards the preferred location. This scheme can improve network performance in hotspot areas and solve the problem that current deployment planning methods cannot meet user performance requirements.

[0064] Figure 1 The specific implementation process of the shipborne network wireless access point deployment planning method in this embodiment of the invention is shown, including the following steps:

[0065] Step 1: Initialization.

[0066] In this step, WLAN network-related parameters are initialized, including the specific scene size, obstacle parameters and locations, the rated effective coverage radius R of the access point, and the number of access points deployed, F, etc.

[0067] Initialize the relevant parameters of the deployment planning algorithm based on the correlation of predicted signals, including the number of candidate solutions popSize, the iteration termination threshold threshold, the maximum number of iterations maxIter, and the maximum movement distance d. max wait.

[0068] Step 2: Randomly initialize the positions of popSize candidate solutions to form the first candidate solution set.

[0069] The first candidate solution set includes popSize candidate solutions, and each candidate solution includes F access points (APs). Let the t-th candidate solution be represented as P. t (X t ,Y t The coordinates of the access point AP of the t-th candidate solution are represented as P. t,i (x t,i ,y t,i ), where t∈[1,popSize], i∈[1,F].

[0070] Step 3: Begin iterative optimization calculation. For each candidate solution's access point AP in the first candidate solution set, calculate the maximum travel distance d. max Randomly select the movement step size d r The access point (AP) is randomly searched in all directions to obtain a second set of candidate solutions.

[0071] The search process in step 3 is as follows:

[0072] x t ′ ,i =x t,i +d r (1)

[0073] y t ′ ,i =y t,i +d r (2)

[0074] Where, d r For the interval [-d max ,d max Random numbers within ]

[0075] Step 4: For the second candidate solution set, predict the signal strength of the access point (AP) at the receiving point in the candidate solution.

[0076] In this step, the signal strength at the receiving point is predicted based on the SBR / IM-based ray tracing algorithm.

[0077] Step 5: Based on the location relationship and signal strength of the access points (APs) in the candidate solutions, calculate the attractive force between the access points (APs) and the hotspot areas, as well as the repulsive force between the access points (APs).

[0078] In this step, the magnitude and direction of the repulsive and attractive forces acting on the AP nodes of each candidate solution are calculated based on the association relationships of AP deployments. Here, the attractive force refers to the attraction between the hotspot area and the AP, while the repulsive force refers to the repulsive force relationship between APs.

[0079] Set up access point (AP) q The attractive force experienced at the receiving point H for:

[0080]

[0081] Among them, AP q Defined as the q-th access point AP after sorting the signal strength at the receiving point H, where q represents the AP's sequence number; k1 is the attraction coefficient, used to adjust the strength of the attraction; R th The set receiving sensitivity; Indicates access point (AP) q Distance between the receiving point H and the receiving point H; Defined as an access point (AP) q The direction of the force between the receiving point H and the receiving point H. When Less than the set receiving sensitivity R th At that time, AP will be generated q The attractive force pulls towards location H. As signal strength decreases and distance increases, the magnitude of the attractive force gradually increases to enhance the signal strength at that location and meet the c-degree coverage requirement.

[0082] If only the attraction of hotspot locations to access points (APs) is considered, multiple APs may become excessively close together. This not only causes severe signal interference but may also affect coverage in other locations. Therefore, a repulsive force between APs needs to be introduced into the correlation analysis. The purpose of this repulsive force is to ensure that APs maintain a certain distance from each other and avoid excessive concentration. This repulsive force can be calculated based on the Euclidean distance between APs; the closer the distance, the greater the repulsive force. (Two access point APs) i With AP j The repulsive forces between them are as follows:

[0083]

[0084] Where r represents the repulsive force coefficient, which can be selected through simulation. d th Represented as distance threshold; d ij Represented as Access Point (AP) i With AP j The Euclidean distance between them; α ij Indicated by access point AP i Pointing to AP j The direction, α ij +π represents the direction of the repulsive force. From the above formula, it can be seen that when the distance between two access points (APs) is less than the distance threshold d... th When the distance between APs decreases, a repulsive force is generated. The repulsive force increases as the distance between APs decreases, preventing the APs from getting too close.

[0085] Step 6: Decompose the force on the access point AP, calculate the distance and direction of movement, and adjust the position of the candidate solutions to obtain the third candidate solution set.

[0086] In this step, the resultant force of the repulsive and attractive forces on the access point AP is decomposed into xy-axis components in a planar coordinate system. The movement distance of the access point AP is determined based on the xy-axis components and applied to the second candidate solution set to obtain the third candidate solution set.

[0087] In a preferred embodiment, during the force decomposition process, the attractive and repulsive forces experienced by each access point need to be projected onto the x-axis and y-axis of the planar coordinate system, respectively, to calculate the forces experienced by each access point in the x and y directions. Figure 2 Analyze the forces acting on the access point (AP), assuming the attractive force of the reference point H on access point AP1 is... Its direction points towards the reference point H. This force can be decomposed into components in the x and y directions, as shown below:

[0088]

[0089] in, and They are respectively The components in the x and y directions; θ is The angle with the x-axis; for The amplitude, x H y H The coordinates of the reference point H; The coordinates of access point AP1; This is the distance between access point AP1 and reference point H.

[0090] Assume that the repulsive force of access point AP3 on access point AP1 is The direction points towards AP3. Similarly, this force can be decomposed into components F in the x and y directions.13_x and F 13_y The details are as follows:

[0091]

[0092] Where x3 and y3 are the coordinates of access point AP3, x1 and y1 are the coordinates of access point AP1, and d 13 This represents the distance between access point AP3 and access point AP1.

[0093] After obtaining the x and y components of all the relationships, the magnitude of the resultant force in each direction can be calculated. Figure 3 As shown, the x-axis components of both the attractive and repulsive forces are positive, the y-axis component of the attractive force is positive, and the y-axis component of the repulsive force is negative. Therefore, the resultant force along the x and y axes is as follows:

[0094]

[0095]

[0096] When the connection point is subjected to multiple forces, the forces are decomposed in the same way, and then the magnitude of the resultant force in each direction is calculated, as shown below:

[0097]

[0098] Among them, F q_x and F q_y These represent access points (APs). q The magnitude of the forces acting on the x-axis and y-axis; and These represent the hotspot locations relative to the access points (APs). q The magnitude of the attractive force F experiences along the x-axis and y-axis. qj_x and F qj_y These represent access points (APs). j Access Point (AP) q The magnitude of the repulsive force along the x-axis and y-axis. Since the repulsive force is... The opposite direction of the force is represented by a negative number. Note that during implementation, it is necessary to determine the direction of each force.

[0099] Finally, the resultant force needs to be converted into the actual distances that AP should move in the x and y directions. For AP q In terms of the net force acting on it, the magnitude is:

[0100]

[0101] The distance moved is directly proportional to the magnitude of the force. However, directly using the magnitude of the resultant force may lead to excessive movement, thus affecting the stability of the algorithm. Therefore, this paper introduces an exponential function to avoid this phenomenon and sets a threshold F to limit the movement. th Only when the component size is greater than the threshold will it affect the location of AP. Assume AP... q The initial coordinates are (x q ,y q The coordinates after the movement are (x′). q ,y′ q The distance traveled can be calculated using the following formula:

[0102]

[0103] Where, d max This represents the maximum travel distance of the AP node.

[0104] Step 7: For the third set of candidate solutions, calculate the objective value S for each candidate solution. t The process involves selecting the candidate solution with the largest objective value. This is then compared to the objective value of the best candidate solution from the previous cycle. If the objective value of this candidate solution is greater than that of the best candidate solution from the previous cycle, then this candidate solution becomes the new best candidate solution. The objective value S of this best candidate solution is recorded. t The position will become the initial position of the candidate solution in the next round of iterative search.

[0105] In this step, the objective value of the candidate solutions is calculated, and the candidate solution with the largest objective value is selected to obtain the current optimal candidate solution. The objective value S of the candidate solution is... t and location P t (X t ,Y t The expression is as follows:

[0106]

[0107] Where F(c) represents the c-degree coverage of the entire receiving area. F(c) represents the normalized average signal strength; α and β are the weighting values ​​for the c-degree coverage and the average signal strength. Since F(c) and The values ​​of both are within [0,1]. In a preferred embodiment, both are considered equally important in the optimization objective, so the weight coefficients are all set to 1. max I represents the optimal solution. best This indicates the location of the optimal solution, max(S) t ) indicates taking the maximum S t The candidate solutions corresponding to the values ​​and their access point locations.

[0108] Where F(c) is the c-degree coverage of the entire region, as shown below:

[0109]

[0110] Where M and N are the maximum number of rows and columns in the receiving point grid within the receiving area, respectively; f(c,i,j) indicates whether the receiving point at coordinates (i,j) satisfies c-degree coverage. c-degree coverage means that a receiving point at a certain location can receive signals from at least c different APs, and the strength of these signals is higher than the set receiving sensitivity R. min This coverage measurement method ensures that every location in the wireless network receives a sufficiently strong signal from at least c access points, thereby improving network redundancy and reliability, and significantly enhancing network performance and user experience. The criteria for this judgment are as follows:

[0111]

[0112] in The normalized average signal strength is shown below:

[0113]

[0114] Where, r ij r represents the signal strength value before normalization at (i,j); max r represents the maximum absolute value in the RSSI matrix; ij ′ represents the normalized signal strength value.

[0115] In a preferred embodiment, the receiving sensitivity R min The selection method is as follows: based on the modulation and coding scheme (MCS) index of the access area where the access point (AP) is located, the selection is performed by querying the pre-configured MCS index and the receiver sensitivity R. min The corresponding relationship is used to obtain the information, thus making it suitable for differentiated deployment scenarios.

[0116] Specifically, considering the varying network performance requirements of users in different regions, this invention flexibly sets receiver sensitivity for different regions based on the Modulation and Coding Scheme (MCS) index to achieve differentiated deployment. In the 802.11 standard, MCS is a crucial parameter specifying the modulation and coding scheme of wireless signals. It not only determines the data transmission rate and reliability but is also closely related to receiver sensitivity. Higher-order MCS indices result in higher transmission rates, but also require stronger received signal strength to achieve these higher rates. The frequency bandwidth of the WLAN network has a certain impact on receiver sensitivity; Table 1 shows the receiver sensitivity corresponding to different MCS indices under different bandwidths.

[0117] Table 1. Receiver Sensitivity Corresponding to MCS Index

[0118]

[0119] In differentiated deployment scenarios, high-end MCS solutions are needed for hotspot areas with high data transmission demands to improve data transmission efficiency and stability. For example, the MCS15 solution has a receiver sensitivity of -61dBm and can provide speeds up to 300Mbps. In contrast, low-end MCS solutions are used for ordinary areas with low data transmission demands to optimize AP deployment range. For example, the MCS8 solution has a receiver sensitivity of -79dBm and can provide speeds up to 30Mbps.

[0120] Step 8: Repeat steps 3 to 7 to iterate the candidate solution positions and obtain the next best candidate solution. Repeat this process to complete the deployment of the access point.

[0121] The iteration can be terminated when the preset maximum number of iterations (maxIter) is reached or when the maximum objective value of the candidate solution fluctuates less than the threshold (threshold).

[0122] Step 9: Perform redundancy detection on the AP access point layout obtained in Step 8 to obtain the pseudo-optimal layout of the WLAN network.

[0123] This step involves redundancy detection in the AP (Access Point) layout, aiming to meet coverage requirements while reducing deployment costs. The Pearson correlation coefficient is a widely used metric in statistics, primarily used to measure the degree of linear correlation between two variables. In a WLAN deployment scenario, the similarity between access points is analyzed by calculating the Pearson correlation vector to identify and optimize the deployment of redundant nodes.

[0124] The Pearson correlation coefficient is a statistic that measures the degree of linear correlation between two n-dimensional vectors. It is defined as the ratio of the product of the covariance of two variables X and Y to the product of their respective standard deviations, as shown below:

[0125]

[0126] Where, ρ XY Here, cov(X,Y) represents the Pearson correlation coefficient, and cov(X,Y) represents the covariance of variables X and Y; σ X and σ Y These represent the standard deviations of variables X and Y, respectively. The Pearson correlation coefficient typically ranges from -1 to 1, where 1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no correlation.

[0127] In redundancy detection for WLAN deployments, node redundancy is determined by comparing the Pearson correlation coefficients of the signal strength matrices of two access points (APs). In this scenario, calculating the Pearson correlation coefficients of different columns of the two signal strength matrices is meaningless because AP redundancy detection requires determining whether the coverage areas of the two APs are highly similar, i.e., whether their signal strengths are similar at the same location. Therefore, for APs X and Y, it is necessary to calculate the Pearson correlation coefficient ρ of the same column. XY It can be expressed mathematically as:

[0128]

[0129] in, The Pearson correlation coefficient represents the i-th column of AP nodes X and Y.

[0130] The implementation steps are as follows: First, calculate the mean value of the signal strength matrix for each access point, and then sort the access points in ascending order according to the mean value. Then, start with the access point (AP) with the poorest signal coverage. x Starting from this point, calculate the Pearson correlation coefficient vector with other access points sequentially. If the similarity coefficient is high, for example, above the similarity threshold ρ, then... th If this is the case, it indicates that redundant deployment may have occurred. At this point, it is still necessary to check whether network performance is affected after deleting the access point. Network performance can be assessed using c-degree coverage. If network performance does not change significantly, for example, if the change is less than a set value, then it indicates that redundant deployment has indeed occurred, and the node AP can be removed. x Remove the affected components to ensure that the network performance of the deployment plan is not impacted.

[0131] The present invention will be described below using specific data in Example 1.

[0132] This embodiment 1 proposes a deployment planning algorithm based on the correlation of predicted signals (VF-FOA for short), which includes the following steps:

[0133] The first step is to model the WLAN network scenario and initialize relevant parameters. The modeling results are as follows: Figure 4 As shown in Table 2, the simulation scene parameter configuration table is shown in Table 3, and the obstacle electromagnetic parameters are shown in Table 3.

[0134] Table 2 Simulation Scene Parameter Configuration Table

[0135]

[0136] Table 3 Electromagnetic Parameters

[0137]

[0138] The second step is to initialize the relevant parameters of the deployment planning algorithm, as shown in Table 4.

[0139] Table 4 Algorithm Parameter Configuration Table

[0140]

[0141] The third step is to randomly initialize the positions of the candidate solutions, obtaining the positions P of popSize candidate solutions. t (X t ,Y t The AP node of the t-th candidate solution is denoted as P. t,i (x t,i ,y t,i ), where t∈[1,popSize], i∈[1,N].

[0142] The fourth step is to iteratively optimize the candidate solutions by randomly searching in all directions.

[0143] The fifth step is to predict the signal strength at the receiving point based on the SBR / IM-based ray tracing algorithm.

[0144] The sixth step is to calculate the magnitude and direction of the repulsive and attractive forces acting on the AP nodes of each candidate solution based on the association relationships of AP deployment.

[0145] The seventh step is to decompose the forces acting on node AP, calculate the distance and direction of movement, and adjust the positions of candidate solutions.

[0146] Step 8: The objective value S of the candidate solution obtained in step S6 is... t Perform calculations and select the candidate solution with the largest objective value. Then, compare this candidate solution's objective value with the objective value of the best candidate solution from the previous cycle. If the objective value of this candidate solution is greater than that of the best candidate solution from the previous cycle, then this candidate solution becomes the new best candidate solution. Record the objective value S of this candidate solution. t and location P t (X t ,Y t This position will become the initial position of the candidate solution for the next round of iterative search.

[0147] Step 9: Iterate over the candidate solution positions, repeating steps 4 to 9 until the preset maximum number of iterations maxIter is reached or the maximum flavor concentration fluctuation of the candidate solution is less than the threshold threshold, then end the iteration.

[0148] Step 10: Perform redundancy detection on the AP access point layout obtained in step 9, identify and delete unnecessary AP access points.

[0149] This invention simulates the VF-FOA deployment planning algorithm with different numbers of access points (APs). To test the effectiveness of the VF-FOA algorithm in optimizing wireless LAN coverage, it is compared with three other algorithms: I-FOA (Fine-grained Fruit Fly Optimization), VFA (Virtual Force-based Node Deployment Algorithm), and PSO-DFOA (Particle Swarm Optimization-based Fruit Fly Optimization). The tests primarily focus on two key indicators: coverage rate and average signal strength, comparing results in overall areas and hotspot areas to observe the algorithm's effectiveness. Each algorithm underwent 10 independent simulation runs to obtain stable statistical data.

[0150] (1) Coverage Comparison

[0151] The overall regional coverage comparison results and the hotspot regional coverage comparison results are as follows: Figure 5 and Figure 6 As shown. By Figure 5 It is evident that as the number of access points (APs) increases from 8 to 16, the coverage of all four deployment optimization algorithms gradually increases with the increase in access points. When the number of nodes increases to 14 and 16, the I-FOA algorithm seems to be trapped in a local optimum, with its coverage ranking last among the four algorithms, while the VF-FOA and PSO-DFOA algorithms both achieve 100% coverage, i.e., complete coverage. Under the same conditions, the overall coverage of the I-FOA and VFA algorithms is significantly lower than that of the VF-FOA and PSO-DFOA algorithms, while the VF-FOA algorithm has a slightly higher coverage than the PSO-DFOA algorithm. Furthermore, observation... Figure 6 It can be observed that, under the same conditions, the VF-FOA algorithm significantly improves the coverage of hotspot areas. The PSO-DFOA and I-FOA algorithms do not differentiate their deployment in hotspot areas, resulting in insufficient coverage. While the VFA algorithm adjusts AP deployment locations to consider the coverage needs of hotspot areas and improves their coverage, its overall coverage remains low, and its global search capability is poor. Therefore, compared to other algorithms, VF-FOA significantly improves the coverage of hotspot areas while maintaining overall coverage, demonstrating strong differentiated deployment capabilities for specific regions.

[0152] (2) Network performance comparison

[0153] By comparing the average signal strength of the networks obtained from four deployment optimization algorithms, the network performance of the optimal deployment scheme obtained by the four algorithms can be determined. Average signal strength is an important indicator for measuring network performance, directly reflecting the quality and stability of signal transmission in the network. The simulation results comparing the average signal strength of the overall area and hotspot areas are as follows: Figure 7 and Figure 8 As shown.

[0154] Analysis and comparison show that, under the same conditions, the PSO-DFOA algorithm has the best global network performance, followed by the VF-FOA algorithm, while the I-FOA and VFA algorithms have relatively poor network performance. However, considering network performance in hotspot areas, the VF-FOA algorithm is the best, followed by the VFA algorithm. The VFA algorithm takes into account the impact of user distribution on AP deployment locations and makes corresponding optimizations, but its improvement in overall network performance is still limited and fails to achieve the desired effect. While the overall regional network performance of the VF-FOA algorithm is slightly inferior to that of the PSO-DFOA algorithm, it significantly improves network performance in hotspot areas, meeting the needs of users in these areas without significantly impacting overall performance, and can better adapt to users' differentiated network requirements. Other algorithms, because they did not impose additional limitations on the receiver sensitivity in hotspot areas during optimization, ultimately fail to meet user needs in these areas.

[0155] (3) Stability comparison

[0156] Statistical analysis was performed on the results of ten trials for each algorithm, and the results were plotted as follows: Figure 9 The box plots shown indicate that the VFA algorithm exhibits high stability, followed by the VF-FOA, PSO-DFOA, and I-FOA algorithms. In comparison, the I-FOA algorithm shows poor stability. The PSO-DFOA algorithm, optimized using particle swarm optimization, has improved stability to some extent, but still exhibits significant fluctuations. The VFA algorithm, not using a swarm-based approach, shows strong consistency in results across multiple runs, demonstrating good stability. The VF-FOA algorithm, optimized by constructing associations for AP deployment planning, improves stability and global search capabilities, showing strong consistency in results across multiple runs and effectively enhancing network performance.

[0157] (4) Redundancy detection

[0158] The tenth step was tested with different numbers of APs. Ten simulation experiments were conducted with the number of APs set to 8, 10, 12, 14, and 16 respectively. The changes in the number of nodes, network coverage, and average signal strength before and after running the tenth step are shown in Table 4.

[0159] Table 5. Performance changes of the deployment scheme before and after step 10.

[0160]

[0161] When the number of APs is set to 8, 10, or 12, the redundant access points in the deployment scheme are not identified in the tenth step, and the algorithm performance remains unchanged before and after running the tenth step. When the number of APs is set to 14 or 16, the redundant access points in the deployment scheme are identified in the tenth step, and the PCC-RDA algorithm deletes the redundant access points. After deletion, the coverage of the deployment scheme is not affected. However, due to the deletion of redundant access points, the maximum signal strength of some receiving locations decreases, which slightly reduces the average signal strength of the network, but does not have a significant impact on network performance.

[0162] Simulation results show that step ten can effectively identify redundant access points in the deployment scheme while ensuring overall network coverage. In practical applications, the number of APs can be reasonably reduced according to network requirements and deployment environment to balance the contradiction between network performance and deployment cost.

[0163] In summary, the near-optimal deployment scheme for the scenario was obtained. This scheme requires a minimum of 13 access points to ensure global network coverage. The AP deployment locations meet the set receiver sensitivity requirements for different areas. Signal strength in hotspot areas is greater than -61dBm, while signal strength in ordinary areas is greater than -79dBm. The signal coverage heatmap is shown below. Figure 10 As shown.

[0164] The specific embodiments described above only illustrate the design principles of the present invention. The shapes and names of the components in this description may differ and are not limited. Therefore, those skilled in the art can modify or make equivalent substitutions to the technical solutions described in the foregoing embodiments; and these modifications and substitutions do not depart from the inventive spirit and technical solutions of the present invention, and should all fall within the protection scope of the present invention.

Claims

1. A method for planning the deployment of wireless access points (APs) in a shipborne network, comprising candidate solutions composed of APs and iteratively optimizing the candidate solutions; characterized in that: During iterative optimization, the movement distance is constructed based on the attractiveness between the access point (AP) and the hotspot area, as well as the repulsive force between the APs, and the access point is moved accordingly. The c-degree coverage and average signal strength are used as optimization objectives to screen the best candidate solution. This process is repeated to complete the access point deployment. The calculation methods for the attractive force and the repulsive force are as follows: Set up access point (AP) q The attractive force experienced at the receiving point H for: Among them, AP q Defined as the q-th access point after sorting the signal strength of each access point AP at the receiving point H; k1 is the attraction coefficient, used to adjust the strength of the attraction. Access Point (AP) q Signal strength at receiving point H; R th The set receiving sensitivity; Indicates access point (AP) q Distance between the receiving point H and the receiving point H; Indicates access point (AP) q The direction of the force between the receiving point H and the receiving point H; when Less than the set receiving sensitivity R th At that time, the access point AP will be generated. q The attractive force pulling towards the receiving point H; Set up access point (AP) i With AP j Repulsive force between for: Where r represents the repulsive force coefficient; d th Represented as distance threshold; d ij Indicates access point (AP) i With AP j The Euclidean distance between them; α ij Indicates that it is from the access point AP i Pointing to AP j The direction, α ij +π represents the direction of the repulsive force; when the distance between two access points is less than the distance threshold d... th When this occurs, a repulsive force is generated.

2. The method as described in claim 1, characterized in that, The constructed movement distance, the movement of the access point is as follows: Let F be the x-axis component of the resultant force of the repulsive and attractive forces acting on the access point AP. q_x The y-axis component is F q_y Then, the distances Δx and Δy represented by exponential functions are: Among them, F q For F q_x and F q_y The resultant force, d max The maximum distance to move is set. The access point is moved using the moving distances Δx and Δy.

3. The method as described in claim 1, characterized in that, The method further includes: before applying the movement distance constructed based on attractive and repulsive forces to the access point, determining whether the xy-axis component of the resultant force of the repulsive and attractive forces is greater than a set limit movement threshold F. th If so, update the location of the access point (AP) by moving the distance; otherwise, keep the AP's location unchanged.

4. The method according to any one of claims 1-3, characterized in that, Before moving the access point of the candidate solution based on attraction and repulsion, a random move is also performed.

5. The method as described in claim 1, characterized in that, The optimization objective, which utilizes c-degree coverage and average signal strength, is as follows: Among them, S t Let F(c) be the objective value for optimizing the candidate solution, and F(c) be the c-degree coverage of the entire receiving area. α represents the normalized average signal strength; α and β are the importance weights. Where M and N are the maximum number of rows and columns of the receiving point grid within the receiving area, respectively; f(c,i,j) indicates whether the receiving point at coordinates (i,j) satisfies c-degree coverage, if the receiving point (i,j) can receive signals with a strength higher than the set receiving sensitivity R from at least c different access points (APs). min If the signal is c, then it is determined that the c-degree coverage is satisfied, and f(c,i,j)=1; otherwise, f(c,i,j)=0. Where, r ij r represents the signal strength value before normalization at (i,j); max r′ represents the maximum absolute value in the RSSI matrix. ij This is the normalized signal strength value.

6. The method as described in claim 5, characterized in that, The receiving sensitivity R min The method of obtaining it is: Based on the Modulation and Coding Strategy (MCS) index of the access area where the access point (AP) is located, the pre-configured MCS index and the receiver sensitivity R are queried. min We obtain the corresponding relationship.

7. The method as described in claim 1, characterized in that, The method for selecting the best candidate solution is as follows: select the candidate solution whose optimization objective value is the largest in this round and is greater than the largest optimization objective value in the previous round, and use it as the initial position of the candidate solution for the next round of iteration, and repeat the process.

8. The method as described in claim 1, characterized in that, The signal strength at the access point is obtained by using a ray tracing algorithm based on SBR / IM to predict the signal strength at the receiving point.

9. The method as described in claim 1, characterized in that, After the iterative optimization is completed, redundancy detection and optimization are further performed on the access point (AP) layout corresponding to the iterative optimization results, specifically as follows: Calculate the average signal strength of each access point and sort the access points in ascending order according to the average value; then start from the access point with the poorest signal coverage. x Starting from this point, the Pearson correlation coefficient vectors with other access points are calculated sequentially; if the similarity is higher than a set value, further verification is performed to delete the access point AP. x Whether the network performance is affected afterward, if the impact is less than the set value, then it is confirmed that redundant deployment has occurred, and only then can the access point AP be removed. x delete.

Images