A signal planning method, device and medium based on pre-deployment and fine-tuning strategy

Abstract

The present invention discloses a signal planning method, device and medium based on a pre-deployment and fine-tuning strategy, and relates to the technical field of wireless communication. In the pre-deployment stage, the signal overlapping area between all wireless access points in a target detection area is obtained, the minimum area of ​​the signal overlapping area between the wireless access points is defined as a first optimization target, the positions of the wireless access points are globally adjusted, the optimal position with the minimum overlapping area is obtained, the overlapping area between the wireless access points is minimized, and on the basis of the optimal position, the wall loss and the regional priority are considered, and the second optimization target is defined to search for the precise position. Based on the second optimization target, the optimal position is adjusted in combination with the quasi-Newton method L-BFGS to obtain the final coordinates of the wireless access point, so that the deployment position of the wireless access point can be determined more accurately, the most advanced signal quality is achieved, a high signal coverage rate is realized, the signal coverage dead zone is reduced, and the deployment efficiency is ensured.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a signal planning method, device and medium based on pre-deployment and fine-tuning strategy. Background Art

[0002] The deployment of access points (APs) is crucial to ensure wireless signal coverage in local areas. Wi-Fi-based technologies, especially the currently popular Wi-Fi 6 standard, have become an important way to access the Internet, with a maximum rate of up to 9.6 Gbps. For 60 GHz millimeter wave (mmWave) and terahertz (THz) wireless networks, although these networks have potential for applications with high bandwidth requirements in indoor scenarios, they have not yet been widely used due to their poor propagation characteristics. To this end, various strategies for deploying access points have been proposed to adapt to the needs of these networks. Therefore, how to provide an intelligent deployment method to achieve high signal quality has become a research topic in Wi-Fi network optimization.

[0003] Wi-Fi deployment refers to placing a certain number of candidate APs in suitable locations to maximize signal coverage, avoid signal dead spots, and ensure average signal strength. This work is quite challenging because it relies on the movement patterns of multiple wireless APs. Related research work includes: 1) Key attenuation factors of signal propagation. Signal attenuation caused by signal propagation determines the candidate locations of APs. Line-of-sight (LoS) or non-line-of-sight (NLoS) conditions are also key factors for signal propagation. These key factors should be considered when deploying and planning APs to ensure that signal coverage is free of dead spots. 2) Deployment methods of high-frequency wireless local area networks (WLANs). Compared with low-frequency band networks such as millimeter wave and terahertz networks, high-frequency communications can provide higher data transmission rates and capacity. Terahertz communications also have great potential in using ultra-wide bandwidths. 3) Deployment methods of Wi-Fi. Due to the short propagation distance of millimeter waves and terahertz, Wi-Fi technology is widely used in indoor and outdoor wireless applications such as IoT systems, healthcare, robotic systems, and smart vehicles. In recent years, APs supporting Wi-Fi communication have been deployed in wireless LANs to achieve good system performance.

[0004] In order to achieve effective Wi-Fi coverage, the impact of irregular coverage of Wi-Fi signals caused by obstacles inside buildings must be considered. To ensure that each area can obtain sufficient wireless signal coverage, the following points become the key points of research: 1) Uneven signal coverage is caused by factors such as walls, dense population and obstacle reflection. The optimal AP position can be calculated through the signal propagation formula to improve deployment efficiency and accuracy. 2) In areas with dense population or high signal strength requirements, it is important to ensure sufficient signal strength, and the number of APs may need to be increased to meet user wireless service needs. 3) Attention should be paid to the irregular coverage of signals caused by walls to avoid signal blind spots and ensure network stability and availability.

[0005] When considering wireless AP deployment, existing methods limit signal strength and coverage due to the discreteness of candidate locations. 1) First, AP deployment research in Wi-Fi environments often uses methods that randomly move AP positions, but such methods cannot accurately find the best position, which limits the optimization of signal strength and coverage, and the lack of coordinated movement between APs makes it difficult to place all APs in the best position at the same time. The accuracy of such methods depends only on the size of the mobile AP step. 2) Secondly, the quality of wireless network signals is affected by many factors, such as room shape, area, obstacle location, etc., and these factors need to be considered comprehensively during research.

[0006] To solve the above problems, two key research issues need to be addressed: 1) How to deploy Wi-Fi with a limited number of APs to minimize the AP signal overlap area while considering key factors; 2) How to deploy with sufficient APs to maximize the weakest signal strength. The research on these two issues is crucial to the development of efficient wireless network deployment strategies. Summary of the invention

[0007] The technical problem to be solved by the present invention is that the factors considered are single, the optimal position cannot be accurately found, and the optimization of signal strength and coverage is limited. The purpose is to provide a signal planning method, device and medium based on pre-deployment and fine-tuning strategies, which globally adjust the position of the wireless access point in the pre-deployment stage to obtain the optimal position with the smallest overlapping area, minimize the overlapping area between the wireless access points, and on the basis of the optimal position, consider the wall loss and regional priority, minimize the sum of the inverse of RSS to search for the precise position, and can more accurately determine the deployment position of the wireless access point, achieve the most advanced signal quality, realize high signal coverage, reduce the signal coverage dead zone, and ensure deployment efficiency.

[0008] The present invention is achieved through the following technical solutions:

[0009] A first aspect of the present invention provides a signal planning method based on a pre-deployment and fine-tuning strategy, comprising the following specific steps:

[0010] Obtain the signal overlap area between all wireless access points in the target detection area;

[0011] Divide the target detection area into grids, define the minimum signal overlap area between wireless access points as the first optimization goal, adjust the position of the wireless access point according to the first optimization goal, and obtain the position of the wireless access point with the minimum signal overlap area as the optimal position;

[0012] Calculate the signal value of the wireless access point at the best position, and define the maximization of the sum of the signal values ​​of all grid points as the second optimization goal;

[0013] Based on the second optimization objective, the optimal position is adjusted in combination with the quasi-Newton method L-BFGS to obtain the final coordinates of the wireless access point.

[0014] Furthermore, the obtaining of the signal overlap area between all wireless access points in the target detection area specifically includes:

[0015] The signal coverage radius of the current wireless access point without considering the attenuation of obstacles is obtained according to the signal strength, and the signal coverage area of ​​all wireless access points in the target area is obtained in combination with the position of the wireless access point;

[0016] The signal overlap area between all wireless access points in the target detection area is determined based on the signal coverage area.

[0017] Further, the defining of minimizing the signal overlap area between wireless access points as the first optimization goal specifically includes:

[0018] ;

[0019] ;

[0020] in, Indicates i Wireless access points and j Wireless access points The overlapping area between q Indicates that all wireless access points are x and y The position vector on the axis, r represents the signal radius vector, Indicates i Wireless access points The signal radius, Indicates i Wireless access points The signal radius, k Indicates the total number of wireless access points. Indicatesi Wireless access points and j Wireless access points The Euclidean distance between them in a two-dimensional plane.

[0021] Further, the adjusting the position of the wireless access point according to the first optimization target specifically includes:

[0022] Initialize the grid, randomly select a wireless access point, obtain the grid where the selected wireless access point is located, and determine whether the grid point is in the prohibited area:

[0023] If it is in the prohibited area, the current grid point is re-obtained to determine whether it is in the prohibited area;

[0024] If it is not in the prohibited area, the next grid point of the current grid point is obtained. If the current grid point is in the allowed placement area, the candidate position is obtained, the overlapping area is initialized, and the position of the current wireless access point is updated;

[0025] The position of the wireless access point is repeatedly adjusted until the area of ​​the signal overlap region reaches a set threshold, and the position of the wireless access point with the smallest signal overlap region is obtained as the optimal position.

[0026] Furthermore, the initialization of the overlapping area specifically includes:

[0027] The overlapping area between the wireless access points is calculated traversally. If the overlapping area between the wireless access points calculated this time is less than the threshold of the compared overlapping area, the minimum overlapping area value is updated.

[0028] Furthermore, the step of calculating the signal value of the wireless access point at the best position specifically includes:

[0029] Calculate the i Signal value of wireless access point:

[0030] ;

[0031] According to i The signal value of the wireless access point is used to calculate the maximum signal value at the target location:

[0032] ;

[0033] in, Indicates that the point ( x , y , z ) i The RSS value of each wireless access point, ( x , y ,z ) represents the position of the point considered when calculating the RSS value, Indicates i The power of each wireless access point, C The parameter indicating the basis for adjusting the RSS value. Indicates i The signal from the wireless access point passes through the j Signal loss when facing a wall, Indicates i The signal of the wireless access point passes through n Total signal loss when facing a wall, Indicates i The signal wavelength of each wireless access point, θ represents the signal path loss coefficient, i Wireless access points The position is defined as .

[0034] Furthermore, the definition of maximizing the sum of the signal values ​​of all grid points as the second optimization objective specifically includes:

[0035] ;

[0036] ;

[0037] in, Indicates priority Φ is the regional priority, is the location ( x , y , z ) signal value, indicating that at position ( x , y , z ), given the network configuration q The signal strength at and Respectively represent the number of grids in the length and width directions of the indoor map, Indicates that the function is found Get the minimum value The value of .

[0038] Furthermore, based on the second optimization objective, the optimal position is adjusted in combination with the quasi-Newton method L-BFGS to obtain the final coordinates of the wireless access point, specifically including:

[0039] Initialize the optimal coordinates, get the initial coordinates and a positive number as the threshold of the gradient norm;

[0040] In the preset iteration rounds, the initial coordinates are updated, and the gradient of the current updated initial coordinates is calculated in each iteration round;

[0041] The gradient of the currently updated initial coordinates is judged: if the norm of the gradient is not greater than the threshold of the gradient norm, the iteration is stopped and the updated initial coordinates are returned as the optimal solution; if the norm of the gradient is greater than the threshold of the gradient norm, the search direction of the gradient is calculated, and the step size is determined by a straight line search;

[0042] Get the inverse of the updated Hessian matrix in the current iteration according to the difference of the gradient, the initial coordinates, and the difference of the coordinates of the t-th iteration; recalculate the gradient based on the inverse of the updated Hessian matrix and update the gradient.

[0043] A second aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a signal planning method based on a pre-deployment and fine-tuning strategy is implemented.

[0044] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a signal planning method based on a pre-deployment and fine-tuning strategy.

[0045] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0046] 1. A new algorithm is proposed for deploying a given number of wireless access points. The algorithm is not limited by the traditional step size and allows searching for wireless access points at any location, thereby achieving more accurate location selection. Compared with previous studies, the new algorithm focuses on the precise location of wireless access points rather than discrete locations for the first time while ensuring deployment efficiency, avoiding inaccurate deployment determined by the step size, while also providing first-class signal quality, high signal coverage, and low dead zone rate;

[0047] 2. An optimization objective is designed to minimize the sum of the inverses of the priority signal strengths at all grid points, thereby improving the overall signal quality;

[0048] 3. In response to different signal coverage requirements in complex office scenarios, the invention proposes a regional priority strategy to adapt to the specific signal coverage requirements of different areas;

[0049] 4. Consider key factors such as three-dimensional distance loss, wall loss and wireless power in signal propagation calculations to obtain more accurate measurement results. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0051] Figure 1 A main process for signal planning in an embodiment of the present invention;

[0052] Figure 2 A signal planning method in an embodiment of the present invention;

[0053] Figure 3 is an indoor map to be deployed in an embodiment of the present invention;

[0054] Figure 4 is the change of the average RSS in the pre-deployment stage in the embodiment of the present invention;

[0055] Figure 5 is the AP deployment result in the pre-deployment stage in the embodiment of the present invention;

[0056] Figure 6 is the change of the average RSS in the fine-tuning stage in the embodiment of the present invention;

[0057] Figure 7 is a heat map of deployed APs in an embodiment of the present invention;

[0058] Figure 8 The following are deployment results of different numbers of APs in the embodiments of the present invention;

[0059] Fig. 9 is the RSS of different grid sizes in the embodiment of the present invention;

[0060] Fig.10 The following are deployment results compared with different strategies in the embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0062] As a possible implementation, Figure 1 and Figure 2As shown, this embodiment provides a signal planning method based on pre-deployment and fine-tuning strategy, including the following specific steps: obtaining the signal overlap area between all wireless access points in the target detection area; dividing the target detection area into grids, defining the minimum signal overlap area between wireless access points as the first optimization goal, adjusting the position of the wireless access point according to the first optimization goal, and obtaining the position of the wireless access point with the minimum signal overlap area as the optimal position; calculating the signal value of the wireless access point at the optimal position, defining the maximum sum of the signal values ​​of all grid points as the second optimization goal; based on the second optimization goal, combining the quasi-Newton method L-BFGS to adjust the optimal position, and obtaining the final coordinates of the wireless access point. By globally adjusting the position of the wireless access point in the pre-deployment stage, the optimal position with the minimum overlap area is obtained, and the overlap area between the wireless access points is minimized. On the basis of the optimal position, the wall loss and the regional priority are considered, and the sum of the inverses of the RSS is minimized to search for the precise position, so that the deployment position of the wireless access point can be determined more accurately, the most advanced signal quality is achieved, high signal coverage is achieved, the signal coverage dead zone is reduced, and the deployment efficiency is ensured.

[0063] In this embodiment, it specifically consists of three steps, including: algorithm and data definition, pre-deployment planning and fine-tuning planning.

[0064] S1. Algorithm and data initialization.

[0065] S101: Define indoor grid. Assume that the number of deployed wireless access points is n , the length and width of the indoor map are l Mihe w meters, indoor area is l × w， Given a grid size of g meters, Indicates the length of the room. Indicates the width of the room. Indicates the size or interval of the segmentation, and is evenly marked on the indoor map. grid points, where and , and Respectively represent the number of grids in the length and width directions of the indoor map.

[0066] S102: Define signal value. a represents the AP vector, express signal coverage area; p represents the position vector of the AP, express The positions are Coordinates on the x, y, and z axes. and is randomly generated. Assume that the AP is located on the ceiling at a height of 3 meters. is set to a constant, i.e. the height of the ceiling is 3 meters. It represents the signal coverage radius without considering the attenuation caused by obstacles. express At point ( x , y , z )’s signal strength RSS, z is the height of the signal measurement point from the ground. Assume that the signal measurement point is on the ground, and z Set it as a constant, i.e. 0 meters; q Represents the position vectors of all APs on the x-axis and y-axis, r Represents the signal radius vector.

[0067] S103: Define the indoor map area. s(q,r) represents the area covered by the Wi-Fi signal, u(q,r) represents the uncovered area, and s represents the sum of s(q,r) and u(q,r), which is the total coverage area of ​​a certain area. Therefore:

[0068] s=s(q,r)+u(q,r) ;

[0069] S104: Define coverage area. Let s(q,r) represent the sum of coverage areas of all APs. Subtract the sum of the overlapping areas between the two, l(q,r) represents the sum of all overlapping areas, then:

[0070] ;

[0071] S105: Define the first optimization goal. Maximizing the signal coverage area is equivalent to minimizing the overlap area. Therefore, our first optimization goal seeks to minimize the overlap area, that is:

[0072] ;

[0073] ;

[0074] in, Indicates i Wireless access points and j Wireless access points The overlapping area between q Indicates that all wireless access points are x and y The position vector on the axis, r represents the signal radius vector, Indicatesi Wireless access points The signal radius, Indicates i Wireless access points The signal radius, k Indicates the total number of wireless access points. Indicates i Wireless access points and j Wireless access points The Euclidean distance between them in a two-dimensional plane.

[0075] S2: Pre-deployment planning. Pre-deployment planning adjusts the positions of APs globally. First, the grid points are initialized according to the grid size. Second, the initial position of each AP is taken as the optimal position. For each AP on the grid point, the sum of the overlapping areas is calculated to obtain the optimal position with the smallest overlapping area. Finally, a new deployment position is formed.

[0076] S201: Initialize the grid. According to step S101, the target detection area is initialized and divided into a number of grid points.

[0077] S202: Plan AP locations. n For each AP, calculate the best position of each AP according to steps S203-S208:

[0078] S203: Initialize the best AP position.

[0079] S204: Obtain the current first grid point from the grid point set;

[0080] S205: Determine the prohibited area. If the current grid point is in the prohibited area, repeat S204 and take the next grid point of the current grid point. If the current grid point is in the allowed placement area, proceed to S206.

[0081] S206: Obtain candidate positions;

[0082] S207: Initialize the overlapping area value;

[0083] Calculate the intersection area between APs in sequence;

[0084] If the intersection area obtained this time is less than the minimum known intersection area, then:

[0085] Update the minimum intersection area and the best AP position;

[0086] S208: Update the position of the current AP, and repeat S202 until the best position of each AP is calculated.

[0087] S209: Under the set conditions, repeat S202-S208, and take the one with the smallest overlapping area as the best planning position.

[0088] S3, fine-tuning planning. The purpose of fine-tuning planning is to find the precise location of APs based on pre-deployment.

[0089] S301: Calculate the signal value. This algorithm calculates the signal strength of the two-dimensional plane area on the ground. According to the grid point ( x , y , z )and The Euclidean distance between the points ( x , y , z )of RSS value :

[0090] ;

[0091] in, Indicates that the point ( x , y , z ) i The RSS value of each wireless access point, ( x , y , z ) represents the position of the point considered when calculating the RSS value, Indicates i Wireless access points The power, C The parameter indicating the basis for adjusting the RSS value. Indicates i The signal from the wireless access point passes through the j Signal loss when facing a wall, Indicates i The signal of the wireless access point passes through n Total signal loss when facing a wall, Indicates i The signal wavelength of each wireless access point, θ represents the signal path loss coefficient, i Wireless access points The position is defined as .

[0092] S302: Calculate the maximum value of the current position signal. For any position ( x , y , z ), signal value Take the maximum signal value of all APs at the current location, is the maximum value of the current position signal, and its calculation formula is:

[0093] ;

[0094] S303: Define the second optimization objective. To maximize the sum of the signal values ​​of all grid points, the position of each AP must be adjusted to obtain each point ( x , y , z ) In other words, Therefore, the goal of AP planning is to minimize the The sum of , the formula is:

[0095] ;

[0096] in, Indicates priority Φ is the regional priority, is the location ( x , y , z ) signal value, indicating that at position ( x , y , z ), given the network configuration q The signal strength at and Respectively represent the number of grids in the length and width directions of the indoor map. We set the second optimization goal as follows:

[0097] ;

[0098] in, Indicates that the function is found Get the minimum value The value of .

[0099] S304: Use the quasi-Newton method L-BFGS to find the exact coordinates. The quasi-Newton method belongs to the unconstrained objective optimization method using first-order and second-order derivatives. The basic idea is that in each iteration, the search direction is updated with the Newton direction. The initial coordinates calculated in the pre-deployment phase , Newton's method iteratively updates q in the following way in the preset round T:

[0100] ;

[0101] in, t represents the number of iterations, t ∈[1, T ], represents the t-th position, represents the t-1th position, Indicated in coordinates Department The gradient vector of Indicated in Function The Hessian matrix of the function The second-order derivative matrix of Indicated in Function The inverse matrix of the Hessian matrix;

[0102] ;

[0103] in, Representation function exist The gradient at Representation function exist The gradient at

[0104] However, calculation It will consume a lot of computing resources because its time complexity is The quasi-Newton method is generally used , through the formula Iterative Updates Approach , rather than directly calculating , as shown below:

[0105] ;

[0106] ;

[0107] ;

[0108] S3041: Initialization , from the starting point of the optimization process Start, and the threshold ;

[0109] S3042: Calculate gradient .if , then execute step 3, if not, end the calculation and return ;

[0110] S3043: Calculate gradient , and find the step size by linear search ;

[0111] ;

[0112] S3044: Passed , and calculate , then go to step S3042;

[0113] In the above steps, It means that the time complexity of calculating the inverse of the Hessian matrix is ​​cubic. Representation function The inverse matrix of the Hessian matrix is, represents the inverse of the initial approximate Hessian matrix, usually set to the identity matrix, represents the inverse of the approximate Hessian matrix at the tth iteration, and Respectively represent the position and gradient differences at the tth iteration, and For update , Indicated in t The inverse of the approximate Hessian matrix at +1 iterations, Indicated in t The parameters are changed from Update to The amount of change, Indicates that at t t The function at the iteration exist and The gradient difference of the point, represents the search direction at step t, Indicates the current iteration point. represents the step length at step t, represents the step size, determined by straight line search, Represents the threshold of the gradient norm, that is, the approximate inverse Hessian matrix is ​​updated iteratively and search direction , gradually approximating the function The minimum value of Representation function exist The gradient at .

[0114] As a possible implementation method, this embodiment deploys different numbers of wireless access points and meets different signal coverage requirements through experiments based on a real geographic background map. The results show that the invented algorithm (PDFT) achieves the following results: it reaches the most advanced signal quality, achieves high signal coverage, reduces signal coverage dead zones, and ensures deployment efficiency.

[0115] As a possible implementation, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a signal planning method based on a pre-deployment and fine-tuning strategy is implemented.

[0116] As a possible implementation, this embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, a signal planning method based on a pre-deployment and fine-tuning strategy is implemented.

[0117] As a possible implementation method, in order to further verify the research and development functions and results, the following Figure 3 The practice base,an indoor map with a size of 67.1m×22.6m,is used to evaluate the performance of the proposed algorithm.,In order to maximize the sum of the signal values ​​of all grid points,,we adopted an optimization algorithm to search for the best AP location.,We developed a simulation environment.

[0118] There are 7 rooms in the indoor map, including 2 large rooms and 1 small room. We assume that APs can only be placed in rooms, not in corridors. Therefore, the regional priority Φ of rooms 2-5 is set to 2, and the regional priority Φ of corridors is set to 0.1. To check the effectiveness of regional priorities, the smallest room 1 is used as a high-priority deployment area to ensure good signal quality. The Φ of room 1 is set to 300, which is higher than other rooms. Since the area of ​​rooms 6-7 is larger than other rooms and may require more signal coverage, the Φ of rooms 6-7 is set to 10.

[0119] like Figure 4 As shown in Figure 1, the change of average RSS in the pre-deployment stage is described. The focus of this stage is to adjust the global position of APs without considering wall loss and regional priority. The optimization goal of the pre-deployment strategy is to minimize the signal overlap area between APs, thereby maximizing the signal coverage of a single AP. In the case of redundant APs, the signal coverage strength of the AP is maximized. Before the pre-deployment algorithm, all APs were randomly assigned positions on the indoor map. The average RSS value of Room 1 was 48.87dBm, Rooms 2-5 was -42.66dBm, Rooms 6-7 was -46.72dBm, and the corridor was -46.98dBm. The average RSS of Room 1 first reached the maximum value of -40.40dBm after the first round of iterations. After the second round of iterations, the average RSS of Rooms 2-5, Rooms 6-7, and the corridor reached -41.86dBm, -40.82dBm, and -48.7dBm, respectively. It can be concluded that our pre-deployment strategy can quickly reach stability after the second round of iterations.

[0120] like Figure 5 The figure shows the deployment results using the pre-deployment strategy. It can be concluded that the coverage area of ​​a single AP is the largest when the signal overlap area is the smallest. Without considering wall loss and area priority, the AP position is globally optimized.

[0121] In order to take into account wall loss and area priority, it is necessary to fine-tune the position based on the output of the pre-deployment phase. Our fine-tuning algorithm adjusts the AP position in the local area so that R x,y In other words, our optimization goal is to minimize R x,y In the simulation, Φ of Room 1 is 300, which is the largest coefficient of all rooms, Φ of Rooms 6-7 is 10, and Φ of Rooms 2-5 is 2. All rooms are bonus areas that need better signal coverage. The corridor has a Φ of 0.1, which is a restricted area in the simulation environment.

[0122] like Figure 6 As shown in the figure, the change of the average RSS in the fine-tuning stage is described. First, the fine-tuning algorithm adjusts the AP position to give priority to the signal distribution of Room 1. After the 4th iteration, the average RSS rate of Room 1 first reaches the maximum value of -39.29dBm. Secondly, the fine-tuning algorithm will meet the signal coverage requirements of Rooms 6-7, reaching the maximum value of -40.21dBm after the 8th iteration. At the same time, the average RSS of Rooms 2-5 reaches the maximum value of -41.63dBm after the 8th iteration. Finally, the corridor reaches the minimum value of -50.35dBm after the 10th epoch. It can be concluded that the fine-tuning algorithm makes the RSS of the reward area gradually increase at each time point, while the RSS of the prohibited area gradually decreases.

[0123] from Figure 4-6 We can draw the following conclusions.

[0124] 1) The pre-deployment strategy can adjust the global location of APs well. However, it cannot achieve the best signal coverage because it does not consider wall loss and area priority.

[0125] 2) The fine-tuning strategy can adjust the position locally and also support the signal coverage requirements of Φ in different areas.

[0126] 3) The two strategies work well together.

[0127] like Figure 7As shown, the maximum signal strength of each point for all APs in the indoor simulation can be clearly displayed. To further highlight the quality of our algorithm, we calculated the RSS after the final deployment. The average RSS in the indoor map is -40.95dBm, and the signal coverage is 100%. The average RSS in room 1 is -39.29dBm, -41.63dBm in rooms 2-5, -40.21dBm in rooms 6-7, and -50.35dBm in the corridor. Obviously, at all indoor grid points, the RSS meets the minimum signal strength requirement of -60dBm. The coverage rate reaches 100%.

[0128] Room 1 has the smallest area among all the rooms, but its coverage priority is higher, with a regional priority Φ of 300, so our method automatically deploys an AP in room 1. The regional priority Φ of rooms 6-7 is 10, which is higher than the regional priority Φ of rooms 2-5, so we deploy two APs here to ensure sufficient signal coverage. In the corridor, the regional priority Φ is 0.1, and finally we can only rely on the APs in the room to achieve signal coverage. Obviously, the method proposed in this paper can automatically deploy APs according to the signal coverage requirements. Compared with previous studies, our algorithm is more intelligent and can meet the requirements of practical applications.

[0129] Our approach also includes an accurate algorithm to calculate the 3D signal propagation and use the 3D propagation distance in the 2D signal plane. This is an important feature that is often overlooked in previous studies. In indoor environments, path loss depends on distance and physical obstacles. In our simulations, the farther away from the AP, the lower the RSS value, and the signal is significantly attenuated after passing through walls.

[0130] We observed the following points from the results.

[0131] 1)RSS decreases as the signal propagation distance increases.

[0132] 2) PDFT can effectively deploy APs according to signal coverage requirements.

[0133] 3) High signal coverage, reaching 100% in a simulated environment.

[0134] Therefore, the effect of the algorithm is shown in Table 1 and Table 2 below:

[0135] Table 1 Signal strength performance comparison;

[0136]

[0137] Table 1 compares our algorithm with two popular algorithms on RSS. To further validate our algorithm, different numbers of APs are selected in the experiment to evaluate its deployment performance. The ZOA algorithm and the FOA algorithm generate random positions for APs at each iteration and then calculate the R x,y If the value is less than the previous value, the current position is assigned for deployment in this iteration. Since the position of the AP is randomly generated, it is difficult to place each AP in the correct position, and the accuracy of the ZOA and FOA methods is low. Our algorithm first sets the optimization object with a minimization function, and then finds the best position based on the gradient according to the minimized objective function. It can find the best deployment position of the AP within the iteration range, and the accuracy is better than the algorithm based on random positions.

[0138] Based on these results, we can conclude that:

[0139] 1) Since the zone priority Φ of Room 1 is 300, which is higher than other rooms, the signal strength requirement must be much higher than that of other rooms. Compared with previous work, our algorithm is able to identify deployment priorities. The average RSS of our algorithm for 3 APs, 6 APs, and 9 APs in Room 1 is -39.30dBm, -39.29dBm, and -39.29dBm, respectively, which is better than the ZOA algorithm and the FOA algorithm. Especially in the case of 3 APs, only our algorithm can give priority to meeting the signal coverage requirements of the high-priority room 1.

[0140] 2) In the simulation, APs are expected to be deployed in rooms, not corridors. The average RSS of our algorithm in rooms (1-7) are -50.43dBm, -44.04dBm, and -41.14dBm, respectively, which are larger than those of other algorithms; the average RSS in corridors are -52.12dBm, -51.33dBm, and -50.39dBm, respectively, which are smaller than those of other algorithms. This shows that the proposed work is able to maximize the average signal strength in rooms with high regional priority and minimize the average signal strength in corridors with regional priority.

[0141] 3) The average RSS of Room 1, Rooms 2-5, and Rooms 6-7 are all larger than those of other algorithms, and the average RSS of the corridor is smaller than that of other algorithms, indicating that when there are a large number of APs, our algorithm has the strongest fine-tuning ability and can adjust each AP to the most appropriate position to achieve the best signal coverage performance.

[0142] Table 2 Comparison of average signal coverage;

[0143]

[0144] Table 2 compares the average signal coverage and average RSS of five identical experiments. Coverage is an important indicator for evaluating the performance of AP deployment algorithms. If the number of APs is small, it is easy for APs to achieve effective signal coverage in any room, so it is not meaningful to compare the signal coverage of APs with a small number of APs. Since there are 7 rooms on the map, it is best to place an AP in each room considering the wall attenuation factor. In addition, the location of APs in each room is also a key factor affecting the signal coverage. In order to achieve differentiated signal coverage for all algorithms, we reduced the AP power to verify whether obvious AP signal coverage can be achieved. Therefore, the number of APs deployed in the simulation environment is set to 7, the AP power is uniformly set to 10dBm, and the effective threshold of signal strength is set to -50dBm. In order to verify whether the pre-deployment strategy is effective, we added three groups of experiments, one group is a combination of PD and ZOA, one group is a combination of PD and FOA, and the last group is only using the FT strategy. In order to verify the stability of the algorithm, we conducted five identical experiments and used the standard deviation (STD) to verify the stability of the algorithm.

[0145] We can observe the following points from the results.

[0146] 1) In room 1 with higher priority, the average signal coverage of PDFT is 100% and the average RSS is -43.29dBm. To meet the signal coverage of high-priority areas, the performance of PDFT is better than that of ZOA and FOA methods.

[0147] 2) In all rooms 1-7, the average signal coverage and average RSS of PDFT are better than those of ZOA and FOA methods.

[0148] 3) In most cases, the STD of PDFT is 0, so it is more stable than ZOA and FOA, and can help AP always converge stably on the same candidate position.

[0149] 4) The PD strategy cannot be used well with ZOA and FOA. Due to the randomness of the position movement of the swarm intelligence algorithm, it cannot significantly improve their signal coverage and RSS.

[0150] 5) Without the PD strategy, the performance of the FT strategy is also poor.

[0151] 1. Parameter Analysis

[0152] Figure 8 We describe the deployment of indoor maps by inputting different numbers of APs from 1 to 9. In practical applications, users usually decide how many APs to deploy based on indoor signal coverage quality requirements and economic budget, so the algorithm must be able to adapt to changes in the number of APs. We also analyze the impact of the number of APs. Figure 8The numbers (a) and (b) in the figure show that PDFT first deploys APs in the high coverage priority area. The signal coverage priority of Room 1 with a regional priority of Φ300 and Rooms 6-7 with a regional priority of Φ10 is higher than that of other rooms, so the algorithm prioritizes ensuring basic signal coverage in these areas. On the basis of ensuring coverage of Room 1 and Rooms 6-7, Figure 8 The numbers (c)-(h) describe how to deploy APs in rooms 2 to 5. Since the area of ​​room 6-7 is the largest, the algorithm will give priority to multi-AP coverage of room 6-7, while there are some redundant APs. Figure 8 The number (f)-7=(i) indicates that two wireless APs are deployed in room (6) and room (7).

[0153] Our deployment algorithm is based on the zone priority Φ. APs are deployed first in areas with higher Φ, then in areas with lower Φ, and no APs are deployed in prohibited areas. In addition, the algorithm can automatically detect the size of the area. If there are redundant APs, more APs are deployed in larger rooms.

[0154] Fig. 9 The change of average RSS in rooms with different grid sizes is shown. Since the AP movement direction of the ZOA algorithm and FOA algorithm is random, there is a large uncertainty in their results. Therefore, we take the average of the five experimental results as comparison data to observe the performance of different algorithms.

[0155] We can observe the following points from the results.

[0156] 1) In terms of the average RSS of all rooms under different grid sizes, the PDFT algorithm outperforms other algorithms.

[0157] 2) Even if we take the average of multiple experimental values ​​of the ZOA algorithm and the FOA algorithm, we can still see that the swarm intelligence algorithm is not very stable, especially for the average RSS of room 1, which varies greatly with the grid size. On the other hand, the signal value of the PDFT algorithm does not change much with the grid size.

[0158] 3) As the grid size increases, the running time of the algorithm decreases significantly, but the PDFT algorithm is relatively stable, so in practical applications, we can choose a larger grid size when selecting to obtain the maximum average signal strength with the shortest running time.

[0159] 2. Ablation Study

[0160] Fig.10The number (a) shows the deployment results using only the fine-tuning strategy, and the number (b) shows the deployment results using both strategies. The fine-tuning strategy is suitable for moving the AP position in a local area according to the optimization goal. If the random initial position does not meet the global optimum, a large deviation will occur. In order to better understand the pre-deployment strategy and fine-tuning strategy designed in the algorithm, we further studied the impact of the pre-deployment strategy on the algorithm performance.

[0161] Without the pre-deployment strategy, the fine-tuning strategy cannot move the initial position of the AP from the corridor to the room, resulting in the AP being deployed in some restricted areas. Although the wall loss and reward factor are set, the fine-tuning algorithm can only move the AP locally due to the uneven distribution of the initial positions, resulting in one AP being deployed in the corridor, two APs being deployed in room (5), and no APs being deployed in rooms (3) and (4). After adding the pre-deployment strategy, the fine-tuning strategy can well deploy the APs to the expected positions.

[0162] We observed the following points from the results.

[0163] 1) Without considering wall loss and area priority, the pre-deployment strategy can evenly deploy APs in the global area and maximize signal coverage. Fig.10 The middle mark (a) verifies this point.

[0164] 2) The fine-tuning strategy can only fine-tune the AP to minimize the sum of 1Rx,y in the local area, and also calculates wall loss and regional priority. The fine-tuning strategy works best when used in conjunction with the pre-deployment strategy. Fig.10 This is verified by the figure (b).

[0165] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A signal planning method based on pre-deployment and fine-tuning strategy, characterized in that: The specific steps include: Obtain the signal overlap area between all wireless access points in the target detection area; Divide the target detection area into grids, define the minimum signal overlap area between wireless access points as the first optimization goal, adjust the position of the wireless access point according to the first optimization goal, and obtain the position of the wireless access point with the minimum signal overlap area as the optimal position; Defining the minimum signal overlap area between wireless access points as the first optimization goal specifically includes: ； ； in, Indicates i Wireless access points and j Wireless access points The overlapping area between q Indicates that all wireless access points are x and y The position vector on the axis, r represents the signal radius vector, Indicates i Wireless access points The signal radius, Indicates i Wireless access points The signal radius, k Indicates the total number of wireless access points. Indicates i Wireless access points and j Wireless access points The Euclidean distance between them in the two-dimensional plane; The adjusting the position of the wireless access point according to the first optimization goal specifically includes: initializing the grid, randomly selecting a wireless access point, obtaining the grid where the selected wireless access point is located, and judging whether the grid is in the prohibited area: if it is in the prohibited area, re-obtaining the current grid to judge whether it is in the prohibited area; if it is not in the prohibited area, obtaining the next grid of the current grid, and if the current grid is in the allowed placement area, obtaining the candidate position, initializing the overlapping area, and updating the position of the current wireless access point; repeatedly adjusting the position of the wireless access point until the area of ​​the signal overlapping area reaches a set threshold, and obtaining the position of the wireless access point with the smallest signal overlapping area as the optimal position; calculating the signal value of the wireless access point at the optimal position, and defining the maximum sum of the signal values ​​of all grids as the second optimization goal; Based on the second optimization objective, the best position is adjusted in combination with the quasi-Newton method L-BFGS to obtain the final coordinates of the wireless access point, specifically including: initializing the best coordinates, obtaining the initial coordinates and a positive number as a threshold of the gradient norm; updating the initial coordinates in a preset iteration round, and calculating the gradient of the current updated initial coordinates in each iteration round; judging the gradient of the current updated initial coordinates: if the norm of the gradient is not greater than the threshold of the gradient norm, stopping the iteration and returning the updated initial coordinates as the optimal solution; if the norm of the gradient is greater than the threshold of the gradient norm, calculating the search direction of the gradient, and determining the step size through a straight line search; obtaining the inverse of the Hessian matrix updated in the current iteration round according to the difference of the gradient, the initial coordinates and the difference of the coordinates of the t-th iteration; recalculating the gradient based on the inverse of the updated Hessian matrix, and updating the gradient.

2. The signal planning method based on pre-deployment and fine-tuning strategy according to claim 1 is characterized in that: The obtaining of the signal overlap area between all wireless access points in the target detection area specifically includes: The signal coverage radius of the current wireless access point without considering the attenuation of obstacles is obtained according to the signal strength, and the signal coverage area of ​​all wireless access points in the target area is obtained in combination with the position of the wireless access point; The signal overlap area between all wireless access points in the target detection area is determined based on the signal coverage area.

3. The signal planning method based on pre-deployment and fine-tuning strategy according to claim 1 is characterized in that: The initialization overlap area specifically includes: The overlapping area between the wireless access points is calculated traversally. If the overlapping area between the wireless access points calculated this time is less than the threshold of the compared overlapping area, the minimum overlapping area value is updated.

4. The signal planning method based on pre-deployment and fine-tuning strategy according to claim 1 is characterized in that: The step of calculating the signal value of the wireless access point at the best position specifically includes: Calculate the i Signal value of wireless access point: ； According to i The signal value of the wireless access point is used to calculate the maximum signal value at the target location: ； in, Indicates that the point ( x , y , z ) i The RSS value of each wireless access point, ( x , y , z ) represents the position of the point considered when calculating the RSS value, Indicates i The power of each wireless access point, C The parameter indicating the basis for adjusting the RSS value. Indicates i The signal from the wireless access point passes through the j Signal loss when facing a wall, Indicates i The signal of the wireless access point passes through n Total signal loss when facing a wall, Indicates i The signal wavelength of each wireless access point, θ represents the signal path loss coefficient, i Wireless access points The position is defined as .

5. The signal planning method based on pre-deployment and fine-tuning strategy according to claim 1 is characterized in that: The second optimization objective is to maximize the sum of the signal values ​​of all grids, which specifically includes: ； ； in, Indicates priority Φ is the regional priority, is the location ( x , y , z ) signal value, indicating that at position ( x , y , z ), given the network configuration q The signal strength at and Respectively represent the number of grids in the length and width directions of the indoor map, Indicates that the function is found Get the minimum value The value of .

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the signal planning method based on pre-deployment and fine-tuning strategy as described in any one of claims 1 to 5 is implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the signal planning method based on pre-deployment and fine-tuning strategy as described in any one of claims 1 to 5 is implemented.

Images