Method and device for determining access point deployment scheme, equipment, medium and product
By using scenario prediction models and neighborhood search technology in large venues, the deployment scheme of access points was determined, which solved the problem of uneven Wi-Fi signal coverage and improved network signal quality and the Internet experience of terminal devices.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-22
- Publication Date
- 2026-04-14
AI Technical Summary
In large venues, existing technologies struggle to quickly and accurately determine the optimal layout of access points, resulting in uneven Wi-Fi signal coverage and impacting the internet experience of terminal devices.
By acquiring the floor plan and input vector of the target location, an initial deployment plan is generated using a pre-trained scheme prediction model. Then, a neighborhood search is performed in the search space to generate multiple candidate deployment plans. Finally, the target deployment plan is determined based on the signal quality prediction value.
It improves the network signal quality in the target location, avoids uneven signal coverage, and thus enhances the internet experience of terminal devices.
Smart Images
Figure CN121865282A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of communication technology, and in particular to a method, apparatus, device, medium and product for determining an access point deployment scheme. Background Technology
[0002] In wireless network deployment, personnel often expend significant time and effort on-site for surveying, measurement, and solution development. In large venues such as hotels, office buildings, smart factories, and smart parks, especially in complex environments with multiple floors and walls, the proper placement of access points (APs) is highly dependent on the actual site conditions. Manual surveying and solution design are time-consuming and struggle to accurately simulate real-world signal propagation and interference, potentially leading to uneven Wi-Fi signal coverage and negatively impacting the internet experience of end devices. Summary of the Invention
[0003] This application provides a method, apparatus, device, medium, and product for determining an access point deployment scheme, so as to at least solve the problem of uneven Wi-Fi signal coverage after AP deployment.
[0004] To solve the above-mentioned technical problems, this application is implemented as follows: In a first aspect, embodiments of this application provide a method for determining an access point deployment scheme, comprising: acquiring a floor plan of a target location and an input vector, wherein the input vector includes the area of the target location, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP); obtaining an initial deployment scheme by inputting the floor plan and the input vector into a pre-trained scheme prediction model, wherein the initial deployment scheme includes multiple deployment areas within the target location and AP parameter information corresponding to each deployment area; determining a search space corresponding to each deployment area based on the AP parameter information, generating multiple candidate deployment schemes by performing a neighborhood search within the search space; and determining a target deployment scheme based on the signal quality prediction values of the multiple candidate deployment schemes. Secondly, embodiments of this application provide an apparatus for determining an access point deployment scheme, comprising: an acquisition module for acquiring a floor plan of a target location and an input vector, wherein the input vector includes the area of the target location, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP); a prediction module for obtaining an initial deployment scheme by inputting the floor plan and the input vector into a pre-trained scheme prediction model, wherein the initial deployment scheme includes multiple deployment areas within the target location and AP parameter information corresponding to each deployment area; a search module for determining a search space corresponding to each deployment area based on the AP parameter information, and generating multiple candidate deployment schemes by performing a neighborhood search within the search space; and a determination module for determining a target deployment scheme based on the signal quality estimates of the multiple candidate deployment schemes. Thirdly, embodiments of this application provide an electronic device, which includes a processor and a memory. The memory stores programs or instructions that can run on the processor, and when the programs or instructions are executed by the processor, they implement the steps of the method described in the first aspect above.
[0005] Fourthly, embodiments of this application provide a computer-readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect above.
[0006] Fifthly, embodiments of this application provide a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform the steps of the method described in the first aspect above.
[0007] In this embodiment, a floor plan and input vector of the target location are obtained. An initial deployment plan is obtained by inputting the floor plan and input vector into a pre-trained scheme prediction model. This initial deployment plan includes multiple deployment areas within the target location and AP parameter information corresponding to each deployment area. A search space corresponding to each deployment area is determined based on the AP parameter information, and multiple candidate deployment plans are generated by performing a neighborhood search within the search space. The target deployment plan is determined based on the signal quality estimates of the multiple candidate deployment plans. In this way, by determining multiple deployment areas within the target location through the scheme prediction model, setting corresponding AP parameter information for each deployment area, and performing a neighborhood search within the search space, the network signal quality within the target location can be maximized, avoiding uneven signal coverage and thus improving the internet experience of terminal devices.
[0008] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description
[0009] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.
[0010] Figure 1 A flowchart illustrating a method for determining an access point deployment scheme provided in some embodiments of this application is shown. Figure 2 Example diagrams showing the floor plan recognition results provided in some embodiments of this application are shown; Figure 3 Example diagrams illustrating the outer contour of a floor plan provided in some embodiments of this application are shown; Figure 4 Example diagrams of solution space point sets provided in some embodiments of this application are shown; Figure 5 A schematic diagram of the structure of an access point deployment scheme determination device provided in some embodiments of this application is shown; Figure 6 The diagram shows a schematic representation of the structure of an electronic device provided in some embodiments of this application. Detailed Implementation
[0011] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.
[0012] In the deployment of APs in large venues such as hotels, buildings, smart factories, and smart parks, relevant personnel need to spend a lot of time on site surveys and plan development. Furthermore, the final deployment plan may be based on the surveyors' daily work experience, which may not meet the needs of network signal coverage or lead to waste due to the deployment of too many AP hotspots. In high-density networking environments, the close proximity of APs and mutual interference between channels can also affect the Internet experience of terminal devices, and its importance may even exceed that of signal strength.
[0013] In existing AP deployment methods, heuristic algorithms, such as genetic algorithms or their improved versions, are used based on floor plans to calculate AP locations, aiming to achieve maximum signal coverage with the fewest APs. However, in high-density network environments, the actual number of APs required may far exceed the number needed to achieve sufficient signal coverage due to the limited number of terminals that can connect to a single AP. Furthermore, this method optimizes based solely on device location, neglecting actual signal propagation and interference, which may result in uneven Wi-Fi signal coverage after deployment, negatively impacting the internet experience of terminal devices.
[0014] To address the aforementioned problems in AP deployment, this application provides a method for determining an access point deployment scheme. This method uses a scheme prediction model to determine multiple deployment areas within a target location, sets corresponding AP parameter information for each deployment area, and performs a neighborhood search within the search space to maximize the network signal quality within the target location.
[0015] Please see Figure 1 , Figure 1 This document illustrates a flowchart of a method for determining an access point deployment scheme according to some embodiments of this application. The execution subject of this method can be a terminal device or a server. The terminal device can be a personal computer, a mobile terminal device such as a mobile phone or tablet, or a user-used terminal device. The server can be a standalone server or a server cluster composed of multiple servers. Furthermore, the server can be a backend server for a specific service, or a backend server for a platform or application (e.g., an IoT platform, a wireless network optimization platform, a building management system, etc.). This embodiment uses a server as the execution subject for illustration. For the case of a terminal device, the following related content can be used, and will not be elaborated further here. Figure 1 As shown, the method 100 may include the following steps: Step 101: Obtain the floor plan and input vector of the target location. The input vector includes the area of the target location, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP).
[0016] In one exemplary embodiment, floor plans of target locations such as hotels, buildings, smart factories, and smart parks are acquired. These floor plans can be engineering drawings including building structures and room layouts, or image data generated from panoramic images of the target location. The floor plans are converted into a d×d×3 matrix m based on RGB values, where d is a preset floor plan resolution (e.g., 1024). Furthermore, according to actual needs, the area 'a' of the target location, the number of terminal devices, the network frequency range 'f' (e.g., 0 representing 2.4GHz and 1 representing 5GHz), and the transmission power range of the access point (AP) are acquired (e.g., upper and lower limits k1 and k2, respectively).
[0017] In some possible implementations, a panoramic image of the target location can be acquired, and a pre-trained floor plan recognition model can be used to process the panoramic image to obtain a floor plan. This floor plan recognition model can be built based on convolutional neural networks, generative adversarial networks, region convolutional neural networks, image segmentation networks (such as U-Net), etc.
[0018] For example, the floor plan recognition model uses the U-Net model, which consists of 27 convolutional layers. The model input is a 512×512 three-channel color image. After convolution and pooling operations in the left half of the model, the shallow features of the image are extracted to deep features. After each convolutional module operation, the image size is halved and the number of channels is increased. Dropout is added to the last three convolutional modules to enhance the model's generalization ability and prevent overfitting. The entire left half of the model can be understood as the image encoding process. The right half of the model is convolution and upsampling. The outputs of the corresponding convolutional modules on the left and right sides of the model are concatenated through the concatenation operation to restore the feature information obtained in the image encoding process to the original image size, ensuring that the image features at each stage of encoding are effectively preserved in the process of restoring the original image size. The entire right half of the model can be understood as the image decoding process. According to the requirements of the task, each pixel in the feature map obtained after the encoding-decoding process is classified, and finally, the floor plan corresponding to the original image can be obtained.
[0019] The training sample set can be obtained by labeling the original floor plan. Each pixel in the floor plan sample image is labeled as a load-bearing wall (0 black) or a non-wall (1 white), thereby generating a label image corresponding to the sample image.
[0020] like Figure 2As shown in Figure 3, the user-defined floor plan recognition model identifies the original floor plan as a wall diagram. Starting from the first wall pixel in the lower left corner, the outer boundary of the floor plan is derived by circling the entire floor plan. A ray casting method is used to distinguish the internal points of the floor plan. The points outside the walls constitute the solution space for the application programming interface (AP), as shown in the gray area in Figure 4. The basic idea of the ray casting method is to start from the point to be judged and emit a ray in any direction, then observe the number of intersections between this ray and the polygon boundary. If the number of intersections is odd, the point is inside the polygon; if the number of intersections is even, the point is outside the polygon. This method is applicable to both convex and non-convex polygons, with a time complexity of O(N), where N is the number of sides of the polygon.
[0021] Step 102: The initial deployment plan is obtained by inputting the floor plan and input vector into the pre-trained scheme prediction model.
[0022] The initial deployment plan includes multiple deployment areas within the target site and AP parameter information corresponding to each deployment area.
[0023] Continuing with the above embodiment, the obtained floor plan and input vector are input into a pre-trained deployment prediction model. This model is trained based on historical deployment data and generates an initial deployment plan based on the floor plan and input vector. The initial deployment plan includes the following information: 1) Multiple deployment areas within the target site; 2) AP parameter information for each deployment area, such as AP power and AP channel.
[0024] Step 103: Determine the search space corresponding to each deployment area based on AP parameter information, and generate multiple candidate deployment schemes by performing neighborhood search within the search space.
[0025] Continuing with the above embodiments, based on the AP parameter information in the initial deployment scheme, a search space is determined for each deployment area. This search space is determined considering factors such as the power and channel of each AP. By performing a neighborhood search within each search space, the deployment location, power, and channel of the APs are continuously adjusted to obtain multiple candidate deployment schemes.
[0026] Step 104: Determine the target deployment scheme based on the signal quality estimates of multiple candidate deployment schemes.
[0027] Continuing with the above embodiments, the multiple candidate deployment schemes generated can be evaluated using a signal quality prediction model to obtain a signal quality prediction value. This model estimates the signal quality in different areas of the target location based on the AP deployment, power, and channel of each candidate deployment scheme, thus obtaining the signal quality prediction value. The optimal candidate deployment scheme is then determined as the target deployment scheme based on the signal quality prediction values of each scheme.
[0028] This application provides a method for determining an access point deployment scheme. The method involves obtaining a floor plan and input vector of a target location; inputting the floor plan and input vector into a pre-trained scheme prediction model to obtain an initial deployment scheme, which includes multiple deployment areas within the target location and AP parameter information corresponding to each deployment area; determining a search space for each deployment area based on the AP parameter information; generating multiple candidate deployment schemes by performing a neighborhood search within the search space; and determining the target deployment scheme based on the signal quality estimates of the multiple candidate deployment schemes. In this way, by determining multiple deployment areas within the target location through a scheme prediction model, setting corresponding AP parameter information for each deployment area, and performing a neighborhood search within the search space, the network signal quality within the target location can be maximized, avoiding uneven signal coverage and thus improving the internet experience of terminal devices.
[0029] In some embodiments, the AP parameter information mentioned above includes the probability of deploying APs in each deployment region and the performance parameters of the APs; in step 103 above, the search space corresponding to each deployment region is determined based on the AP parameter information, and multiple candidate deployment schemes are generated by performing a neighborhood search within the search space, including: Deployment areas where the probability of deploying an AP is greater than a preset probability threshold are defined as the location search space; the performance search space is determined based on the performance parameters of the APs corresponding to the location search space; and multiple candidate deployment schemes are determined based on the location search space and the performance search space corresponding to the location search space.
[0030] In an exemplary embodiment, the scheme prediction model divides the target location into multiple deployment areas using a floor plan, for example, dividing the target location into multiple grids. Based on the input vector, the probability of deploying an AP in each grid and the AP's performance parameters such as channel and power are determined. All grids with an AP deployment probability greater than a preset probability threshold P1 are defined as the AP deployment location search space, where the preset probability threshold P1 can be set according to actual needs, for example, P1=0.7. Based on the performance parameters of the APs corresponding to the location search space, a performance search space is determined. Based on the location search space and the performance search space, multiple candidate deployment schemes are determined.
[0031] In some possible implementations, the aforementioned AP performance parameters include the AP's power and the AP's channel; the aforementioned performance search space includes a power search space and a channel search space; the aforementioned determination of the performance search space based on the AP's performance parameters corresponding to the location search space includes: The power search space is determined based on the power of the AP corresponding to the location search space and the preset power offset; the channel search space is determined based on the channel of the AP corresponding to the location search space and the preset channel offset.
[0032] Continuing with the above embodiments, the preset power offset t can be set according to actual needs, for example, ±3dBm, defining the power k of the AP corresponding to the location search space and its range within ±3dBm as the power search space. The preset channel offset can be set according to actual needs, for example, 1, defining the channel C of the AP corresponding to the location search space and its adjacent channels as the channel search space.
[0033] In some embodiments, step 104 above, determining the target deployment scheme based on the signal quality estimates of multiple candidate deployment schemes, includes: Based on the preset signal attenuation model and the AP parameter information corresponding to each deployment area in multiple candidate deployment schemes, the signal quality information of multiple sampling points in the target location is determined; based on the signal quality information of multiple sampling points, the signal quality estimate is determined; and through a heuristic algorithm, with the goal of minimizing the signal quality estimate, the target deployment scheme among the multiple candidate deployment schemes is determined.
[0034] In an exemplary embodiment, a signal attenuation model can be constructed based on the signal intensity attenuation process in space. The formula for the signal attenuation model is as follows: ; in, RSSI For signal strength, This refers to the device's transmit power (dBm). For free space signal loss, D is the distance. Where is the signal frequency (MHz), and N is the thickness of the obstacle (m). The building medium loss coefficient per unit thickness (dBm / m) varies depending on the building material.
[0035] Based on the signal attenuation model described above and the AP parameter information corresponding to each deployment area in multiple candidate deployment schemes, the signal quality information of multiple sampling points within the target location is determined. This signal quality information may include the signal coverage score h, the signal interference score i, and the AP distribution uniformity score g.
[0036] For example, the signal coverage score h can be obtained by calculating the signal coverage rate of each sampling point within the floor plan. Coverage rate e = number of detection points with signal strength greater than d / total number of detection points. The signal strength of the sampling point Dj = Max(Si), where Si is the signal strength of the i-th AP at the sampling point, and the score h = (1-e) × 10000.
[0037] The signal interference score *i* can be obtained by calculating the signal interference rate at each sampling point. The interference rate *f* = number of sampling points with signal interference / total number of sampling points. The criterion for determining whether a sampling point has interference is whether it receives signals from multiple access points (APs). For example, at least two APs must have signal strengths greater than 70 dBm, and the strength difference between the two signals must be less than 20 dBm. The score *i* = *f* × 10000.
[0038] The AP distribution uniformity score g can be obtained by calculating the standard deviation j of the number of sampling points covered by each AP. Each sampling point is only covered by the AP with the highest signal strength at that sampling point. The number of sampling points is m, and the score g = j / m.
[0039] Furthermore, based on the signal quality information from multiple sampling points, a signal quality estimate is determined. For example, the signal quality information may include a weighted sum of the signal coverage score h, the signal interference score i, and the AP distribution uniformity score g, which is then used to determine the signal quality estimate.
[0040] Using heuristic algorithms, with the goal of minimizing the predicted signal quality, a target deployment scheme is determined from among multiple candidate deployment schemes. Heuristic algorithms include genetic algorithms, particle swarm optimization, and simulated annealing. The heuristic algorithm searches the search space, optimizing for minimizing the predicted signal quality within the target location. Through a finite number of searches, the target deployment scheme is obtained.
[0041] In this embodiment, the floor plan is identified by the scheme prediction model to predict multiple deployment areas and the power and channel of the APs corresponding to each deployment area. Based on the approximate deployment scheme predicted by this model, a heuristic algorithm is used to perform neighborhood search, which can quickly find the approximate optimal deployment scheme of APs in high-density networking scenarios.
[0042] In some embodiments, the above-described scheme prediction model is trained in the following manner; Obtain a training sample set, which includes multiple training samples and label vectors corresponding to each training sample. Each training sample includes a floor plan sample of the location sample and an input vector sample. The input vector sample includes the area of the location sample, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP). The pre-defined convolutional neural network model is trained using a training sample set. During the training process, the convolutional kernels are dynamically adjusted according to the input vector samples to obtain the weighting coefficients of each convolutional kernel. Based on each convolutional kernel and its corresponding weighting coefficients, deep features are determined. Based on the deep features, the output vector is determined. The output vector includes multiple deployment area samples from the floor plan sample and AP parameter sample information corresponding to each deployment area sample. Based on the AP parameter sample information, multiple deployment area samples are aggregated to obtain the target output vector; The loss value is determined based on the target output vector, the label vector, and the preset loss function; The network parameters of the convolutional neural network are adjusted based on the loss value to obtain the scheme prediction model.
[0043] In one exemplary embodiment, the convolutional neural network model includes an input layer, a backbone network, and an output layer. The input layer may employ CondConv, dynamically adjusting the convolution kernel by using the input vector sample x as a conditional vector, and using this convolution kernel in subsequent convolution operations. Specifically, the network will have a set of basic convolution kernels. (Each convolutional kernel represents an "expert"), and then the weighting coefficients for these basic convolutional kernels are calculated using conditional vectors. The calculation formula is similar to: ; in, It is a condition vector x The weights are calculated using a learnable function.
[0044] The backbone network can use EfficientNet, which reduces the computational cost and number of parameters while maintaining high accuracy, making it more efficient. The activation function used is the swish function, defined as f(x) = x × σ(βx), which offers better performance and faster convergence compared to ReLU.
[0045] The output of the convolutional neural network is an n×n×v matrix M, where n is a value less than d1 and divisible by d1 (a fixed value is set according to the scenario), representing the division of the floor plan into n×n grids. The vector v contains the probability p of deploying an AP in that grid, the AP's power k, and the AP's channel c. Channel c can be represented using one-hot encoding. 2.4G has three non-interfering channels, and 5G has five available channels in China.
[0046] The convolutional neural network model described above is trained using the training sample set to obtain the scheme prediction model.
[0047] Since the output is an n×n×v matrix, directly calculating the loss function for each output item would be computationally intensive, and the loss would be spread across every cell, affecting the model's accuracy and convergence speed. Therefore, the model aggregates multiple deployment area samples in the output vector based on AP parameter sample information to obtain the target data vector. The loss value is determined based on the target output vector, label vector, and a preset loss function. However, conventional loss functions cannot accurately quantify the differences between two AP deployment schemes. Calculating the signal score difference across the entire floor plan as the loss function would be too complex, and the relationship between signal score and deployment location would be too implicit, preventing model convergence. Therefore, this embodiment of the application designs a special loss function calculation method for AP deployment, as follows: First, the b target output vectors of the result set are paired one by one with the b label vectors of the sample using the Hungarian algorithm (with coordinate distance as the influence factor).
[0048] loss function . For pairing i The loss value.
[0049] Paired loss value . in, d 2 represents the distance between the two paired positions; k This represents the difference in power k between the paired components; This represents the channel spacing between pairs. α and β are correction factors that can be adjusted according to specific circumstances.
[0050] This loss function primarily considers the distance between access points (APs), supplemented by power and the number of channels as secondary influencing factors. This approach improves model accuracy while maintaining training speed. The network parameters of the convolutional neural network are then adjusted based on the loss value to obtain the scheme prediction model.
[0051] In one possible implementation, the AP parameter sample information mentioned above includes the probability of deploying APs in each deployment area sample and the performance parameters of the APs; the above-mentioned aggregation processing of multiple deployment area samples based on the AP parameter sample information to obtain an output vector includes: Sort the probability of deploying APs in the deployment area sample from high to low, and select the deployment areas with the highest probability as candidate target deployment areas. Filter out deployment area samples whose distance from the selected candidate target deployment areas is less than the target distance, and obtain updated candidate deployment area samples. Continue until the number of candidate deployment area samples equals the number of APs, and determine the final candidate deployment area samples as the output vector. The target distance is the ratio of the area of the location corresponding to the floor plan sample to the number of access points (APs), and the number of APs is the ratio of the number of terminal devices in the input vector sample to the preset maximum number of access terminals per AP.
[0052] In one exemplary embodiment, the above aggregation steps are as follows: 1) Sort the probability p of each cell (i.e., the deployment area sample), take the vector v of the cell with the highest probability value and add the coordinate values x, y of the cell to generate vector V as one of the elements of the result set; 2) Remove cells from the remaining cells that are less than [the distance to the cell]. d A grid of 3, where, , where a is the area of the target location and b is the number of APs; where the number of APs is the ratio of the number of terminal devices in the input vector sample to the preset maximum number of connected terminals per AP; 3) If the number of elements in the result set reaches the AP value, then the process ends; otherwise, continue with steps 1) and 2) above.
[0053] After performing the above aggregation steps, the output result set R is obtained, where R is a set of vectors V of size b.
[0054] In this embodiment, conditional convolution is used to transform the input during the construction of the convolutional neural network, which can improve the model's generalization ability. Furthermore, by aggregating the output results and defining an improved loss function, the computational cost of the model can be reduced, and the model's accuracy can be improved.
[0055] The training sample set mentioned above can be manually labeled for the floor plan samples to obtain the label vector corresponding to the floor plan sample; or it can be identified by the convolutional neural network to obtain a standard binary floor plan, and the AP deployment scheme (i.e., label vector) can be obtained by using a heuristic algorithm optimization method. Training data with the original floor plan and high-density network deployment requirements as input and the AP deployment scheme as output can be generated in batches for the prediction of approximate AP deployment scheme. Compared with manual labeling, this method can quickly construct a large amount of effective training data without spending a lot of manpower to survey and label AP location, power, signal and other information.
[0056] This application provides a method for determining access point deployment schemes. This method can be applied to the planning and formulation of wireless network deployment schemes in commercial office areas such as hotels and buildings, as well as large venues such as factories, parks, and stadiums. It can also be used for quickly estimating scheme quotes and guiding installation and maintenance engineers in installation and deployment. Compared with traditional network planning methods, this method is more efficient, requires less manpower, and has a wider range of commercial applications and greater commercial value.
[0057] Please refer to Figure 5. Figure 5 A schematic diagram of the structure of an access point deployment scheme determination device provided in some embodiments of this application is shown. This access point deployment scheme determination device can achieve the following: Figure 1 The access point deployment scheme determination device 500, comprising all or part of the contents shown in the embodiment, includes: The acquisition module 510 is used to acquire the floor plan of the target location and the input vector, wherein the input vector includes the area of the target location, the number of terminal devices, the network frequency range, and the transmission power range of the access point (AP). The prediction module 520 is used to obtain an initial deployment plan by inputting the floor plan and the input vector into a pre-trained scheme prediction model. The initial deployment plan includes multiple deployment areas within the target site and AP parameter information corresponding to each deployment area. Search module 530 is used to determine the search space corresponding to each deployment area based on the AP parameter information, and generate multiple candidate deployment schemes by performing neighborhood search within the search space. The determination module 540 is used to determine the target deployment scheme based on the signal quality prediction values of the multiple candidate deployment schemes.
[0058] In some embodiments, the AP parameter information includes the probability of deploying an AP in each deployment area and the performance parameters of the AP; the search module 530, when determining the search space corresponding to each deployment area based on the AP parameter information, and generating multiple candidate deployment schemes by performing a neighborhood search within the search space, is specifically used for: Deployment areas with a probability of deploying APs greater than a preset probability threshold in multiple deployment areas are defined as the location search space; The performance search space is determined based on the performance parameters of the AP corresponding to the location search space; Based on the location search space and the corresponding performance search space, multiple candidate deployment schemes are determined.
[0059] In some possible implementations, the performance parameters of the AP include the AP's power and the AP's channel; the performance search space includes a power search space and a channel search space; the search module 530 described above, when determining the performance search space based on the performance parameters of the AP corresponding to the location search space, is specifically used for: The power search space is determined based on the power of the AP corresponding to the location search space and the preset power offset; The channel search space is determined based on the channel of the AP corresponding to the location search space and the preset channel offset.
[0060] In some embodiments, the determining module 540, when determining the target deployment scheme based on the signal quality estimates of the plurality of candidate deployment schemes, is specifically configured to: Based on the preset signal attenuation model and the AP parameter information corresponding to each deployment area in multiple candidate deployment schemes, the signal quality information of multiple sampling points in the target location is determined; Based on the signal quality information from the multiple sampling points, a signal quality estimate is determined; Using a heuristic algorithm, the target deployment scheme is determined from among the multiple candidate deployment schemes with the goal of minimizing the signal quality prediction.
[0061] In some embodiments, the scheme prediction model is trained in the following manner; Obtain a training sample set, which includes multiple training samples and label vectors corresponding to each training sample. Each training sample includes a floor plan sample of the location sample and an input vector sample. The input vector sample includes the area of the location sample, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP). The training sample set is used to train a preset convolutional neural network model. During the training process, the convolution kernels are dynamically adjusted according to the input vector samples to obtain the weighting coefficients of each convolution kernel. Based on each convolution kernel and its corresponding weighting coefficients, deep features are determined. Based on the deep features, an output vector is determined. The output vector includes multiple deployment area samples from the floor plan sample and AP parameter sample information corresponding to each deployment area sample. Based on the AP parameter sample information, the multiple deployment area samples are aggregated to obtain the target output vector; The loss value is determined based on the target output vector, the label vector, and the preset loss function; The network parameters of the convolutional neural network are adjusted based on the loss value to obtain the scheme prediction model.
[0062] In some embodiments, the AP parameter sample information mentioned above includes the probability of deploying APs in each deployment area sample and the performance parameters of the APs; the above-mentioned aggregation processing of the multiple deployment area samples based on the AP parameter sample information to obtain an output vector includes: The probability of deploying APs in the deployment area sample is sorted from high to low, and the deployment areas with the highest probabilities are selected as candidate target deployment areas. Filter out deployment area samples in the deployment area sample whose distance to the selected candidate target deployment area is less than the target distance, and obtain updated candidate deployment area samples. Continue until the number of candidate deployment area samples is equal to the number of APs, and then determine the final candidate deployment area samples as the output vector. Wherein, the target distance is the ratio of the area of the location corresponding to the floor plan sample to the number of APs, and the number of APs is the ratio of the number of terminal devices in the input vector sample to the preset maximum number of terminals that can be accessed by each AP.
[0063] This application provides an apparatus for determining an access point deployment scheme, including a vector determination module, a scheme prediction module, a neighborhood search module, and a scheme determination module. The vector determination module acquires a floor plan of the target location and an input vector, the input vector including the area of the target location, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP). The scheme prediction module inputs the floor plan and the input vector into a pre-trained scheme prediction model to obtain an initial deployment scheme, which includes multiple deployment areas within the target location and AP parameter information corresponding to each deployment area. The neighborhood search module determines the search space corresponding to each deployment area based on the AP parameter information, and generates multiple candidate deployment schemes by performing a neighborhood search within the search space. The scheme determination module determines the target deployment scheme based on the signal quality estimates of the multiple candidate deployment schemes. In this way, by determining multiple deployment areas within the target location through the scheme prediction model, setting corresponding AP parameter information for each deployment area, and performing a neighborhood search within the search space, the network signal quality within the target location can be maximized, avoiding uneven signal coverage and thus improving the internet experience of terminal devices.
[0064] Figure 6The diagram illustrates the structure of an electronic device according to some embodiments of this application. Referring to the diagram, at the hardware level, the electronic device 600 includes a processor 610, and optionally includes an internal bus 620, a network interface 630, and a memory. The memory may include main memory 641, such as high-speed random-access memory (RAM), and may also include non-volatile memory 642, such as at least one disk storage device. Of course, the electronic device may also include other hardware required for other services.
[0065] The processor 610, network interface 630, and memory can be interconnected via an internal bus 620. This internal bus 620 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be categorized as an address bus, data bus, control bus, etc. For ease of illustration, only a single bidirectional arrow is used in this diagram, but this does not imply that there is only one bus or one type of bus.
[0066] The memory stores programs. Specifically, the program may include program code, which includes computer operation instructions. The memory may include main memory 641 and non-volatile memory 642, and provides instructions and data to the processor 610.
[0067] Processor 610 reads the corresponding computer program from non-volatile memory 642 into memory and then runs it, forming a device for locating the target user at the logical level. Processor 610 executes the program stored in memory and specifically performs the following: Figure 1 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods described in the preceding method embodiments, and will not be repeated here.
[0068] The above is as stated in this application. Figure 1The methods disclosed in the illustrated embodiments can be applied to or implemented by processor 610. Processor 610 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above methods can be completed by integrated logic circuits in the hardware or by instructions in software form within processor 610. Processor 610 can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.
[0069] The computer device can also execute the methods described in the preceding method embodiments and achieve the functions and beneficial effects of the methods described in the preceding method embodiments, which will not be repeated here.
[0070] Of course, in addition to software implementation, the electronic device of this application does not exclude other implementation methods, such as logic devices or a combination of hardware and software, etc. In other words, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.
[0071] This application also proposes a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform... Figure 1 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods described in the preceding method embodiments, and will not be repeated here.
[0072] The computer-readable storage medium mentioned above includes read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc.
[0073] Furthermore, embodiments of this application also provide a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, implement the following process: Figure 1 The methods disclosed in the embodiments shown achieve the functions and beneficial effects of the methods described in the preceding method embodiments, and will not be repeated here.
[0074] The embodiments of this application can be applied to various scenarios of electronic device collaboration or interconnection, including: collaborative interconnection between mobile phones and laptops / tablets; collaborative interconnection between mobile sampling terminals and smart TVs / monitors; collaborative interconnection between mobile phones or tablets and in-vehicle entertainment systems; collaborative interconnection between mobile sampling terminals and smart conferencing systems, etc. This satisfies users' diverse needs in smart home, smart office, and smart travel scenarios.
[0075] In summary, the above description is merely a preferred embodiment of this application and does not limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.
[0076] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.
[0077] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0078] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0079] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.
Claims
1. A method for determining an access point deployment scheme, characterized in that, include: Obtain the floor plan and input vector of the target location. The input vector includes the area of the target location, the number of terminal devices, the network frequency range, and the transmission power range of the access point (AP). By inputting the floor plan and the input vector into a pre-trained scheme prediction model, an initial deployment scheme is obtained. The initial deployment scheme includes multiple deployment areas within the target location and AP parameter information corresponding to each deployment area. Based on the AP parameter information, the search space corresponding to each deployment area is determined, and multiple candidate deployment schemes are generated by performing a neighborhood search within the search space. The target deployment scheme is determined based on the signal quality estimates of the multiple candidate deployment schemes.
2. The method according to claim 1, characterized in that, The AP parameter information includes the probability of deploying an AP in each deployment region and the performance parameters of the AP; the step of determining the search space corresponding to each deployment region based on the AP parameter information, and generating multiple candidate deployment schemes by performing a neighborhood search within the search space, includes: Deployment areas with a probability of deploying APs greater than a preset probability threshold in multiple deployment areas are defined as the location search space; The performance search space is determined based on the performance parameters of the AP corresponding to the location search space; Based on the location search space and the corresponding performance search space, multiple candidate deployment schemes are determined.
3. The method according to claim 2, characterized in that, The performance parameters of the AP include the AP's power and the AP's channel; the performance search space includes the power search space and the channel search space; The step of determining the performance search space based on the performance parameters of the AP corresponding to the location search space includes: The power search space is determined based on the power of the AP corresponding to the location search space and the preset power offset; The channel search space is determined based on the channel of the AP corresponding to the location search space and the preset channel offset.
4. The method according to claim 1, characterized in that, The step of determining the target deployment scheme based on the signal quality prediction values of the multiple candidate deployment schemes includes: Based on the preset signal attenuation model and the AP parameter information corresponding to each deployment area in multiple candidate deployment schemes, the signal quality information of multiple sampling points in the target location is determined; Based on the signal quality information from the multiple sampling points, a signal quality estimate is determined; Using a heuristic algorithm, the target deployment scheme is determined from among the multiple candidate deployment schemes with the goal of minimizing the signal quality prediction.
5. The method according to claim 1, characterized in that, The prediction model for the proposed scheme is trained using the following method; Obtain a training sample set, which includes multiple training samples and label vectors corresponding to each training sample. Each training sample includes a floor plan sample of the location sample and an input vector sample. The input vector sample includes the area of the location sample, the number of terminal devices, the network frequency range, and the transmit power range of the access point (AP). The preset convolutional neural network model is trained using the training sample set. During the training process, the convolutional kernels are dynamically adjusted according to the input vector samples to obtain the weighting coefficients of each convolutional kernel. Based on each convolutional kernel and its corresponding weighting coefficients, deep features are determined. Based on the deep features, an output vector is determined, which includes multiple deployment area samples from the floor plan sample and AP parameter sample information corresponding to each deployment area sample. Based on the AP parameter sample information, the multiple deployment area samples are aggregated to obtain the target output vector; The loss value is determined based on the target output vector, the label vector, and the preset loss function; The network parameters of the convolutional neural network are adjusted based on the loss value to obtain the scheme prediction model.
6. The method according to claim 5, characterized in that, The AP parameter sample information includes the probability of deploying an AP in each deployment area and the performance parameters of the AP; the aggregation process of the multiple deployment area samples based on the AP parameter sample information to obtain an output vector includes: The probability of deploying APs in the deployment area sample is sorted from high to low, and the deployment areas with the highest probabilities are selected as candidate target deployment areas. Filter out deployment area samples in the deployment area sample whose distance to the selected candidate target deployment area is less than the target distance, and obtain updated candidate deployment area samples. Continue until the number of candidate deployment area samples is equal to the number of APs, and then determine the final candidate deployment area samples as the output vector. Wherein, the target distance is the ratio of the area of the location corresponding to the floor plan sample to the number of APs, and the number of APs is the ratio of the number of terminal devices in the input vector sample to the preset maximum number of terminals that can be accessed by each AP.
7. A device for determining an access point deployment scheme, characterized in that, include: The acquisition module is used to acquire the floor plan of the target location and the input vector, wherein the input vector includes the area of the target location, the number of terminal devices, the network frequency range, and the transmission power range of the access point (AP). The prediction module is used to obtain an initial deployment plan by inputting the floor plan and the input vector into a pre-trained scheme prediction model. The initial deployment plan includes multiple deployment areas within the target site and AP parameter information corresponding to each deployment area. The search module is used to determine the search space corresponding to each deployment area based on the AP parameter information, and generate multiple candidate deployment schemes by performing a neighborhood search within the search space. The determination module is used to determine the target deployment scheme based on the signal quality prediction values of the multiple candidate deployment schemes.
8. An electronic device, characterized in that, The electronic device includes a processor and a memory, the memory storing programs or instructions that can run on the processor, the programs or instructions being executed by the processor to implement the steps of the method as described in any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program or instructions that, when executed by a processor, implement the steps of the method as described in any one of claims 1 to 6.
10. A computer program product, characterized in that, The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions that, when executed by a computer, cause the computer to perform the steps of the method as described in any one of claims 1 to 6.