Wireless indoor positioning method and device based on DAE-ALSTM, equipment and medium

Abstract

The invention discloses a wireless indoor positioning method and device based on DAE-ALSTM, equipment and a medium, and the method comprises the steps: collecting coordinates and RSSI data of each point location in an environment, and constructing a signal route map; preprocessing the signal route map to obtain a data set, and performing offline training and verification on the DAE-ALSTM model by using the data set to obtain a trained DAE-ALSTM model; and performing online position prediction according to the target RSSI data by using the trained DAE-ALSTM model. According to the wireless indoor positioning method based on the DAE-ALSTM, the dimension of an RSSI sample is reduced by using the DAE, so that stable characteristics are input into an ALSTM network, main characteristics of input data are extracted from the sample, and stable and accurate position prediction is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless Wi-Fi signal indoor positioning, and in particular to a wireless indoor positioning method, device, equipment and medium based on DAE-ALS™. Background Art

[0002] For years, research and exploration have been ongoing into using AP technology to provide indoor positioning services. Leveraging existing AP infrastructure in indoor environments can significantly reduce deployment costs. AP-based indoor positioning can be further categorized into two main categories: channel state information (CSI) and received signal strength indicator (RSSI). In small-scale indoor environments, CSI-based methods provide more accurate target positioning than RSSI-based systems. This improved accuracy is attributed to orthogonal frequency division multiplexing (OFDM) and multiple-input, multiple-output (MIMO). However, CSI-based methods require specialized hardware to implement. RSSI-based AP fingerprinting, on the other hand, does not require specialized hardware for indoor positioning. AP fingerprinting methods have been successful for indoor location-based services (LBS) in small, single-story buildings. However, these techniques face challenges in accuracy and robustness in large, complex scenarios.

[0003] Existing research has demonstrated the feasibility of deep learning-based methods for indoor positioning. However, these studies have not considered the temporal variation of RSSI patterns, the spatial uncertainty of fingerprints, and the sparse RSSI samples. They have leveraged the combined strengths of denoising autoencoders (DAEs), attention mechanisms, and long short-term memory (LSTMs) to overcome the practical challenges associated with AP signal fingerprinting in indoor positioning schemes.

[0004] For example, the invention application with application number 202110305445.0 discloses an indoor positioning method and system based on the RSSI characteristics of Wi-Fi signals. The positioning method includes the following process: This application solution can reduce labeling costs and ensure security and privacy. However, its solution also has the following problems: it does not take into account the temporal variation of RSSI patterns, the uncertainty of fingerprint space, and the sparse RSSI samples, making it difficult to obtain accurate positioning results. Summary of the Invention

[0005] The purpose of the present invention is to provide a wireless indoor positioning method, device, equipment and medium based on DAE-ALSTM, combining the comprehensive advantages of denoising autoencoder (DAE), attention mechanism and long short-term memory (LSTM), using the trained DAE-ALSTM model to perform online position prediction based on target RSSI data.

[0006] The embodiments of the present invention provide a wireless indoor positioning method, apparatus, device and medium based on DAE-ALS™.

[0007] A first aspect: A wireless indoor positioning method based on DAE-ALSTM, comprising the steps of:

[0008] S1. Deploy multiple AP access points in the target environment, collect coordinates and RSSI data of each point in the environment, and build a signal route map.

[0009] S2. Preprocess the signal route map to obtain a data set, and divide the data set into a training set and a validation set;

[0010] S3. Use the dataset to perform offline training and verification on the DAE-ALSTM model to obtain the trained DAE-ALSTM model.

[0011] S4. Use the trained DAE-ALSTM model to perform online location prediction based on the target RSSI data.

[0012] Furthermore, the signal route map is constructed as follows:

[0013] X={(P1,RSSI1,(x1,y1)),…,(P i ,RSSI i ,(x i ,y i )),…,(P N ,RSSI N ,(x N ,y N ))}

[0014] Among them, X represents the RSSI data of the point, P i is the i-th AP access point, RSSI i is the i-th AP access point at position (x i ,y i )’s signal receiving strength, and N represents the number of AP access points.

[0015] Furthermore, the DAE-ALSTM model includes a denoising autoencoder DAE, the DAE includes multiple layers of encoding, and the DAE processing of data includes:

[0016] S21. Introduce noise into the input data to obtain noise data. The formula is expressed as:

[0017]

[0018] Where Σ is the covariance matrix, which represents the correlation between noise in different dimensions, ∈~N(0,Σ) means that the noise is sampled from a normal distribution;

[0019] S22. Input the noise data into the encoder to obtain the key feature data of dimensionality reduction. The formula is expressed as:

[0020]

[0021] Among them, Z represents the key feature data of dimensionality reduction, W enc represents the weight of the encoder, b enc represents the bias of the encoder, L represents the number of layers of the multi-layer encoder, γ i Represents the regularization coefficient of each layer encoder, ReLU is the activation function applied to each residual module, W i is the weight in each layer of residual network, b i For each bias.

[0022] Furthermore, the DAE-ALSTM model includes an ALSTM network, and the ALSTM network includes multiple LSTM layers and an Attention layer.

[0023] Furthermore, the Attention layer introduces an attention mechanism to assign variable weights to different features of the input data according to their relative importance, thereby optimizing the prediction accuracy of the target position. The following steps are included:

[0024] S31. Introduce an attention mechanism to focus on the most important signal sources and time steps of the input data for target position prediction;

[0025] S32. Calculate the attention weight based on the correlation between the input data and the access point position, indicating the relative importance of the time step to the target position prediction;

[0026] S33. According to the calculated attention weight, the predicted features of each time step output by the ALSTM model are weighted to obtain the optimized features. The formula is expressed as: Z′ t =α t ·Z t

[0027] Among them, α t represents the attention weight, Z t Denotes the predicted features of the improved ALSTM model for time step t, Z′ t Indicates the features after the attention mechanism is introduced;

[0028] S34, the weighted feature Z' t Input to the fully connected layer, perform classification regression, and obtain the target position coordinate (x, y) output. The formula is expressed as:

[0029] (x,y)=f regression (Z′)

[0030] Among them, Z′ represents the weighted feature set of all time steps, f regression Represents a regression classification function.

[0031] Furthermore, a Dropout layer is provided between the multiple LSTM layers to randomly discard a portion of neurons during data processing to reduce overfitting.

[0032] A second aspect: A wireless indoor positioning device based on DAE-ALS™, comprising:

[0033] Data construction module, used to collect coordinates and RSSI data of various points in the environment and build a signal route map;

[0034] A preprocessing module is used to preprocess the signal route map to obtain a data set;

[0035] The DAE-ALSTM module uses the DAE-ALSTM model to perform online location prediction based on target RSSI data.

[0036] A third aspect: 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, the steps of the method provided in the first aspect are implemented.

[0037] A fourth aspect: A non-transitory computer-readable storage medium having a computer program stored thereon, which implements the steps of the method provided in the first aspect when executed by a processor.

[0038] Beneficial effects of the present invention:

[0039] 1. The wireless indoor positioning method based on DAE-ALSTM of the present invention uses DAE to reduce the dimension of RSSI samples, thereby achieving stable feature input into the ALSTM network and extracting the main features from the samples, which can achieve robust and improved location prediction.

[0040] 2. The present invention learns the nonlinear complexity between RSSI samples and location coordinates in the regression task, and focuses on the masked temporal changes of RSSI samples through ALSTM, which makes up for the problem of insufficient accuracy of traditional LSTM networks, thereby improving accuracy and efficiency and reducing computational costs.

[0041] 3. The wireless indoor positioning method based on DAE-ALSTM in the present invention adopts a stratified sampling method to select a validation set from the entire data set during the training process. This method solves the problem of sample imbalance caused by random selection, ensures that the validation set has the same distribution characteristics as the overall data, and has similar RSSI sample distribution and characteristics, thereby improving the model generalization ability.

[0042] 4. This invention innovatively integrates the attention mechanism with the LSTM network (ALSTM), strengthens the learning of key time step features through dynamic weight allocation, effectively captures the temporal variation of RSSI signals, and improves positioning accuracy in complex environments. At the same time, it introduces the Dropout layer for regularization processing to prevent model overfitting and enhance system robustness.

[0043] 5. The end-to-end framework DALLoc of the present invention realizes the joint optimization of feature extraction and position prediction, which improves the computational efficiency by more than 30% compared with traditional methods. In large and complex scene tests, the positioning error is reduced to within 0.8 meters, which is 42% higher than the accuracy of a single LSTM model.

[0044] 6. The method and device of the present invention have strong engineering practicality and are compatible with existing Wi-Fi infrastructure. They do not require dedicated hardware equipment to support dynamic environment adaptation. When the building structure changes, only part of the database needs to be updated, reducing maintenance costs by 60%. This solution effectively solves industry pain points such as fingerprint spatial uncertainty, signal time-varying and sample sparsity through deep learning methods, providing a cost-effective technical solution for application scenarios such as smart cities and indoor navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 Schematic diagram of the process of the wireless indoor positioning method based on DAE-ALS™ of the present invention;

[0046] Figure 2 Flowchart showing the principle of the wireless indoor positioning method based on DAE-ALS™ of the present invention;

[0047] Figure 3 This is a structural framework diagram of the DAE-ALS™ of the present invention;

[0048] Figure 4 This is a structural framework diagram of the wireless indoor positioning device based on DAE-ALS™ of the present invention;

[0049] Figure 5 Schematic diagram of the structure of the electronic device of the present invention.

[0050] For ease of understanding, the Chinese and English translations of the attached figures are listed below: DALLoc SystemOverviewDALLoc System Overview;

[0051] DAE-ALSTM ModelDAE-ALSTM model;

[0052] AP wireless access point;

[0053] Database;

[0054] Data Preprocessing

[0055] Raw Input

[0056] All Dataset all data sets;

[0057] Normalization;

[0058] Random Sampling

[0059] Training Set

[0060] Validation Set

[0061] Training stage (offline) offline training stage;

[0062] Gaussian NoiseGaussian noise;

[0063] Encoder Layer encoding layer;

[0064] LSTM LayerLSTM layer;

[0065] Dropout Layer discards the layer;

[0066] Attention Layer

[0067] Drawer Delay drawer delay;

[0068] Predict stage (online) online prediction stage;

[0069] Testing Set

[0070] Dense Layer fully connected layer;

[0071] Output Layer output layer;

[0072] Predicted Location Predicted location. DETAILED DESCRIPTION

[0073] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0074] The current Wi-Fi fingerprint recognition technology based on the received signal strength indicator (RSSI) has problems with accuracy and robustness when applied in large-scale and complex scenarios.

[0075] In view of the above problems, the present invention provides a wireless indoor positioning method based on DAE-ALSTM. Figure 1 A flow chart of a wireless indoor positioning method based on DAE-ALS™ provided in an embodiment of the present invention includes the following steps:

[0076] S1. Deploy multiple AP access points in the target environment, collect the coordinates and RSSI data of each point in the environment, and build a signal route map.

[0077] like Figure 2 As shown in the figure, the APs in the target environment are used to collect the combined data of RSSI and location coordinates to build a signal route map.

[0078] Signal route map, the formula can be expressed as:

[0079] X={(P1,RSSI1,(x1,y1)),...,(P i ,RSSI i ,(x i ,y i )),...,(P N ,RSSI N ,(x N ,y N ))}

[0080] Among them, X represents the RSSI data of the point, P i is the i-th AP access point, RSSI i is the i-th AP access point at position (x i ,y i )’s signal receiving strength, and N represents the number of AP access points.

[0081] S2. Preprocess the signal route map to obtain a data set, and divide the data set into a training set and a validation set;

[0082] like Figure 2As shown in the figure, the preprocessing process includes: first, integrating the full dataset (AllDataset) of the signal route maps of all APs; then, eliminating dimensional differences and unifying the data range for normalization (Normalization); then, enhancing the model generalization ability and avoiding overfitting, random sampling (Random Sampling) is performed to obtain the dataset; finally, the dataset is divided into a training set (Training Set), a validation set (Validation Set) for offline training, and a testing set (Testing Set) for online prediction.

[0083] S3. Use the dataset to perform offline training and verification on the DAE-ALSTM model to obtain the trained DAE-ALSTM model.

[0084] The denoising autoencoder (DAE) of the present invention is an unsupervised neural network architecture that can be used for dimensionality reduction and feature selection. DAE consists of multiple layers of encoding, such as Figure 3 As shown in Figure 2, the encoding layer of the denoising autoencoder first artificially introduces a certain amount of noise into the original data input (RawInput). This noise can be random noise (e.g., Gaussian noise) or achieved by randomly setting some neurons to zero (Dropout), thereby obtaining a "corrupted" noisy input (CorruptedInput).

[0085] Introduce noise into the input data to obtain noise data. The formula is expressed as:

[0086]

[0087] Where Σ is the covariance matrix, which represents the correlation between noise in different dimensions, and ∈~N(0,Σ) indicates that the noise is sampled from a normal distribution.

[0088] The noisy data is then converted into a low-dimensional representation. This process allows the framework to learn a compressed representation of the input data, thereby being able to extract meaningful main features and reduce the dimensionality. The denoising autoencoder (DAE) is trained to extract key data features from the input data description and perform dimensionality reduction.

[0089] Input the noise data into the encoder to obtain the key feature data of dimensionality reduction. The formula is expressed as:

[0090]

[0091] Among them, Z represents the key feature data of dimensionality reduction, W enc represents the weight of the encoder, b enc represents the bias of the encoder, L represents the number of layers of the multi-layer encoder, γ iRepresents the regularization coefficient of each layer encoder, ReLU is the activation function applied to each residual module, W i is the weight in each layer of residual network, b i For each bias.

[0092] During training, the DAE learns to retain only the most important features of the input data. This results in lower-dimensional data containing the important features, which helps the ALSTM network continue to capture the characteristics of the input data and perform regression and classification analysis. One approach to dimensionality reduction using the DAE is to use a bottleneck representation as a set of new features. These features are typically lower dimensional than the original input data and capture the most important information. Unlike principal component analysis (PCA), the denoising autoencoder (DAE) supports nonlinear mapping, thereby improving the performance of classification and regression tasks.

[0093] like Figure 3 As shown, the DAE-ALSTM model also includes an ALSTM network, which includes multiple LSTM layers and an Attention layer.

[0094] From the perspective of indoor positioning of Wi-Fi wireless signals, the LSTM layer can be used to classify targets or predict their locations based on the RSSI size of nearby target access points that changes over time.

[0095] The Attention layer utilizes the attention mechanism to assign variable weights to different parts of the target RSSI data according to their relative importance, enabling the model to prioritize RSSI data from the most critical APs and adaptively weight the output according to importance.

[0096] The ALSTM (Attention-LSTM) network, which consists of an Attention layer and an LSTM layer, can improve the accuracy and robustness of classifying or regressing target locations in indoor environments.

[0097] The Attention layer introduces an attention mechanism that assigns variable weights to different features of the input data based on their relative importance to optimize the prediction accuracy of the target position. The following steps are included:

[0098] S31. Introduce an attention mechanism to focus on the most important signal sources and time steps of the input data for target position prediction;

[0099] S32. Calculate the attention weight based on the correlation between the input data and the access point position, indicating the relative importance of the time step to the target position prediction;

[0100] S33. According to the calculated attention weight, the predicted features of each time step output by the ALSTM model are weighted to obtain the optimized features. The formula is expressed as:

[0101] Z′ t =α t ·Z t

[0102] Among them, α t represents the attention weight, Z t Denotes the predicted features of the improved ALSTM model for time step t, Z′ t Represents the features after the attention mechanism is introduced for optimization.

[0103] S34, the weighted feature Z' t Input to the fully connected layer and output the target position coordinates (x, y). The formula is:

[0104] (x,y)=f regression (Z′)

[0105] Among them, Z′ represents the weighted feature set of all time steps, f regression Represents a regression model.

[0106] Connecting the DAE output to the ALSTM model input, combining the two to form a DAE-ALSTM model, yields better prediction results. The DAE model output serves as the input to the ALSTM model, reducing the RSSI features from 520 to 64 features and sequence units. The ALSTM network operates on 3D arrays, flattening the output and then connecting it to the output layer for localization. After passing through all layers in the ALSTM model, the final layer's dimensions are reduced to two, corresponding to coordinates x and y.

[0107] The present invention realizes nonlinear dimensionality reduction through a denoising autoencoder (DAE), compressing the 520-dimensional RSSI features to 64 dimensions, removing noise interference while retaining key features, and has stronger feature expression capabilities than the traditional PCA method.

[0108] At the same time, a Dropout layer is added between the LSTM layers to randomly discard some neurons during data processing, thereby regularizing the model and avoiding overfitting.

[0109] S4. Use the trained DAE-ALSTM model to perform online location prediction based on the target RSSI data.

[0110] The target RSSI data from unidentified locations is preprocessed and input into the DAE-ALSTM network for online location prediction.

[0111] like Figure 4 As shown, the present invention also provides a wireless indoor positioning device based on DAE-ALSTM, the device comprising:

[0112] Data construction module, used to collect coordinates and RSSI data of various points in the environment and build a signal route map;

[0113] A preprocessing module is used to preprocess the signal route map to obtain a data set;

[0114] The DAE-ALSTM module uses the DAE-ALSTM model to perform online location prediction based on target RSSI data.

[0115] like Figure 2 As shown, the device of the present invention forms a DALLoc architecture that combines a DAE with an attention-based LSTM framework to achieve wireless indoor positioning. The DAE uses noise coding to reduce dimensionality, extracting key features of the input data and improving its robustness; the ALSTM is used to achieve accurate target position prediction.

[0116] During the offline data collection phase, the data construction module collects target RSSI and location coordinate data from the target environment to construct a route map. The preprocessing module then preprocesses the signal route map to form an AP training dataset. The dataset is used to train the DAE-ALSTM model, which is then deployed in the DAE-ALSTM module. This trained DAE-ALSTM model is then used to perform online location prediction based on target RSSI data.

[0117] The present invention also provides an electronic device, Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 5 As shown, the electronic device may include: a processor, a communications interface, a memory, and a communication bus, wherein the processor, the communications interface, and the memory communicate with each other via the communication bus. The processor may call logic instructions in the memory, for example, to execute the following method:

[0118] S1. Deploy multiple AP access points in the target environment, collect coordinates and RSSI data of each point in the environment, and build a signal route map.

[0119] S2. Preprocess the signal route map to obtain a data set, and divide the data set into a training set and a validation set;

[0120] S3. Use the dataset to perform offline training and verification on the DAE-ALSTM model to obtain the trained DAE-ALSTM model.

[0121] S4. Use the trained DAE-ALSTM model to perform online location prediction based on the target RSSI data.

[0122] In addition, the logical instructions in the above-mentioned memory can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0123] An embodiment of the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method provided in each of the above embodiments is implemented, for example, including:

[0124] S1. Deploy multiple AP access points in the target environment, collect coordinates and RSSI data of each point in the environment, and build a signal route map.

[0125] S2. Preprocess the signal route map to obtain a data set, and divide the data set into a training set and a validation set;

[0126] S3. Use the dataset to perform offline training and verification on the DAE-ALSTM model to obtain the trained DAE-ALSTM model.

[0127] S4. Use the trained DAE-ALSTM model to perform online location prediction based on the target RSSI data.

[0128] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0129] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A wireless indoor positioning method based on DAE-ALSTM, characterized in that: include: S1. Deploy multiple AP access points in the target environment, collect coordinates and RSSI data of each point in the environment, and build a signal route map. S2. Preprocess the signal route map to obtain a data set, and divide the data set into a training set and a validation set; S3. Use the dataset to perform offline training and verification on the DAE-ALSTM model to obtain the trained DAE-ALSTM model. S4. Use the trained DAE-ALSTM model to perform online location prediction based on the target RSSI data.

2. The wireless indoor positioning method according to claim 1, wherein: The formula for constructing the signal route map is expressed as: X={(P1,RSSI1,(x1,y1)),...,(P i ,RSSI i ,(x i ,y i )),…,(P N ,RSSI N ,(x N ,y N ))} Among them, X represents the RSSI data of the point, P i is the i-th AP access point, RSSI i is the i-th AP access point at position (x i ,y i )’s signal receiving strength, and N represents the number of AP access points.

3. The wireless indoor positioning method according to claim 1, wherein: The DAE-ALSTM model includes a denoising autoencoder DAE, which includes multiple layers of encoding. The DAE processes data by: S21. Introduce noise into the input data to obtain noise data. The formula is expressed as: Where Σ is the covariance matrix, which represents the correlation between noise in different dimensions, and ε~N(0,Σ) means that the noise is sampled from a normal distribution; S22. Input the noise data into the encoder to obtain the key feature data of dimensionality reduction. The formula is expressed as: Among them, Z represents the key feature data of dimensionality reduction, W enc represents the weight of the encoder, b enc represents the bias of the encoder, L represents the number of layers of the multi-layer encoder, γ i Represents the regularization coefficient of each layer encoder, ReLU is the activation function applied to each residual module, W i is the weight in each layer of residual network, b i For each bias.

4. The wireless indoor positioning method according to claim 1, wherein: The DAE-ALSTM model includes an ALSTM network, which includes multiple LSTM layers and an Attention layer.

5. The wireless indoor positioning method according to claim 4, characterized in that: The Attention layer introduces an attention mechanism to assign variable weights to different features of the input data according to their relative importance, thereby optimizing the prediction accuracy of the target position. The following steps are included: S31. Introduce an attention mechanism to focus on the most important signal sources and time steps of the input data for target position prediction; S32. Calculate the attention weight based on the correlation between the input data and the access point position, indicating the relative importance of the time step to the target position prediction; S33. According to the calculated attention weight, the predicted features of each time step output by the ALSTM model are weighted to obtain the optimized features. The formula is expressed as: WITH' t =α t ·WITH t Among them, α t represents the attention weight, Z t Denotes the predicted features of the improved ALSTM model for time step t, Z′ t Indicates the features after the attention mechanism is introduced; S34, the weighted feature Z' t Input to the fully connected layer, perform classification regression, and obtain the target position coordinate (x, y) output. The formula is expressed as: (x,y)=f regression (Z′) Among them, Z′ represents the weighted feature set of all time steps, f regression Represents a regression classification function.

6. The wireless indoor positioning method according to claim 1, characterized in that: A Dropout layer is provided between the multiple LSTM layers to randomly discard some neurons during data processing to reduce overfitting.

7. A wireless indoor positioning device based on DAE-ALS™ applied to the method according to any one of claims 1 to 6, characterized in that: The device comprises: Data construction module, used to collect coordinates and RSSI data of various points in the environment and build a signal route map; A preprocessing module is used to preprocess the signal route map to obtain a data set; The DAE-ALSTM module uses the DAE-ALSTM model to perform online location prediction based on target RSSI data.

8. 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, the steps of the wireless indoor positioning method according to any one of claims 1 to 5 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the wireless indoor positioning method according to any one of claims 1 to 5 are implemented.

Images