Beidou pseudolite indoor positioning system multi-level credibility evaluation method

Abstract

This invention discloses a multi-level reliability evaluation method for BeiDou pseudosatellite indoor positioning systems, belonging to the field of pseudosatellite indoor positioning technology. This method can evaluate the reliability of indoor positioning systems in real time, solving the problem of unreliable positioning results caused by poor pseudosatellite data quality and external environmental interference, thus ensuring the reliability requirements of high-end applications based on indoor location services. The system includes a reliability evaluation model and a geographic prior information evaluation strategy, evaluating from both the data layer and the result layer, providing a relatively comprehensive reliability evaluation mechanism. Furthermore, the proposed method can dynamically evaluate the reliability of the location in real time, demonstrating considerable technical difficulty and innovation from both theoretical research and engineering application perspectives. Experimental results show that adding the reliability evaluation analysis method improves the average positioning accuracy by 74.6% and the reliability by 88.2%, verifying the effectiveness of the method described in this patent.

Description

Technical Field

[0001] This invention relates to the field of pseudo-satellite indoor positioning and navigation technology, and in particular to a multi-level reliability evaluation method for BeiDou pseudo-satellite indoor positioning systems. Background Technology

[0002] The research and industrial application of indoor positioning technology are booming both domestically and internationally, with new technologies and applications constantly emerging. As a supplement to outdoor navigation satellites, pseudosatellites have the ability to transmit the same signals as navigation satellites, providing stable and reliable navigation and positioning signals for indoor environments. This allows for seamless indoor and outdoor positioning and navigation capabilities based on existing navigation receivers, meeting the needs of most scenarios and applications. However, as the application areas of pseudosatellite positioning systems continue to expand, more and more users are beginning to focus on the reliability indicators of positioning systems to ensure the reliability of location service capabilities.

[0003] However, the performance of BeiDou pseudosatellite indoor positioning systems is often unreliable due to poor data quality and external environmental interference, making it difficult to meet the needs of high-end applications. Moreover, in practical applications, the lack of a reliability evaluation mechanism poses risks to subsequent missions, making it difficult for pseudosatellite-based indoor positioning to be widely used. Summary of the Invention

[0004] In view of this, the present invention proposes a multi-level reliability evaluation method for BeiDou pseudo-satellite indoor positioning systems. This method can evaluate the reliability of pseudo-satellite indoor positioning systems in real time, providing reliability assurance for positioning systems in location-based service applications.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A multi-level reliability evaluation method for a BeiDou pseudo-satellite indoor positioning system includes the following steps:

[0007] (1) In an indoor environment where a Beidou pseudo-satellite positioning system is deployed, pseudo-satellite observation data are collected using a navigation receiver to construct a training dataset;

[0008] (2) Design a reliable evaluation model for indoor positioning of pseudo-satellites, and train the model using the training dataset constructed in step (1). Use the trained model to evaluate the reliability of pseudo-satellite observation data from the data level. When the evaluation result exceeds the set threshold, the corresponding pseudo-satellite observation data is unreliable data.

[0009] (3) Extract prior information of building structure from the map in the indoor environment, optimize the weight of the particle filter method, and use the optimized particle filter method to evaluate the credibility of the positioning results from the result level.

[0010] Furthermore, the specific method of step (1) is as follows:

[0011] In an indoor environment, various types of navigation receivers are used to collect pseudo-satellite observation data in the area to be located. The collected dataset includes timestamps, carrier phase, carrier-to-noise ratio, pseudorange, and satellite number information.

[0012] During the data collection process, the location results are observed in real time for any jumps or drifts. Data within the time period corresponding to the jumps or drifts are considered unreliable datasets, while other data are considered reliable datasets. The time period is a fixed time window.

[0013] Furthermore, the specific method of step (2) is as follows:

[0014] (201) Design a variational autoencoder (VAE) model for pseudosatellite data quality assessment. The VAE model is a probabilistic graphical model with an encoder and a decoder, which are used for the extraction of deep features of pseudosatellite data and the generation of new data, respectively.

[0015] (202) Set a sliding window to divide the trusted dataset into multiple subsequences x. i , i∈[1,N], where N is the window number, and each subsequence x i Both are used as input to the VAE model. The encoder maps the data to latent features Z, and the decoder reconstructs the data x from the latent features Z. i ′, and by minimizing the reconstructed data x i ′ and input data x i The differences were analyzed to obtain a credibility evaluation model, which was then used to assess the credibility of pseudosatellite observation data from a data perspective.

[0016] (203) The reconstruction probability threshold of the credible assessment is determined by offline positioning experiment. During real-time assessment, when the reconstruction probability of pseudo-satellite observation data in the window exceeds the reconstruction probability threshold after passing through the credible assessment model, the pseudo-satellite observation data in the corresponding window is regarded as unreliable data.

[0017] Furthermore, step (3) is specifically implemented as follows:

[0018] By obtaining location-related structural information in the environment through building maps, including corridors, walls, and doors and windows, the acquired information is used as the basis for weight updates in the particle filter method. The updated particle filter method is then used to evaluate the reliability of the localization results from the result level.

[0019] The beneficial effects of this invention compared to the prior art are as follows:

[0020] (1) This invention proposes a multi-level reliability evaluation method for Beidou pseudo-satellite indoor positioning system, which can provide location reliability evaluation for pseudo-satellite-based indoor positioning system and ensure the reliability of location-based services.

[0021] (2) The multi-level credibility evaluation method proposed in this invention includes a variational autoencoder credibility evaluation model and a geographic prior information evaluation strategy, which evaluate from the data layer and the result layer respectively, and is a relatively comprehensive credibility evaluation mechanism.

[0022] (3) The reliability evaluation system designed and proposed in this invention can dynamically evaluate the reliability of a location in real time without a real reference value. This places high demands on the real-time performance, reliability and accuracy of the technology. Therefore, this location reliability evaluation technology has certain technical difficulties and innovations in terms of theoretical research and engineering application. Attached Figure Description

[0023] Figure 1 This is a flowchart of the BeiDou pseudo-satellite dataset processing in this patent embodiment.

[0024] Figure 2 This is a flowchart of the VAE credibility evaluation model in an embodiment of the present invention.

[0025] Figure 3 This is a trajectory diagram of an experiment to verify the effectiveness of the credible evaluation method in this embodiment of the invention. Detailed Implementation

[0026] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0027] A multi-level reliability evaluation method for a BeiDou pseudo-satellite indoor positioning system includes the following steps:

[0028] (1) In an indoor environment, use a commercial navigation receiver to collect pseudo-satellite observation data in the area to be located, and construct a training dataset; such as Figure 1 As shown, specifically:

[0029] Step 1: To ensure dataset diversity, various types of commercial receivers are selected for acquisition. The dataset includes timestamps, carrier phase, carrier-to-noise ratio, pseudorange, and the number of satellites received. Step 2: After acquiring the raw observation data, preprocessing is performed to obtain an intermediate data format containing position labels and inter-satellite differences, improving the convergence speed during model training. Step 3: During the acquisition process, real-time observation is conducted to check for jumps and drifts in the positioning results. Data within time periods exhibiting jumps and drifts are designated as untrusted datasets, while other data are designated as trusted datasets. A portion of the trusted data is used as the training set, and the other portion as the test set; the time periods are fixed time windows.

[0030] (2) Design a reliable evaluation model for indoor pseudo-satellite positioning, and train the model using the training set constructed in step (1). Use the trained model to evaluate the reliability of pseudo-satellite observation data in the test set from a data perspective. When the evaluation result exceeds a set threshold, the corresponding pseudo-satellite observation data is considered unreliable. Specifically, in the offline stage, the distribution model of pseudo-satellite signals in the environment is obtained by analyzing the dataset described in step (1). Generally speaking, after the model training is completed, the reconstruction probability of normal data is relatively low, while the reconstruction probability of abnormal data is relatively high. In view of this, by setting a reconstruction probability threshold to identify abnormal data, the impact of abnormal data caused by environmental factors and equipment factors can be effectively reduced, thereby improving the reliability of the positioning system. The workflow of the reliable evaluation model is as follows: Figure 2 As shown. The specific method is as follows:

[0031] (201) Design a variational autoencoder (VAE) model for pseudosatellite data quality assessment. The VAE model is a probabilistic graphical model with an encoder and a decoder, which are used for the extraction of deep features of pseudosatellite data and the generation of new data, respectively.

[0032] (202) Set a sliding window to divide the trusted dataset into multiple subsequences x. i , i∈[1,N], where N is the window number, and each subsequence x i Both are used as input to the VAE model. The encoder maps the data to latent features Z, and the decoder reconstructs the data x from the latent features Z. i ′, and by minimizing the reconstructed data x i ′ and input data x i The differences were analyzed to obtain a credibility evaluation model, which was then used to assess the credibility of pseudosatellite observation data from a data perspective.

[0033] (203) The reconstruction probability threshold of the credible assessment is determined by offline positioning experiment. During real-time assessment, when the reconstruction probability of pseudo-satellite observation data in the window exceeds the reconstruction probability threshold after passing through the credible assessment model, the pseudo-satellite observation data in the corresponding window is regarded as unreliable data.

[0034] (3) Obtain location-related structural information in the environment through building maps, including corridors, walls, and doors and windows. Use this information as the basis for weight updates in the particle filter method, and then use the updated particle filter method to evaluate the reliability of the positioning results from the result level. In actual positioning, the actual location of the target will not be subject to unreliable situations such as "passing through walls". Therefore, in the weight update stage, real geographic environment information is considered as the basis for reliability evaluation to further improve the reliability of the positioning system.

[0035] The above steps achieved a multi-level positioning reliability evaluation analysis. To verify the effectiveness of the evaluation method on a practical pseudosatellite positioning system, a positioning experiment was conducted in an experimental environment with and without the participation of the reliability evaluation method. The experimental results are as follows: Figure 3 As shown, the square marker trajectory represents the pseudo-satellite positioning trajectory without a credible evaluation method, the circular marker trajectory represents the positioning trajectory after adding a VAE credible evaluation model, and the triangular marker trajectory represents the positioning trajectory after adding geographic prior information. The horizontal and vertical coordinates represent the actual dimensions in the test environment, in meters. After adding the credible evaluation method, the positioning accuracy and reliability of the pseudo-satellite positioning system were significantly improved. The average positioning error was 0.56m, the maximum positioning error was 1.25m, and 95.2% of the errors were less than 1m. Compared with the positioning results without the credible evaluation, the average positioning accuracy improved by 74.6%. The percentage of the total observation time in a specified area within a certain period of time is the time when the system positioning accuracy is less than the reliability threshold. When the reliability threshold is set to 1m, the positioning reliability improved by 88.2% after adding the credible evaluation, verifying the effectiveness of the credible evaluation method described in this patent.

[0036] In summary, the multi-level reliability assessment method proposed in this invention solves the problem of decreased positioning reliability caused by interference from factors such as the sensor itself and the environment when pseudo-satellite indoor positioning observation data are easily affected. It evaluates positioning results from multiple aspects, including pseudo-satellite data quality assessment and weight update strategy assisted by environmental prior information, to achieve multi-level evaluation at the data layer and result layer. This further improves the positioning accuracy and reliability of the positioning results, provides reliability assurance for high-end applications of location-based services, and has high application value and promotion significance.

Claims

1. A multi-level reliability evaluation method for a BeiDou pseudo-satellite indoor positioning system, characterized in that, Includes the following steps: (1) In an indoor environment where a Beidou pseudo-satellite positioning system is deployed, pseudo-satellite observation data is collected using a navigation receiver to construct a training dataset; (2) Design a reliable evaluation model for indoor positioning of pseudo-satellites, and train the model using the training dataset constructed in step (1). Use the trained model to evaluate the reliability of pseudo-satellite observation data from the data level. When the evaluation result exceeds the set threshold, the corresponding pseudo-satellite observation data is unreliable data. (3) Extract prior information of building structure from the map in the indoor environment, optimize the weight of the particle filter method, and use the optimized particle filter method to evaluate the credibility of the positioning results from the result level. The specific method of step (2) is as follows: (201) Design a variational autoencoder (VAE) model for pseudosatellite data quality assessment. The VAE model is a probabilistic graphical model with an encoder and a decoder, which are used for the extraction of deep features of pseudosatellite data and the generation of new data, respectively. (202) Set a sliding window to divide the trusted dataset into multiple subsequences. , For each subsequence, the number of windows is [number]. All of these are used as input to the VAE model, and the encoder maps the data to latent features. Then, the decoder is used to extract the hidden features. Reconstructing into data And by minimizing the reconstructed data With input data The differences were analyzed to obtain a credibility evaluation model, which was then used to assess the credibility of pseudosatellite observation data from a data perspective. (203) The reconstruction probability threshold of the credible assessment is determined by offline positioning experiment. During real-time assessment, when the reconstruction probability of pseudo-satellite observation data in the window exceeds the reconstruction probability threshold after passing through the credible assessment model, the pseudo-satellite observation data in the corresponding window is regarded as unreliable data. The specific implementation method of step (3) is as follows: By obtaining location-related structural information in the environment through building maps, including corridors, walls, and doors and windows, the acquired information is used as the basis for weight updates in the particle filter method. The updated particle filter method is then used to evaluate the reliability of the localization results from the result level.

2. The multi-level reliability evaluation method for a BeiDou pseudo-satellite indoor positioning system according to claim 1, characterized in that, The specific method for step (1) is as follows: In an indoor environment, various types of navigation receivers are used to collect pseudo-satellite observation data in the area to be located. The collected dataset includes timestamps, carrier phase, carrier-to-noise ratio, pseudorange, and satellite number information. During the data collection process, the location results are observed in real time to see if there are any jumps or drifts. Data in the time period corresponding to the jumps and drifts are regarded as untrusted datasets, while other data are regarded as trusted datasets. The time period is a fixed time window.

Images