Cooperative positioning method and apparatus, computer device, and computer readable storage medium

Abstract

The embodiment of the application provides a cooperative positioning method, device, equipment and computer readable storage medium, the method comprises: for the mobile terminal to be positioned in the long and narrow space, using the enabled AP in the effective ranging range of the mobile terminal, the position of the mobile terminal is preliminarily predicted based on the robust adaptive Kalman filtering algorithm;The result of the preliminary prediction is adjusted based on the offset function and the fluctuation function of each AP, and the adjustment result is used as the final positioning result of the mobile terminal;Wherein, the offset function and the fluctuation function of the AP are obtained by pre-fitting according to historical data. The embodiment of the application can optimize the positioning performance in the linear coverage of the positioning AP in the long and narrow terrain, and improve the accuracy and reliability of positioning.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a cooperative positioning method, apparatus, computer device, and computer-readable storage medium. Background Technology

[0002] In open outdoor environments, targets can obtain high-precision positions using the Global Navigation Satellite System (GNSS). However, in narrow, underground indoor environments such as underground utility tunnels, coal mine roadways, and tunnels, GNSS suffers severe degradation or complete failure because satellite signals cannot penetrate the ground. Other outdoor positioning technologies also struggle to adapt well to indoor environments. Currently, achieving location awareness in these narrow indoor spaces through local positioning technology to meet the requirements of automated and intelligent operations has become a hot topic and cutting-edge issue in the field of indoor positioning technology.

[0003] Currently, narrow indoor spaces often rely on linear AP (Access Point) deployments to achieve wireless communication network coverage. However, deploying additional positioning systems requires a trade-off between economic costs, positioning range, and positioning performance. Therefore, leveraging the ranging and positioning potential of APs while providing communication capabilities to achieve low-cost, COTS (Commercial-Off-The-Shelf) one-dimensional positioning technology is of great practical significance.

[0004] WiFi-based positioning technologies have made significant progress in recent years, including RSSI ranging, location fingerprinting, and FTM ranging. RSSI ranging calculates the distance between a target node and a reference node by utilizing the statistical relationship between the intensity attenuation of electromagnetic wave signals during propagation and the signal propagation distance. It typically employs a log-normal distance or a multi-segment model. The advantages of RSSI measurement methods are low power consumption and simple physical implementation. Location fingerprinting is an indirect ranging technology. Its algorithm consists of two stages: the first stage collects fingerprint information (such as RSSI values), and the second stage matches the fingerprint information to locate the target.

[0005] FTM (Fine Time Measurement) is a new method of ranging based on Wi-Fi signals, added in the 802.11mc protocol adopted in December 2016. Its core idea is to calculate the distance by combining the signal flight time between the two parties and the signal's velocity. Calculating the signal flight time requires recording the signal transmission and arrival times of both parties. The FTM ranging result can be described as follows: ;

[0006] In the formula, The ranging result for RTT; The ranging error is caused by the standard time deviation; Ranging error caused by non-line-of-sight (NLOS) propagation. Ranging error caused by standard time deviation. It includes not only the fixed time delay error of the pulse signal in the rover and base station, the device error (of different terminal equipment) and the successive start error, but also the external environment such as the distance and temperature of the pulse signal propagation.

[0007] The Kalman filter algorithm, using state equations as its mathematical tool, is one of the important methods for dynamic data processing and has been widely applied in the field of dynamic navigation and positioning. Among them, the robust adaptive Kalman filter (RAKF) algorithm, with its advantages of high flexibility and robustness, has shown good results in satellite navigation and positioning. The core of the algorithm lies in evaluating two types of anomalies: when anomalies exist in the observations, a robust estimation process is performed; when anomalies exist in the system state model, a uniform adaptive factor is used to adjust the contribution of the system state to the final result.

[0008] While RSSI ranging and positioning is relatively simple to implement, the electromagnetic wave propagation environment is much more complex in narrow indoor spaces used for engineering operations. Multipath effects are more severe, the attenuation of electromagnetic field strength is harder to predict, and signal strength is affected by factors such as shadowing fading and multipath effects, resulting in poor ranging accuracy. Furthermore, the relationship between electromagnetic wave propagation distance and signal reception strength is non-linear; when the positioning distance exceeds a certain range, the accuracy of the RSSI ranging method deteriorates. Since access points (APs) in narrow indoor spaces are typically widely distributed, this method is unsuitable for precise personnel positioning in such scenarios.

[0009] In summary, existing technologies have poor positioning accuracy in narrow terrains and larger errors when there are obstacles or when entering non-line-of-sight (NLOS) areas. Summary of the Invention

[0010] The purpose of this application is to provide a cooperative positioning method, apparatus, computer device, and computer-readable storage medium that can optimize positioning performance under linear coverage conditions of positioning access points in narrow terrain, and improve the accuracy and reliability of positioning; especially for scenarios with unpredictable NLOS areas caused by environmental obstacles, it can improve the accuracy and reliability of positioning.

[0011] One aspect of this application provides a cooperative localization method, including:

[0012] For a mobile terminal to be located in a narrow space, the position of the mobile terminal is initially predicted using an AP enabled within the effective ranging range of the mobile terminal, based on a robust adaptive Kalman filter algorithm.

[0013] The preliminary prediction results are adjusted based on the offset function and fluctuation function of each AP, and the adjusted results are used as the final positioning results of the mobile terminal.

[0014] The offset function and fluctuation function of the AP are pre-fitted based on historical data. The AP activated within the effective ranging range of the mobile terminal performs a preliminary prediction of the mobile terminal's position based on a robust adaptive Kalman filter algorithm, specifically including:

[0015] for The constructed Kalman filter is shown in Equation 1:

[0016] (Formula 1)

[0017] in, , This represents the set of all enabled APs within the effective ranging range of the mobile terminal; for The predicted mobile terminal k The motion state vector at time t, including k Position, velocity, and acceleration at any given moment; for The state transition matrix from the previous state to the next state; for of k The system error vector at time t; for of k The observation location at that moment; for of k The coefficient matrix at time step; for of k The observation noise vector at time step;

[0018] in, ; for of k The distance measurement result at any given time is a scalar. for of k The distance measurement direction vector remains unchanged within a single subinterval. for Position coordinates;

[0019] The preliminary prediction results are calculated according to the following formula 2:

[0020] (Formula 2)

[0021] in, for of k Preliminary predictions for the time. for of k The covariance matrix at time t; for of k The optimal estimate of the state at time -1 for of k The state estimation covariance matrix at time -1 for of k Observe the noise covariance at all times.

[0022] Preferably, the adjustment of the preliminary prediction results based on the offset function and fluctuation function for each AP specifically includes:

[0023] According to the following formula 7 of k Preliminary prediction results of time Adjustments were made to obtain Optimal prediction estimation results :

[0024] (Formula 7)

[0025] in , ; for of k The robustness factor at time; for of k The adaptive factor at time, according to The offset function and the fluctuation function are obtained; for of k The observation noise covariance matrix at time step;

[0026] according to Optimal estimated position The final location result is obtained: ;

[0027] in, ;in, express The fluctuation function.

[0028] Furthermore, after making a preliminary prediction of the location of the mobile terminal, the method further includes:

[0029] Calculated according to the following formula 3 New information vector and the new information covariance matrix :

[0030] (Formula 3)

[0031] in, express of k The observation noise covariance matrix at time t.

[0032] Among them, the Specifically, it is calculated according to the following formula 4:

[0033] (Formula 4)

[0034] in, , , This is a preset constant.

[0035] Among them, the Specifically, it is calculated using the following method:

[0036] according to For each AP k The observation position at any given time has been adjusted. ;

[0037] in, express The offset function of each AP in the mobile terminal is located at the same position. Combination of function values ​​at the location; express Each AP in China k Combination of distance measurement direction vectors at any given time; express Each AP in China k The combination of observation positions at any given time; Indicates the adjusted Each AP in China k The combination of observation positions at any given time;

[0038] according to Obtain the standardized weight vector:

[0039] ;

[0040] in, express The oscillation function of each AP in Combination of function values ​​at the location; ;

[0041] Unify the ranging values ​​of all APs to obtain the updated values ​​for each AP. k Combination of observation positions at different times: ;

[0042] This leads to the updated observation noise covariance matrix: ;in, Weight vector The Item value;

[0043] The updated innovation vector is obtained according to Formula 5 below. New covariance matrix and new information statistics :

[0044] (Formula 5)

[0045] Adaptive factor matrix Determined according to the following formula 6:

[0046] (Formula 6)

[0047] in, .

[0048] One aspect of this application provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the cooperative localization method described above.

[0049] One aspect of this application provides a computer-readable storage medium, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the cooperative localization method described above.

[0050] One aspect of this application provides a cooperative positioning device, including:

[0051] The preliminary positioning module is used to make a preliminary prediction of the position of a mobile terminal to be located in a narrow space by using an AP enabled within the effective ranging range of the mobile terminal and a robust adaptive Kalman filter algorithm.

[0052] The positioning adjustment module is used to adjust the preliminary prediction results based on the offset function and fluctuation function of each AP, and use the adjusted results as the final positioning result of the mobile terminal; wherein, the offset function and fluctuation function of the AP are pre-fitted based on historical data.

[0053] The collaborative positioning method, apparatus, device, and computer-readable storage medium provided in this application, for a mobile terminal to be located in a narrow space, uses the access points (APs) enabled within the effective ranging range of the mobile terminal to make a preliminary prediction of the mobile terminal's position based on a robust adaptive Kalman filter algorithm; the preliminary prediction result is adjusted based on the offset function and fluctuation function of each AP, and the adjusted result is used as the final positioning result of the mobile terminal; wherein, the offset function and fluctuation function of the AP are pre-fitted based on historical data. Because the pre-fitted offset function and fluctuation function of the APs, based on the robust adaptive Kalman filter algorithm, improves the ability to resist abnormal ranging values ​​under conditions of sparse ranging AP coverage, thus improving positioning accuracy, reliability, and precision; especially for unpredictable NLOS areas caused by environmental obstacles in the scene, it improves the accuracy and reliability of positioning.

[0054] Furthermore, the technical solution of this invention performs robust estimation after making a preliminary prediction of the location of the mobile terminal, and predicts the covariance matrix adaptively. The robust estimation process and adaptive process of RAKF are optimized in the case of low-dimensional observation data, so as to further improve the positioning accuracy, precision and reliability.

[0055] Furthermore, by using measurement methods based on non-survey data for ranging bias (offset function) and fluctuation (fluctuation function), more robust outlier detection and processing during movement can be achieved, thereby optimizing the navigation and positioning performance of robust adaptive Kalman filtering and making the technology more adaptable to practical narrow terrain applications. Attached Figure Description

[0056] Figure 1 This illustration schematically shows a linear deployment scenario of AP according to an embodiment of this application;

[0057] Figure 2 A flowchart illustrating a cooperative positioning method for narrow spaces according to an embodiment of this application is shown schematically.

[0058] Figure 3 The schematic diagram illustrates the simulation experiment results according to an embodiment of this application;

[0059] Figure 4 This schematic diagram illustrates the internal structure of a cooperative positioning device according to an embodiment of the present application;

[0060] Figure 5 The illustration shows a schematic diagram of the hardware architecture of a computer device suitable for implementing a cooperative positioning method applicable to narrow spaces, according to an embodiment of this application. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0062] It should be noted that the descriptions involving "first," "second," etc., in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0063] In the description of this application, it should be understood that the numerical labels before the steps do not indicate the order of the steps, but are only used to facilitate the description of this application and to distinguish each step, and therefore should not be construed as a limitation of this application.

[0064] The inventors of this invention considered that long and narrow underground indoor spaces face numerous technical challenges compared to general indoor spaces: First, there is the problem of sparse positioning base stations. Long and narrow underground spaces are characterized by a length much greater than their width and are either enclosed or semi-enclosed. This elongated geometric characteristic means that the geometry formed by reference nodes is often close to a straight line. Due to the limited quantity and low dimensionality of available positioning data, it is difficult to achieve reliable two-dimensional and three-dimensional positioning of targets within the area using only mid-to-long-range wireless technology. Deploying a large number of positioning base stations or tags to achieve full coverage and high-precision positioning would inevitably lead to extremely high deployment and maintenance costs, and this requirement often lacks practical significance. Second, the environment in long and narrow terrain areas is usually more complex. The electromagnetic wave propagation environment differs from that in open space. In such scenarios, electromagnetic radiation and obstacles have a more severe impact on positioning signals, exhibiting significant multipath effects and non-line-of-sight effects. Third, due to economic factors, the current deployment interval of positioning base stations in long and narrow terrains is as long as hundreds of meters, resulting in significant signal attenuation, severe electromagnetic interference, and a high likelihood of significant deviations in remote positioning results. These factors all pose challenges to precise positioning in narrow spaces, and current research on one-dimensional positioning technology is limited by positioning accuracy, manual complexity, and deployment cost, which is insufficient to meet the needs of engineering operations and daily applications.

[0065] For environments with sparse AP coverage, this invention ignores the cross-section of narrow terrain areas and abstracts the positioning scene into a one-dimensional model.

[0066] like Figure 1 As shown, the location network includes several WiFi 802.11mc APs and WiFi 802.11mc mobile terminals. Considering the most economical coverage scheme, the APs are deployed linearly in the scene, with relative position coordinates as follows: As the target moves, its position is observed by APs within the effective ranging range. AP ranging results are independent of each other, and the effective ranging radius is... Assuming all AP ranging intervals are the same, and using this as the basis for time slot division. The ranging process of each AP is carried out in these discrete time slots, starting from the beginning of each time slot and obtaining the result at the end of the time slot.

[0067] Define a set The effective ranging range includes the location of the mobile terminal to be located. All enabled AP sets that satisfy The total area is divided into several sub-segments according to the linearly deployed AP locations and their respective effective ranging radii. Each sub-segment corresponds to the same group of APs responsible for positioning, satisfying... .

[0068] Assume that there are unpredictable NLOS areas in the scene due to environmental obstacles, and the overall scene is a LOS / NLOS mixed scene. The different positions of each AP result in different positioning environments (LOS / NLOS situations).

[0069] This invention addresses the aforementioned narrow terrain environment by proposing a cooperative localization method based on robust adaptive Kalman filtering (SC-RAKF). Employing the concept of federated Kalman filtering, each base station maintains its own state equation, and the final localization result is determined jointly by the optimal state estimates of all base stations. Simultaneously, a ranging state metric method incorporating historical data is proposed to provide environment-based dynamic weights for the cooperative localization process and drive robust estimation and adaptive processes under constrained conditions.

[0070] The technical solutions of the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0071] Figure 2 The schematic diagram illustrates a specific flowchart of a cooperative localization method for narrow spaces according to an embodiment of this application, including the following steps:

[0072] Step S201: For a mobile terminal to be located in a narrow space, use all enabled APs within the effective ranging range of the mobile terminal to make a preliminary prediction of the mobile terminal's position based on a robust adaptive Kalman filter algorithm.

[0073] Specifically, when approximating the trajectory of a moving target through a series of uniform linear motions, a Kalman filter can be used to iteratively update the state of the discrete-time controlled linear dynamic system. One-dimensional localization differs from two-dimensional and three-dimensional localization in that single AP ranging information, which does not include direction, can be completely indicated by offset transformation based on prior information. Based on this characteristic, the SC-RAKF algorithm of this invention can be applied to single sub-intervals... Each AP within (denoted as) Construct Kalman filters respectively; for The constructed Kalman filter is shown in Equation 1:

[0074] (Formula 1)

[0075] in, , This represents the set of all enabled APs within the effective ranging range of the mobile terminal; for The predicted mobile terminal k The motion state vector at time t, including k Position, velocity, and acceleration at any given moment; for The state transition matrix from the previous state to the next state; for of k The system error vector at time t; for of k The observation location at that moment; for of k The coefficient matrix at time step; for of k The observation noise vector at time step;

[0076] in, ; for of k The distance measurement result at any given time is a scalar. for of k The distance measurement direction vector remains unchanged within a single subinterval. for Position coordinates;

[0077] The preliminary prediction results are calculated according to the following formula 2:

[0078] (Formula 2)

[0079] in, for of k Preliminary predictions for the time. for of k The covariance matrix at time t; for of k The optimal estimate of the state at time -1 for of k The state estimation covariance matrix at time -1 for of k Observe the noise covariance at all times.

[0080] Step S202: Perform robust estimation based on the preliminary prediction results, and predict adaptive estimation of the covariance matrix.

[0081] Specifically, the update phase of the SC-RAKF in this invention involves two parts: robust estimation of the observation vector and adaptive estimation of the prediction covariance matrix. In this invention, robust estimation is achieved through a robust factor. To mitigate the impact of NLOS and other ranging anomalies on positioning results, the variance matrix of outlier observations is expanded. Adaptive factors are used for adaptive estimation of the predictive covariance matrix. Reduce the impact of abnormal dynamic model information on the filtering state.

[0082] Based on the preliminary prediction results, the following formula 3 is used to calculate: New information vector and the new information covariance matrix :

[0083] (Formula 3)

[0084] in, express of k The observation noise covariance matrix at time t.

[0085] The innovation vector contains not only observation model information at the current moment but also dynamic model information, thus it can serve as an important indicator for determining whether there are anomalies in the entire Kalman filter model. In this scenario model, the cause of anomalies in the innovation vector is difficult to determine accurately, but there are two main possible reasons: drastic changes in motion state or increased ranging error. However, in real-world environments, AP can obtain relatively high-frequency ranging data using FTM. Within a short timeframe, the changes in the motion state of the target to be located are limited, making observation anomalies caused by NLOS (Neutral Noise Loss) more concerning than errors in the motion state equation.

[0086] Therefore, the SC-RAKF of this invention first performs a robust estimation process, constructs standardized innovation values, and uses the IGGIII scheme to calculate the robustness factor. The calculation is shown in Formula 4:

[0087] (Formula 4)

[0088] in, , , This is a preset constant.

[0089] When there are large abnormal disturbances in the dynamic model (such as sudden acceleration or reversal of personnel), robust estimation based solely on the observations cannot guarantee the robustness of the final positioning result. This is mainly because the system noise setting fails to adapt to the changes in the dynamic model, thus affecting the final positioning result.

[0090] , An approximate ranging state for each AP at any location is given, and this prior information can be used to process the AP ranging data and apply it to the adaptive phase. The SC-RAKF adaptive factor calculation process of this invention will coordinate... The ranging status metric of all APs is performed as follows:

[0091] according to For each AP k The observation position at any given time has been adjusted. ;

[0092] in, express The offset function of each AP in the mobile terminal is located at the same position. Combination of function values ​​at the location; express Each AP in China k Combination of distance measurement direction vectors at any given time; express Each AP in China k The combination of observation positions at any given time; Indicates the adjusted Each AP in China k The combination of observation positions at any given time;

[0093] according to Obtain the standardized weight vector:

[0094] ;

[0095] in, express The oscillation function of each AP in Combination of function values ​​at the location; ;

[0096] Unify the ranging values ​​of all APs to obtain the updated values ​​for each AP. k Combination of observation positions at different times: ;

[0097] This leads to the updated observation noise covariance matrix: ;in, Weight vector The Item value;

[0098] The updated innovation vector is obtained according to Formula 5 below. New covariance matrix and new information statistics :

[0099] (Formula 5)

[0100] Adaptive factor matrix Determined according to the following formula 6:

[0101] (Formula 6)

[0102] in, .

[0103] After using the robust factor and adaptive factor obtained from the above process, the covariance matrix of the optimal estimate and state estimate of each AP state can be calculated in subsequent steps to update the state equation.

[0104] Step S203: Adjust the preliminary prediction results based on the offset function and fluctuation function of each AP, and use the adjusted results as the final positioning results of the mobile terminal.

[0105] The offset function and fluctuation function of the AP represent the expected ranging error and fluctuation degree of the AP estimated from historical data, respectively, and are obtained by pre-fitting; the specific fitting method will be described in detail later.

[0106] Specifically, the RAKF calculation principle is as follows: Therefore, we can use the following formula 7 to... Preliminary prediction results at time k Adjustments were made to obtain Optimal prediction estimation results :

[0107] (Formula 7)

[0108] in , ; for of k The robustness factor at time; for of k The adaptive factor at time is based on The offset function and the fluctuation function are obtained;

[0109] according to Optimal estimated position The final location result is obtained: ;

[0110] in, ;in, express The fluctuation function.

[0111] Since NLOS (Normally Occurring Loss) regions within the scene can cause a positive offset to the ranging results, accurately and sensitively identifying this situation during terminal movement can significantly optimize positioning accuracy. However, the degree of influence varies among different NLOS regions. Furthermore, the volatility of the ranging data itself and the actual ranging frequency of the device also affect the reliability of NLOS state recognition, which cannot be measured by establishing a unified distribution model.

[0112] Based on the RAKF prediction and robust estimation stages, SC-RAKF uses a method that does not require manual location calibration to quantify the error fluctuation of AP ranging results, and then calculates the adaptive factor and the final positioning result.

[0113] for Define an offset function with respect to the relative position of the AP. , wave function , respectively representing estimates obtained from historical data In position The expected ranging error and its fluctuation. A set of metrics functions is maintained for each AP, defined within the effective ranging range of that AP. (Unindexed) , Then it means AP in China The combination of function values ​​at that point.

[0114] The following section uses unsupervised historical data to develop a metric function. , Fitting.

[0115] Ideally, the fluctuation in ranging error reflected by the metric function can be measured by the average offset between the actual ranging result and the true position, as well as the variance of the fluctuation. This approach can be combined with historical measurement data to fit a specific numerical value. However, this method faces the following two problems:

[0116] 1. How to simulate the target's true location. Unless the location is manually marked, the target's true location cannot be known at the time of ranging.

[0117] 2. How to extend discrete ranging data to a continuous interval. The final generated NLOS curve must contain information for every point within the interval.

[0118] The present invention proposes the following solutions to the above problems:

[0119] In addition to ranging data, the preliminary predicted location is introduced by incorporating the system state equation. This invention assists in the judgment process. Furthermore, it selects appropriate ranging data for each real location, rather than calibrating the real location for the ranging data.

[0120] Define historical data format: ;in Let N be the set of ranging APs. The number of APs in the middle, for A joint ranging data vector of the internal AP. To and The corresponding preliminary predicted location combination.

[0121] gather Includes all valid ranging ranges including locations The historical data set of AP.

[0122] Next, based on the preliminary prediction vector For sets Further filtering is performed to obtain the filtered set. : ;

[0123] in , They are vectors The minimum and maximum values ​​of the elements in the set. .

[0124] Thus, we obtain the position defined on a continuous interval. The historical data set contains all historical data within the coverage area of ​​the one-step prediction vector. This data set has a certain reflective value. The ability to determine the ranging status of each AP.

[0125] Next, set Each AP in the data is for a single historical data sample. Position weight. Based on... In the interval The relative positions within the text are used to adjust the corresponding weights. The closer to the predicted position of a certain AP in one step If so, then a higher weight will be assigned to that AP for that sample.

[0126]

[0127] For set All APs within the sample The standardized weight vector. , .

[0128] Considering that the ranging error caused by multiple factors approximately follows a Gaussian distribution, this invention transforms the problem into fitting a Gaussian distribution where the variables are uncorrelated. ,in express The first of all ranging samples A set of items. Under this assumption, the offset function , wave function The estimated values ​​of both are obtained in the following way.

[0129] Next, the objective function is constructed using the form of Weighted Likelihood Estimation (WLE): ;

[0130] in .

[0131] By respectively right , Taking the partial derivative gives the information about , The optimal estimate; where, The optimal estimate The following formula 8 is used for calculation:

[0132] (Formula 8)

[0133] In formula 8, The size of the filtered range measurement sample set, for The first selected Each sample observation value For the corresponding number The weight coefficients of each sample, The coordinates of the point to be fitted;

[0134] The optimal estimate The following formula 9 is used for calculation:

[0135] (Formula 9)

[0136] This invention significantly improves positioning accuracy in narrow terrain environments, as verified by simulation experiments. Figure 3 As shown, compared with the simple average of the distance measurement results on both sides as the positioning result and the traditional robust adaptive Kalman filter algorithm, the accuracy of the SC-RAKF of the present invention is improved by 43.2% and 21.4%, respectively.

[0137] Based on the above-described cooperative positioning method, this embodiment of the invention also provides a cooperative positioning device, the internal structure of which is as follows: Figure 4 As shown, it includes: a preliminary positioning module 401 and a positioning adjustment module 402.

[0138] The preliminary positioning module 401 is used to perform a preliminary prediction of the position of a mobile terminal to be located in a narrow space using an access point (AP) activated within the effective ranging range of the mobile terminal, based on a robust adaptive Kalman filter algorithm. Specifically, the preliminary positioning module 401 is used for... The constructed Kalman filter is shown in Equation 1 above. The preliminary prediction results are shown in Formula 2 above.

[0139] The positioning adjustment module 402 is used to adjust the preliminary prediction results based on the offset function and fluctuation function of each AP, and use the adjustment result as the final positioning result of the mobile terminal; wherein, the offset function and fluctuation function of the AP are pre-fitted based on historical data.

[0140] Specifically, the positioning adjustment module 402 adjusts the position based on the offset function and fluctuation function of each AP according to the formula 7 above. of k Preliminary prediction results of time Adjustments were made to obtain Optimal prediction estimation results And thus obtain Optimal estimated position Thus, the adjusted final positioning result is obtained: .

[0141] Furthermore, the cooperative positioning device provided in this embodiment of the invention may also include a function fitting module 403.

[0142] The function fitting module 403 is used to fit the offset function and fluctuation function of the AP based on historical data;

[0143] Specifically, the function fitting module 403 can obtain the result based on the above formula 8. offset function According to the fitting of formula 9 above, the following is obtained: wave function .

[0144] In the technical solution of this invention, for a mobile terminal to be located in a narrow space, the position of the mobile terminal is initially predicted using all enabled access points (APs) within the effective ranging range of the mobile terminal, based on a robust adaptive Kalman filter algorithm. The initial prediction result is then adjusted based on the offset function and fluctuation function of each AP, and the adjusted result is used as the final positioning result of the mobile terminal. The offset function and fluctuation function of the AP represent the expected ranging error and fluctuation degree of the AP estimated from historical data, respectively, and are pre-fitted. By using the pre-fitted AP offset and fluctuation functions based on the robust adaptive Kalman filter algorithm, the positioning accuracy, reliability, and stability are improved by optimizing the ability to resist abnormal ranging values ​​under conditions of sparse AP coverage. This is especially beneficial for scenarios with unpredictable NLOS (Normally Unstable Out-of-Service) areas caused by environmental obstacles, thus enhancing the accuracy and reliability of positioning.

[0145] Furthermore, the technical solution of this invention performs robust estimation after making a preliminary prediction of the location of the mobile terminal, and predicts the covariance matrix adaptively. The robust estimation process and adaptive process of RAKF are optimized in the case of low-dimensional observation data, so as to further improve the positioning accuracy, precision and reliability.

[0146] Furthermore, by using measurement methods based on non-survey data for ranging bias (offset function) and fluctuation (fluctuation function), more robust outlier detection and processing during movement can be achieved, thereby optimizing the navigation and positioning performance of robust adaptive Kalman filtering and making the technology more adaptable to practical narrow terrain applications.

[0147] Figure 5 This illustration schematically shows a hardware architecture diagram of a computer device 1300 suitable for implementing a cooperative positioning method applicable to narrow spaces according to an embodiment of this application. In this embodiment, the computer device 1300 is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions. For example, it may be a smartphone, tablet computer, laptop computer, desktop computer, rack server, blade server, tower server, or cabinet server (including standalone servers or server clusters composed of multiple servers), etc. Figure 5 As shown, the computer device 1300 includes, but is not limited to, at least: a memory 1310, a processor 1320, and a network interface 1330 that can communicate with each other via a system bus. Wherein:

[0148] The memory 1310 includes at least one type of computer-readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 1310 may be an internal storage module of the computer device 1300, such as the hard disk or memory of the computer device 1300. In other embodiments, the memory 1310 may also be an external storage device of the computer device 1300, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. Of course, the memory 1310 may also include both the internal storage module and the external storage device of the computer device 1300. In this embodiment, the memory 1310 is typically used to store the operating system and various application software installed on the computer device 1300, such as the program code for the channel state information feedback method. In addition, the memory 1310 can also be used to temporarily store various types of data that have been output or will be output.

[0149] In some embodiments, processor 1320 may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip. Processor 1320 is typically used to control the overall operation of computer device 1300, such as performing control and processing related to data interaction or communication with computer device 1300. In this embodiment, processor 1320 is used to run program code stored in memory 1310 or process data.

[0150] Network interface 1330 may include a wireless network interface or a wired network interface, which is typically used to establish a communication link between computer device 1300 and other computer devices. For example, network interface 1330 is used to connect computer device 1300 to an external terminal via a network, establishing a data transmission channel and communication link between computer device 1300 and the external terminal. The network may be an intranet, the Internet, Global System for Mobile Communication (GSM), Wideband Code Division Multiple Access (WCDMA), 4G network, 5G network, Bluetooth, Wi-Fi, or other wireless or wired networks.

[0151] It should be pointed out that, Figure 5 Only a computer device with components 1310-1330 is shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0152] In this embodiment, the cooperative positioning method for narrow spaces stored in memory 1310 can be further divided into one or more program modules and executed by one or more processors (processor 1320 in this embodiment) to complete the embodiment of this application.

[0153] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the cooperative positioning method for narrow spaces described in the embodiments.

[0154] In this embodiment, the computer-readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the computer-readable storage medium can be an internal storage unit of a computer device, such as the hard disk or memory of the computer device. In other embodiments, the computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device. Of course, the computer-readable storage medium can also include both the internal storage unit and the external storage device of the computer device. In this embodiment, the computer-readable storage medium is typically used to store the operating system and various application software installed on the computer device, such as the program code of the channel state information feedback method in the embodiment. In addition, the computer-readable storage medium can also be used to temporarily store various types of data that have been output or will be output.

[0155] Obviously, those skilled in the art should understand that the modules or steps of the embodiments of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the embodiments of this application are not limited to any particular combination of hardware and software.

[0156] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method of cooperative positioning, characterized by, The method comprises the steps of: For a mobile terminal to be positioned in a narrow space, using APs enabled within the effective ranging range of the mobile terminal, a preliminary prediction of the position of the mobile terminal is made based on a robust adaptive Kalman filtering algorithm; The result of the preliminary prediction is adjusted based on the offset function and the fluctuation function of each AP, and the adjusted result is taken as the final positioning result of the mobile terminal; The offset function and the fluctuation function of the AP are pre-fitted according to historical data: Definition of history data format: ; wherein is a set of ranging APs, N is is a number of APs in is is a joint ranging data vector of the APs in is a preliminary predicted position combination corresponding to ; set containing all valid ranging ranges containing positions of aps; and further refining the set according to the preliminary prediction vector to obtain a refined set : , which is a set of historical data for positions defined on a continuous interval; wherein , are the minimum and maximum values of the elements in the vector​ The location weight of each AP in the set is adjusted according to the relative location of the one-step prediction location of the AP in the interval : The closer the one-step prediction location of an AP is to the location of the sample , the higher weight is assigned to the AP for that sample: ; where denotes the normalized weight.​ The objective function is constructed using the weighted maximum likelihood estimation method: ,in, ;in, express The first of all ranging samples A set of items For set Size, for The first selected Each sample observation value For the corresponding number The weight coefficients of each sample, Indicates the first i One AP; express The mean to be estimated; express The standard deviation to be estimated; Further, by respectively taking the partial derivatives of the wavelet function with respect to the parameters , , the migration function , an estimate of the wavelet function is obtained.

2. The method of claim 1, wherein, The use of the AP enabled within the effective ranging range of the mobile terminal, based on the robust adaptive Kalman filtering algorithm, to preliminarily predict the position of the mobile terminal, specifically includes: For The constructed Kalman filter is shown in Equation 1 : (Formula 1) in, , This represents the set of all enabled APs within the effective ranging range of the mobile terminal; for The predicted mobile terminal k The motion state vector at time t, including k Position, velocity, and acceleration at any given moment; for The state transition matrix from the previous state to the next state; for of k The system error vector at time t; for of k The observation location at that moment; for of k The coefficient matrix at time step; for of k The observation noise vector at time step; wherein, ; is the k ranging result at the moment, which is a scalar; is the k ranging direction vector at the moment, which remains unchanged within a single sub-interval; is the position coordinates; The preliminary prediction result is calculated according to the following Equation 2: (Formula 2) wherein is the k initial prediction of the state at time is the k covariance matrix of the state at time is the k optimal estimate of the state at time is the k covariance matrix of the state estimate at time is the k observation noise covariance at time 3. The method of claim 2, wherein, The adjustment of the preliminary prediction result based on the offset function and the fluctuation function of each AP specifically includes: The preliminary prediction result of the time point is adjusted according to the following formula 7 The optimal prediction estimation result of the time point is obtained k The preliminary prediction result of the time point is adjusted according to the following formula 7 The optimal prediction estimation result of the time point is obtained The preliminary prediction result of the time point is adjusted according to the following formula 7 The optimal prediction estimation result of the time point is obtained (Formula 7) wherein , ; is the robustness factor at time k ; is the adaptive factor at time k , obtained from the drift function and the volatility function ; is the observation noise covariance matrix at time k ; According to the optimal estimated position , the final positioning result is obtained: ; wherein ; wherein denotes the wave function of denotes a normalization operation.

4. The method of claim 3, wherein, After the preliminary prediction of the position of the mobile terminal, it further includes: The innovation vector is calculated according to equation 3 below and the innovation covariance matrix is calculated according to equation 4 below (Formula 3) wherein represents the k observation noise covariance matrix at time instant 5. The method of claim 4, wherein, The Specifically, it is calculated according to the following Equation 4: (Formula 4) wherein , , is a preset constant.

6. The method of claim 4, wherein, The Specifically, the calculation is performed according to the following method: According to The observation position at the moment is adjusted: k The observation position at the moment is adjusted: ; wherein, represents the offset function of each AP in the set at the function value combination of the location where the mobile terminal is located; represents the offset function of each AP in the set k at the ranging direction vector combination at the moment; represents the offset function of each AP in the set k at the observation position combination at the moment; represents the adjusted the offset function of each AP in the set k at the observation position combination at the moment; According to The normalized weight vector is obtained: ； wherein represents the function value combination of the fluctuation functions of the respective APs in at the function value of ; Unify the ranging values ​​of all APs to obtain the updated values ​​for each AP. k Combination of observation positions at different times: ; Further, an updated observation noise covariance matrix is obtained: ; wherein is the weight vector of the th item value; The updated innovation vector is obtained according to equation 5 below , the innovation covariance matrix , and the innovation statistics : (Formula 5) Adaptive factor matrix is determined according to the following equation 6: (Formula 6) wherein .

7. The method of claim 1, wherein, The The optimal estimate value of is calculated according to the following Equation 8: (Formula 8) wherein is the optimal estimate value.

8. The method of claim 7, wherein, The The optimal estimate value of the is calculated according to the following Equation 9: (Formula 9) wherein is the optimal estimate of 9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the steps of the cooperative positioning method in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, which can be executed by at least one processor to make the at least one processor execute the steps of the cooperative positioning method in any one of claims 1-8.

11. A co-locating apparatus, comprising: The method comprises the steps of: A preliminary positioning module is configured to, for a mobile terminal to be positioned in a narrow space, use APs enabled within the effective ranging range of the mobile terminal, and preliminarily predict the position of the mobile terminal based on a robust adaptive Kalman filtering algorithm; A positioning adjustment module is configured to adjust the result of the preliminary prediction based on the offset function and the fluctuation function of each AP, and take the adjusted result as the final positioning result of the mobile terminal; wherein the offset function and the fluctuation function of the AP are pre-fitted according to historical data: Definition of history data format: ; wherein is a set of ranging APs, N is a number of APs in is a joint ranging data vector of APs in is a preliminary predicted position combination corresponding to ​ set containing all valid ranging ranges containing positions of aps; further to the preliminary prediction vector further filter the set : as a set of historical data for positions defined on a continuous interval; wherein , are the minimum and maximum values of the elements in the vector , respectively​ Each AP in the set of APs is assigned a weight for a single historical data sample according to its relative position in the interval : The closer the one-step-ahead predicted position of a mobile device is to an AP , the higher weight is assigned to that AP for that sample: ; where is normalized.​ The objective function is constructed using the weighted maximum likelihood estimation method: ,in, ;in, express The first of all ranging samples A set of items For set Size, for The first selected Each sample observation value For the corresponding number The weight coefficients of each sample, Indicates the first i One AP; express The mean to be estimated; express The standard deviation to be estimated; Further, by respectively taking the partial derivatives of the , , , , estimates of the wavelet function

Images