An error estimation method and apparatus for pedestrian navigation systems

Abstract

This disclosure relates to an error estimation method and apparatus for pedestrian navigation systems. The method includes: acquiring detection data from the inertial measurement unit (IMU) of the pedestrian navigation system at the current moment and the navigation information source at the current moment; generating an original trajectory based on the detection data; dynamically selecting a corresponding target error estimation strategy from at least one preset error estimation strategy based on the navigation information source and the original trajectory; determining a target error based on the original trajectory and the target error estimation strategy; and using the target error as the error correction basis for the pedestrian navigation system to generate the original trajectory at the next moment. This allows for comprehensive judgment based on different reference information, determining the corresponding target error estimation strategy from multiple error estimation strategies for error estimation, accurately determining various errors in the pedestrian navigation system, significantly improving the positioning accuracy of the pedestrian navigation system in different scenarios, and balancing robustness and applicability.

Description

Technical Field

[0001] This disclosure relates to the field of pedestrian navigation, and in particular to an error estimation method and apparatus for pedestrian navigation systems. Background Technology

[0002] In related technologies, pedestrian navigation typically relies on Strapdown Inertial Navigation System (SINS) updates or Pedestrian Dead Reckoning (PDR) to estimate position, depending on the detection data from an Inertial Measurement Unit (IMU). However, in practice, measurement errors from the IMU sensors accumulate over time, leading to increasing drift and bias in position estimation. This is particularly problematic in complex environments where simple, effective, and accurate gait detection and actual stride estimation are often difficult, impacting the stability and accuracy of the pedestrian navigation system. Furthermore, many related technologies rely on single sensors or algorithms for error estimation, making accurate and reliable error estimation challenging in various situations. Therefore, accurately determining different types of errors in pedestrian navigation systems based on varying reference information, improving positioning accuracy in different scenarios, and ensuring both robustness and applicability are pressing issues that need to be addressed. Summary of the Invention

[0003] In view of this, this disclosure proposes an error estimation method and device for pedestrian navigation systems. It can accurately determine various errors of pedestrian navigation systems based on different reference information when facing different situations, significantly improve the positioning accuracy of pedestrian navigation systems in different scenarios, and take into account both robustness and applicability.

[0004] According to one aspect of this disclosure, an error estimation method for a pedestrian navigation system is provided. The method includes: acquiring detection data from an inertial measurement unit in the pedestrian navigation system at the current moment and a navigation information source at the current moment, the navigation information source including a digital map and / or satellite navigation information; generating an original trajectory based on the detection data; dynamically selecting a corresponding target error estimation strategy from at least one preset error estimation strategy based on the navigation information source and the original trajectory at the current moment; and determining a target error based on the original trajectory and the target error estimation strategy, the target error being used as an error correction basis for the pedestrian navigation system to generate the original trajectory at the next moment.

[0005] According to another aspect of this disclosure, an error estimation device for a pedestrian navigation system is provided. The device includes: a data acquisition module for acquiring detection data from the inertial measurement unit of the pedestrian navigation system at the current moment and a navigation information source at the current moment, the navigation information source including a digital map and / or satellite navigation information; a trajectory generation module for generating an original trajectory based on the detection data; a strategy selection module for dynamically selecting a corresponding target error estimation strategy from at least one preset error estimation strategy based on the navigation information source and the original trajectory at the current moment; and an error determination module for determining a target error based on the original trajectory and the target error estimation strategy, the target error being used as the error correction basis for the pedestrian navigation system to generate the original trajectory at the next moment.

[0006] According to another aspect of this disclosure, an error estimation apparatus for a pedestrian navigation system is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above method.

[0007] According to another aspect of this disclosure, a non-volatile computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the above-described method.

[0008] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0009] By acquiring the detection data from the inertial measurement unit (IMU) of the pedestrian navigation system at the current moment and the navigation information source at the current moment, an original trajectory is generated based on the detection data. Then, based on the navigation information source at the current moment and the original trajectory, a corresponding target error estimation strategy is dynamically selected from at least one preset error estimation strategy. Finally, based on the original trajectory and the target error estimation strategy, the target error is determined. The target error is used as the error correction basis for the pedestrian navigation system to generate the original trajectory at the next moment. This disclosure is not limited to relying on a single sensor or a single algorithm for error estimation. Instead, when facing various situations, it comprehensively judges based on different forms of external information and scene information, and determines the corresponding target error estimation strategy from multiple error estimation strategies for error estimation. Thus, when facing various situations, it can accurately determine various errors of the pedestrian navigation system based on different reference information, significantly improving the positioning accuracy of the pedestrian navigation system in different scenarios, while taking into account robustness and applicability.

[0010] Other features and aspects of this disclosure will become clear from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0011] The accompanying drawings, which are included in and form part of this specification, illustrate exemplary embodiments, features, and aspects of this disclosure together with the specification and serve to explain the principles of this disclosure.

[0012] Figure 1 A flowchart is shown illustrating an error estimation method for a pedestrian navigation system provided according to an embodiment of the present disclosure;

[0013] Figure 2 This diagram illustrates a comparison between a first reference trajectory and an original trajectory provided according to an embodiment of the present disclosure.

[0014] Figure 3 This diagram illustrates a first reference trajectory similar to the original trajectory provided according to an embodiment of the present disclosure;

[0015] Figure 4 A schematic diagram of a sliding window provided according to an embodiment of the present disclosure is shown;

[0016] Figure 5 A schematic diagram illustrating a typical scenario provided by an embodiment of this disclosure, using a standard 400-meter track;

[0017] Figure 6 A schematic diagram illustrating a pedestrian path feature according to an embodiment of the present disclosure is shown.

[0018] Figure 7 A schematic diagram showing the principal direction vector of a moving target according to an embodiment of the present disclosure is provided;

[0019] Figure 8 A block diagram is shown of an apparatus 1900 for error estimation of a pedestrian navigation system according to an embodiment of the present disclosure. Detailed Implementation

[0020] Various exemplary embodiments, features, and aspects of this disclosure will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0021] As used herein, the terms “comprising,” “including,” “having,” or variations thereof are open-ended and include one or more stated features, integrals, elements, steps, components, or functions, but do not exclude the presence or addition of one or more other features, integrals, elements, steps, components, functions, or groups thereof. When an element is referred to as “connected,” “coupled,” “responsive,” or variations thereof relative to another element, it may be directly connected, coupled, or responsive to the other element, or there may be intermediate elements present. Although the terms first, second, third, etc., may be used herein to describe various elements / operations, these elements / operations should not be limited by these terms. These terms are used only to distinguish one element / operation from another element / operation. Thus, a first element / operation in some embodiments may be referred to as a second element / operation in other embodiments without departing from the teachings of the inventive concept. The term “exemplary” used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0022] Furthermore, to better illustrate this disclosure, numerous specific details are set forth in the following detailed description. Those skilled in the art will understand that this disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art have not been described in detail in order to highlight the main points of this disclosure.

[0023] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant regions.

[0024] To address the aforementioned technical problems, this disclosure provides an error estimation method and apparatus for pedestrian navigation systems. The method acquires detection data from the inertial measurement unit (IMU) of the pedestrian navigation system at the current moment and the navigation information source at the current moment. An original trajectory is generated based on the detection data. Then, based on the navigation information source and the original trajectory at the current moment, a corresponding target error estimation strategy is dynamically selected from at least one preset error estimation strategy. Finally, based on the original trajectory and the target error estimation strategy, a target error is determined. This target error is used as the error correction basis for the pedestrian navigation system to generate the original trajectory for the next moment. This method, by comprehensively judging based on actual conditions and determining the corresponding target error estimation strategy from multiple error estimation strategies for error estimation, can accurately determine various errors in the pedestrian navigation system, significantly improving the positioning accuracy of the pedestrian navigation system, while also considering universality, robustness, and reliability.

[0025] Figure 1 A flowchart illustrating an error estimation method for a pedestrian navigation system according to an embodiment of the present disclosure is shown. Figure 1 As shown, the error estimation method for pedestrian navigation systems provided in this embodiment includes steps S10-S40. This method can be applied to terminal devices, such as smartphones, smart glasses, smartwatches, navigators, tablets, laptops, and other devices that can meet the resource requirements of pedestrian navigation systems. This disclosure does not limit this application.

[0026] Step S10: Obtain the detection data of the inertial measurement unit in the pedestrian navigation system at the current moment and the navigation information source at the current moment, wherein the navigation information source includes digital maps and / or satellite navigation information.

[0027] The pedestrian navigation system includes at least a pedestrian odometry system and an inertial measurement unit (IMU). The pedestrian odometry system calculates displacement vectors based on data from the IMU. The IMU, which includes at least a gyroscope and an accelerometer, is used to detect the actual motion data of moving targets in real time. The detection data may include angular velocity and acceleration. The digital map can be an open-source road network map. The satellite navigation information can be navigation information determined based on systems such as the Global Positioning System (GPS) and the BeiDou Navigation Satellite System (BDS).

[0028] Step S20: Generate the original trajectory based on the detection data.

[0029] The original trajectory can be generated based on the detection data using methods such as strapdown inertial navigation system updates or pedestrian dead reckoning. The original trajectory represents the trajectory generated by the pedestrian navigation system based on the target error at the previous moment and the detection data at the current moment.

[0030] The coordinate system of the original trajectory can be the navigation coordinate system (n-system), which represents the East-North-Sky geographic coordinate system. The coordinate system of the inertial measurement unit can be the carrier coordinate system (b-system), which represents the right-front-up coordinate system, with its origin at the center of gravity of the carrier (i.e., the terminal equipment with the inertial measurement unit). b The axis is to the right along the transverse axis of the carrier, y b The axis moves forward along the longitudinal axis of the carrier, z b If the axis is upward along the vertical axis of the carrier, then the angular velocity and acceleration in the detected data are the real-time angular velocity and acceleration of the moving target carrying the carrier in the b-frame. The coordinate system of the pedestrian odometer can be the pedestrian coordinate system (m-frame), which represents a right-front-up coordinate system, with its origin at the center of gravity of the moving target, x... m The axis moves to the right along the horizontal axis of the moving target, y m The axis moves forward along the longitudinal axis of the moving target, z mIf the axis is aligned upwards along the vertical axis of the moving target, the pedestrian odometer can calculate the cumulative displacement vector of the moving target along its path in the m-frame based on the detection data. The aforementioned coordinate system can be set according to actual needs, and this disclosure does not impose any restrictions on it.

[0031] Step S30: Based on the navigation information source and the original trajectory at the current moment, dynamically select the corresponding target error estimation strategy from at least one preset error estimation strategy.

[0032] Specifically, based on the current navigation information source and the original trajectory, a target error estimation strategy that is more suitable for the current navigation scenario can be dynamically selected from at least one preset error estimation strategy, so as to achieve more accurate and efficient error estimation under different navigation scenarios.

[0033] In order to accurately select the target error estimation strategy, it is also necessary to determine the pedestrian path characteristics based on the original trajectory before making the strategy selection.

[0034] In some embodiments, after determining the original trajectory, multiple positioning points are extracted from the original trajectory at fixed intervals to form a discrete trajectory sequence {p}. i}, p i Let p represent the i-th positioning point. i adjacent positioning points p i+1 p i-1 Each positioning point p is calculated based on the discrete difference approximation. i The local curvature can be calculated using the following formula:

[0035]

[0036]

[0037]

[0038] In the formula, Indicates the location point p i The vertical coordinate value in the navigation coordinate system (n-system); Represents the (i+1)th location point p i+1 The vertical coordinate value in the navigation coordinate system (n-system); p represents the (i-1)th location point i-1 The vertical coordinate value in the navigation coordinate system (n-system); This indicates the interval between adjacent sampling points on the horizontal coordinate in the navigation coordinate system (n-system); This indicates that the original trajectory is at the location point p. i The first-order rate of change at point p is used to represent the location of point p. i The slope of the tangent at that point; This indicates that the original trajectory is at the location point p. i The second rate of change at point p is used to represent the location of point p. i The degree of curvature of the original trajectory; This indicates that the original trajectory is at the location point p. i The curvature at that point.

[0039] Among them, at the positioning point p i curvature satisfy In this case, Given a preset curvature threshold, the moving target is considered to be at the positioning point p. i When in a turning state, position point p i Mark it as a turning point; otherwise, mark it as a straight point. This will result in the discrete trajectory sequence {p}. i The positioning points in the diagram are divided into turning points and straight-ahead points. Based on the turning points, the original trajectory is divided into multiple segments of a specific trajectory, consisting of consecutive turning points or consecutive straight-ahead points. For example, using... Figure 6 Taking the original path shown as an example, p1, p2, and p3 are marked turning points. The original trajectory is divided into four specific trajectories: trajectories 1, 2, 3, and 4. The length characteristics of each specific trajectory segment are calculated. Turning characteristics Characteristics of height variation By doing so, pedestrian path features can be obtained. .in, , …represent the characteristics of specific trajectories 1, 2… determined based on each positioning point p1, p2… That is, ( This represents the characteristics of a specific trajectory i.

[0040] In one possible implementation, step S30 may include: first determining whether the pedestrian path features determined based on the original trajectory match preset typical scene path features. If the pedestrian path features determined based on the original trajectory match the preset typical scene path features, the typical scene error estimation strategy is determined as the target error estimation strategy. Then, if the pedestrian path features do not match the preset typical scene path features, the current navigation information source is determined. If the pedestrian path features do not match the preset typical scene path features and the navigation information source is the digital map, the map error estimation strategy is determined as the target error estimation strategy. If the pedestrian path features do not match the preset typical scene path features and the navigation information source is satellite navigation information, the satellite error estimation strategy is determined as the target error estimation strategy. Furthermore, if no navigation information source is detected and the pedestrian path features do not match the preset typical scene path features, the constraint error estimation strategy is determined as the target error estimation strategy.

[0041] The error estimation strategy for typical scenarios is based on the corresponding typical scenarios, and different typical scenarios require different error estimation strategies. A typical scenario can be a site, road, or building space with a typical geometric layout or fixed structural dimensions, such as a playground or stadium. The path features of a typical scenario can be a set of features that accurately identify the structure of the typical scenario, including path length features. Turning characteristics Characteristics of height variation wait.

[0042] For example, taking a typical scenario like a 400-meter standard track as an example, such as Figure 5 As shown, the main structure of a 400-meter standard track includes two parallel straight sections, each 87 meters long, and two semi-circular curves. The curve radii range from 35.00 meters to 38.00 meters, and the height variation is approximately 0 meters. By storing the features of each segment of the 400-meter standard track, a sequence of typical scenario path features for a "400-meter standard track" is obtained. ,in, , … represents the length characteristic. , …is a turning feature. , …is characterized by high variability. , …represent the characteristics corresponding to paths 1, 2… of a “400-meter standard track”. Then… This represents the feature corresponding to path i in the "400-meter standard track".

[0043] In some embodiments, in step S30, methods such as weighted Euclidean distance can be used to determine whether there is a match between pedestrian path features and typical scene path features.

[0044] Taking the use of weighted Euclidean distance in step S30 as an example, pedestrian path features Typical scenario path characteristics Weighted Euclidean distance between It can be calculated using the following formula:

[0045]

[0046] In the formula, , , Representing length characteristics respectively Turning characteristics Characteristics of height variation The weighting coefficients. By assigning different weights to different features, the accuracy of path feature matching can be enhanced.

[0047] Matching thresholds can be set in advance as needed. Therefore, in determining the weighted Euclidean distance satisfy In the case of a certain condition, the pedestrian path features are considered to match the preset typical scene path features.

[0048] The map error estimation strategy involves using a first reference trajectory obtained from a digital map and applying the similarity principle to estimate the target error on the original trajectory. The satellite error estimation strategy involves using a second reference trajectory obtained from satellite navigation information and employing a Kalman filter algorithm to estimate the target error on the original trajectory. The constraint error estimation strategy involves performing principal component analysis on the position sequence of the moving target, determining that the moving target is in a straight-line state, and then using a Kalman filter algorithm to estimate the target error on the original trajectory. Descriptions of each strategy are provided below and will not be repeated here for brevity.

[0049] This method enables pedestrian navigation systems to no longer rely on a single sensor or algorithm for error estimation. Instead, it can dynamically select the corresponding target error estimation strategy from at least one preset error estimation strategy based on multi-source data such as the original trajectory and digital map generated by the inertial measurement unit, satellite navigation information, and typical scene path characteristics. Specifically, it can automatically select the target error estimation strategy from four strategies: map error estimation, satellite error estimation, typical scene error estimation, and constraint error estimation. This significantly improves the system's universality while enhancing its robustness and stability.

[0050] Step S40: Based on the original trajectory and the target error estimation strategy, the target error is determined. The target error is used as the error correction basis for the pedestrian navigation system to generate the original trajectory at the next moment.

[0051] Among them, the determination method and content of target error differ based on different target error estimation strategies. The following will explain the determination method and content of target error by taking map error estimation strategy, satellite error estimation strategy, typical scenario error estimation strategy and constraint error estimation strategy as examples (i.e., implementation methods one to four).

[0052] Implementation Method 1: The target error estimation strategy is the map error estimation strategy.

[0053] Step S40 may include: when the target error estimation strategy is the map error estimation strategy, determining a first reference trajectory for the original trajectory based on the digital map and the original trajectory; determining a first error of the pedestrian navigation system based on the geometric relationship between the first reference trajectory and the original trajectory, and defining the first error as the target error. The first error includes at least one of the following: the track angle error of the pedestrian navigation system, the calibration coefficient error of the pedestrian odometry, and the zero-bias error of the gyroscope in the inertial measurement unit. Thus, the system's correction of the track angle error based on the first error can effectively suppress the interference of attitude drift on the judgment of the direction of motion, thereby avoiding overall trajectory rotation or distortion; the system's correction of the pedestrian odometry calibration error based on the first error can improve the accuracy of step length calculation and significantly reduce the cumulative distance deviation caused by individual gait differences, walking speed changes, or terrain undulations of the moving target; the system's correction of the gyroscope zero-bias error based on the first error can significantly improve the robustness of the system and reduce the divergence rate of the attitude calculation results. Therefore, by determining the second error as the target error and using it as the basis for error correction in generating the original trajectory at the next moment, the system can significantly improve the positioning accuracy of the pedestrian navigation system.

[0054] In one possible implementation, determining a first reference trajectory for the original trajectory based on the digital map and the original trajectory may include: determining the observation probability and state transition probability of each observation point in the original trajectory relative to a road in the digital map based on the digital map and the original trajectory; and determining the first reference trajectory based on the observation probability and state transition probability corresponding to each observation point.

[0055] For example, a digital map can be an open-source road network map. A road network map is composed of basic data elements, which are the smallest topological and management units constituting the road network map data model. These basic data elements can include point elements, path elements, and relationship elements. Therefore, the basic data elements of a road network map can be divided into spatial data and attribute data. Spatial data can be used to represent the geometric shape and spatial relationship of each basic data element in the road network. Examples include: latitude and longitude coordinates stored as point elements; coordinate sequences composed of multiple point elements stored as road elements; and connection relationships between multiple point elements and / or road elements stored as relationship elements. Attribute data is used to represent the attribute characteristics and constraints of each basic data element in the road network. Examples include: unique identifiers, level types, and various traffic restriction rules for each element. After processing the spatial and attribute data of the road network map, road network map information can be obtained. This information can include: the unique identifier (osm_id) of each element in the road network map, the road name, the road type, and the road geometry.

[0056] In this implementation, when the digital map is a road network map, the observation probability and state transition probability of each observation point in the original trajectory relative to the road in the road network map can be determined based on the road network map information and the original trajectory, using Hidden Markov Models and other methods.

[0057] Based on the Hidden Markov Model, the observation probability can be calculated using the following formula:

[0058]

[0059] In the formula, This represents the standard deviation of the position observation error in a pedestrian navigation system. Represents the position observation point in the original trajectory To the road corresponding to the road network map The nearest great circle distance.

[0060] Based on the Hidden Markov Model, the state transition probability can be calculated using the following formula:

[0061]

[0062] In the formula, This is expressed as follows: If the road where the moving target is located at time t is... Then, at the next moment t+1, the player will move to the road. The probability of.

[0063] Then the Viterbi algorithm can be used to determine the observation probability relative to roads in the road network map. and state transition probability A first reference trajectory is determined. For example, the comparison between the first reference trajectory and the original trajectory is as follows: Figure 2 As shown.

[0064] In one possible implementation, after determining the first reference trajectory, determining the first error of the pedestrian navigation system based on the geometric relationship between the first reference trajectory and the original trajectory may include: determining the cross product of the total displacement of the original trajectory and the total displacement of the first reference trajectory, the first modulus of the total displacement of the original trajectory, and the second modulus of the total displacement of the first reference trajectory based on the geometric relationship between the original trajectory and the first reference trajectory; determining the track angle error of the pedestrian navigation system based on the cross product, the first modulus, and the second modulus; determining the scale coefficient error of the pedestrian odometer based on the ratio of the first modulus to the second modulus; and determining the zero bias error of the gyroscope in the inertial measurement unit based on the track angle error of the pedestrian navigation system when the original trajectory meets preset conditions.

[0065] In this implementation, the cross product of the total displacement of the original trajectory and the total displacement of the first reference trajectory is determined based on the geometric relationship between the original trajectory and the first reference trajectory. The first modulus of the total displacement of the original trajectory and the second modulus of the total displacement of the first reference trajectory This can be achieved through the following process:

[0066] The first step is to determine the moving target corresponding to the first reference trajectory at T. D Reference displacement vector in the n-system over a time period With the second measured displacement vector The relationships between them include:

[0067] Based on the heading installation angle error The moving target corresponding to the original trajectory calculated by the pedestrian odometry at T D The first measured displacement vector in the m-frame within the time period This allows us to obtain the moving target in T. D The second measured displacement vector in the b-frame during the time period It can be calculated using the following formula:

[0068]

[0069] In the formula, This indicates the heading installation angle error of the pedestrian navigation system. Indicates the moving target at T D The first measured displacement vector in the m-frame within the time period. Indicates the moving target at TD The second measured displacement vector in frame b within the time period. This includes the heading installation angle error. It could be the heading installation angle error that occurs when the detection data of the inertial measurement unit is converted from the b-frame to the m-frame where the pedestrian odometer is located in the pedestrian navigation system.

[0070] Then it can be done by transforming the matrix. For the second measured displacement vector After conversion, the moving target in T is obtained. D The third measured displacement vector in the n-system within the time period The conversion process can be calculated using the following formula:

[0071]

[0072] Because the measured displacement vector obtained by the system during the actual movement of the moving target will have various errors, it will differ from the moving target corresponding to the first reference trajectory at T. D Errors exist between reference displacement vectors within a time period, therefore, when considering the attitude angle error of the pedestrian navigation system... Heading installation angle error And pedestrian odometer scale coefficient error In the same case, referring to the above conversion process of the measured displacement vector, the moving target corresponding to the first reference trajectory at T can be obtained similarly. D Reference displacement vector in the n-system over a time period The moving target corresponding to the first reference trajectory at T D Reference displacement vector in frame b within the time period Based on the transformation relationship and the aforementioned error, the reference displacement vector can be determined. With the second measured displacement vector The relationship between them can be calculated using the following formula:

[0073]

[0074] In the formula, This represents the transformation matrix from the b-frame to the n-frame corresponding to the reference displacement vector at the previous moment. This indicates that the moving target corresponding to the first reference trajectory is at T D The reference displacement vector in frame b within the time period. This represents a unit vector. This represents the attitude angle error at the previous moment. In This indicates an antisymmetric matrix operator.

[0075] The second step is to determine the geometric relationship between the original trajectory and the first reference trajectory, including:

[0076] Due to the attitude angle error of the pedestrian navigation system It could be the attitude angle error that occurs when the inertial measurement unit's detection data is converted from the b-frame to the n-frame containing the original trajectory in a pedestrian navigation system. It can be a three-dimensional vector, and can be calculated using the following formula:

[0077]

[0078] In the formula, Indicates pitch angle error. Indicates roll angle error. This indicates the heading angle error.

[0079] The pedestrian odometer scale coefficient error It can be the proportional error between the measured distance increment corresponding to the measured displacement vector and the reference distance increment corresponding to the reference displacement vector.

[0080] To simplify the calculation, attitude angle errors can be ignored. Heading installation angle error And pedestrian odometer scale coefficient error From the higher-order terms, we obtain the reference displacement vector equation:

[0081]

[0082] In the formula, the coefficients .

[0083] Then, based on the above reference displacement vector equation, the dead reckoning error equation can be obtained, which can be calculated using the following formula:

[0084]

[0085] In the formula, Indicates the moving target at T D Displacement error within a time period.

[0086] Based on the above dead reckoning error equation, the moving target at T D Displacement error within a time period Accumulate from time t1 to t j At that moment, the total displacement error of the moving target can be obtained. To simplify the calculation, if the zero-bias error of the gyroscope in the inertial measurement unit is less than a certain small amount over a certain time period, the attitude angle error in the dead reckoning error equation during the motion can be considered constant. Further assuming that the heading installation angle and the pedestrian odometer calibration coefficient error are also constant, the following formula can be used for calculation:

[0087]

[0088] In the formula, It is the total displacement of the moving target.

[0089] Due to the horizontal attitude error angle (i.e., pitch angle error) after the pedestrian navigation system performs self-alignment. and roll angle error The values ​​of the horizontal attitude error angle are relatively small, and the algorithm of the gyroscope in the inertial measurement unit can effectively improve the horizontal attitude. Therefore, the influence of the horizontal attitude error angle can be ignored. Simplifying the above formula again, we obtain the similarity formula between the original trajectory and the first reference trajectory:

[0090]

[0091] In the formula, This represents the error vector in the vertical direction. Thus, the geometric relationship between the original trajectory and the first reference trajectory can be characterized using this similarity formula.

[0092] The third step involves determining the cross product of the total displacement of the original trajectory and the total displacement of the first reference trajectory, the first modulus of the total displacement of the original trajectory, and the second modulus of the total displacement of the first reference trajectory, including:

[0093] like Figure 3 As shown, vector Represents the total displacement of the moving target That is, the total displacement of the original trajectory, in vector form. Reference displacement representing the moving target This refers to the total displacement of the first reference trajectory. The total displacement of the moving target. Reference displacement of the moving target The included angle between them is the track angle error of the pedestrian navigation system. Thus in vector for Vector for In this case, a vector can be obtained. The total displacement error of the moving target. ,Right now .

[0094] Based on the similarity formula (i.e., geometric relationship) between the original trajectory and the first reference trajectory, the cross product of the total displacement of the original trajectory and the total displacement of the first reference trajectory, the first modulus of the total displacement of the original trajectory, and the second modulus of the total displacement of the first reference trajectory can be determined using the following formula:

[0095]

[0096] In the formula, The first modulus represents the total displacement of the original trajectory. The second modulus represents the total displacement of the first reference trajectory. This represents the cross product of the total displacement of the original trajectory and the total displacement of the first reference trajectory.

[0097] In this implementation, the result of the cross product can be used as a basis. First module length Second module length Determine the track angle error of the pedestrian navigation system It can be calculated using the following formula:

[0098]

[0099] In this implementation, the first module length can be used as a reference. Second module length The ratio determines the scale factor error of the pedestrian odometer. It can be calculated using the following formula:

[0100]

[0101] In this implementation, if the original trajectory meets preset conditions, determining the zero-bias error of the gyroscope in the inertial measurement unit based on the track angle error of the pedestrian navigation system can include:

[0102] Assuming installation error at the heading angle If the error has been fully corrected, the above dead reckoning error equation can be simplified to:

[0103]

[0104] Furthermore, if the original trajectory meets the preset conditions, it can be Set as a constant vector. The preset condition is that the changes in velocity and direction of the moving target's displacement over a period of time are both less than a certain small amount. Further assuming that the zero-bias drift of the gyroscope in the inertial measurement unit is also constant, the formula for the change in heading angle error can be obtained:

[0105]

[0106] In the formula, This represents the heading angle error at time t0. Indicates t j The heading angle error at time [time] This indicates the zero bias error of the gyroscope in the inertial measurement unit.

[0107] The above formula for the change in heading angle error is accumulated from time t1 to t. j At any given moment, the total displacement error of the moving target can also be obtained. It can be calculated using the following formula:

[0108]

[0109] Then, assuming the heading angle error is corrected by road sign after zero speed correction or road network matching. If the error has been fully corrected, then the zero bias error of the gyroscope in the inertial measurement unit can be obtained. It can be calculated using the following formula:

[0110]

[0111] This allows the track angle error of the pedestrian navigation system to be reduced. Scale coefficient error of pedestrian odometer Zero bias error of the gyroscope in the inertial measurement unit The first error of the pedestrian navigation system is identified and designated as the target error.

[0112] In the above-described map error estimation strategy, the first reference trajectory obtained from the digital map is used to estimate the target error of the original trajectory based on the similarity principle. This can accurately estimate the track angle error of the pedestrian navigation system, the scale coefficient error of the pedestrian odometer, and the zero bias error of the gyroscope in the inertial measurement unit. This provides an accurate and reliable error correction basis for the pedestrian navigation system to generate the original trajectory for the next moment.

[0113] Implementation Method 2: The target error estimation strategy is the satellite error estimation strategy.

[0114] In one possible implementation, step S40 may include: when the target error estimation strategy is the satellite error estimation strategy, determining a second reference trajectory for the original trajectory based on the satellite navigation information; determining a first trajectory segment corresponding to the original trajectory and a second trajectory segment corresponding to the second reference trajectory based on the second reference trajectory and the original trajectory; determining the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment based on the first trajectory segment and the second trajectory segment; determining a second error of the pedestrian navigation system using a Kalman filter algorithm based on the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment, and determining the second error as the target error; wherein the second error includes at least one of the heading angle error of the pedestrian navigation system, the scaling factor error of the pedestrian odometer, and the zero bias error of the gyroscope in the inertial measurement unit. Thus, the system's correction of heading angle error based on the second error effectively eliminates the cumulative heading drift generated by the inertial measurement unit during long-term operation, thereby ensuring the long-term accuracy of the system's attitude recursion and the stability of the motion direction calculation. The system's correction of pedestrian odometry scaling factor error based on the second error significantly reduces the cumulative deviation in stride length calculation caused by individual gait differences, walking speed variations, or terrain undulations, thereby effectively improving the distance accuracy of pedestrian dead reckoning. The system's correction of gyroscope bias error based on the second error significantly improves the system's robustness and reduces the divergence rate of attitude calculation results. Therefore, by defining the first error as the target error and using it as the error correction basis for generating the original trajectory at the next moment, the system can significantly improve the positioning accuracy of the pedestrian navigation system.

[0115] The satellite navigation information may include various types of satellite navigation observation data, such as pseudorange observations, carrier phase observations, and Doppler frequency shift observations. Based on these various satellite navigation observation data, a second reference trajectory can be determined relative to the original trajectory after processing; this disclosure does not impose any limitations on this.

[0116] In this implementation, during the process of determining the first trajectory segment corresponding to the original trajectory and the second trajectory segment corresponding to the second reference trajectory, multiple sliding windows with fixed durations can be pre-set, and adjacent sliding windows can also have a fixed overlap rate. This allows for the segmentation of continuous trajectory data, facilitating subsequent analysis and calculation. The duration and overlap rate of all sliding windows can be set according to actual needs, and this disclosure does not impose any limitations on this. For example, such as... Figure 4 As shown, the duration of the sliding window is set to 10 seconds, and the overlap rate is set to 70%. Then, multiple sliding windows can be used to record the time period T. j The second reference trajectory L within sjand the original trajectory L pj According to the second reference trajectory L sj and the original trajectory L pj The original trajectory L can be determined by fitting using the least squares method. pj The corresponding first trajectory segment S1 and the second reference trajectory L sj The corresponding second trajectory segment S2.

[0117] In one possible implementation, determining the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment based on the first trajectory segment and the second trajectory segment may include: determining the length ratio of the first trajectory segment to the second trajectory segment and the included angle between the first trajectory segment and the second trajectory segment based on the first trajectory segment and the second trajectory segment; determining the scaling factor error of the pedestrian odometer based on the length ratio; and determining the heading angle error of the pedestrian navigation system based on the included angle between the first trajectory segment and the second trajectory segment.

[0118] In this implementation, the length ratio of the first trajectory segment S1 to the second trajectory segment S2 can be determined first, based on the first trajectory segment S1 and the second trajectory segment S2. and the angle between the first trajectory segment S1 and the second trajectory segment S2. Then, based on the ratio of the lengths of the first trajectory segment to the second trajectory segment... Determine the scaling factor error of the pedestrian odometer It can be calculated using the following formula:

[0119]

[0120] And based on the angle between the first trajectory segment and the second trajectory segment Determine the heading angle error of the pedestrian navigation system It can be calculated using the following formula:

[0121]

[0122] Further, in one possible implementation, determining the second error of the pedestrian navigation system based on the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment using a Kalman filter algorithm, and defining the second error as the target error, may include: determining a first system state transition matrix and a first system observation matrix for the current moment based on the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer; wherein, a pre-built correspondence is established between the first system state transition matrix and the first system state vector, and a pre-built correspondence is established between the first system observation matrix and the first system observation vector; the first system state vector includes at least one of the following: the heading angle error of the pedestrian navigation system, the position information corresponding to the second reference trajectory, the scaling factor error of the pedestrian odometer, and the zero bias error of the gyroscope in the inertial measurement unit; the first system observation vector includes the heading angle error of the pedestrian navigation system, the position information corresponding to the second reference trajectory, the position information corresponding to the second reference trajectory, and the position information corresponding to the second reference trajectory. The system comprises at least one of the following: location information and pedestrian odometry scaling factor error; determining a predicted first system state vector for the current time based on the first system state transition matrix at the current time and the first system state vector at the previous time; determining a predicted first error covariance matrix for the current time based on the first system state transition matrix at the current time and the first error covariance matrix at the previous time, wherein a pre-built correspondence is established between the first error covariance matrix and the first system state vector; determining a first Kalman gain matrix for the current time based on the predicted first error covariance matrix and the first system observation matrix; determining a first system state vector for the current time based on the first Kalman gain matrix, the first system observation vector, the first system observation matrix, and the predicted first system state vector; and determining a second error of the pedestrian navigation system based on the first system state vector at the current time, and defining the second error as the target error.

[0123] In this implementation, the heading angle error of the pedestrian navigation system based on the current time k The proportional factor error of the pedestrian odometer Determine the first system state transition matrix for the current time k. and the first system observation matrix The implementation method is as follows:

[0124] First, determine the first system state transition matrix at the current time k. The implementation method is as follows:

[0125] Wherein, the first system state vector at time k It can be represented as:

[0126]

[0127] In the formula, This represents the heading angle error of the pedestrian navigation system at time k. , Let represent the eastward and northward positions of the second reference trajectory at time k in the n-frame, respectively. This represents the scaling factor error of the pedestrian odometry at time k. , , These represent the zero-bias errors of the gyroscope in the inertial measurement unit at time k in the x-axis, y-axis, and z-axis directions in the b-frame, respectively.

[0128] Among them, the first system state vector Each component in the system can be updated in real time. Taking the update of the first system state vector at time k-th time to obtain the first system state vector at time k+1 as an example...

[0129] The first system state vector at time k can be represented as follows: :

[0130]

[0131] The first system state vector at time k+1 can be represented as: :

[0132]

[0133] Heading angle error of pedestrian navigation system Updates can be calculated using the following formula:

[0134]

[0135] In the formula, This represents the angular velocity measured by the inertial measurement unit at time k. Indicates the sampling period. This represents the process noise associated with the heading angle error.

[0136] The position corresponding to the second reference trajectory Updates can be calculated using the following formula:

[0137]

[0138]

[0139] In the formula, This represents the rangefinder output step size in the pedestrian navigation system at time k. and These represent the process noise at the eastward and northward positions in the second reference trajectory under the n-system, respectively.

[0140] Scale factor error of pedestrian odometer Updates can be calculated using the following formula:

[0141]

[0142] In the formula, This represents the process noise of the proportional factor error in pedestrian odometers.

[0143] Zero bias error of the gyroscope in the inertial measurement unit , , Updates can be calculated using the following formula:

[0144]

[0145]

[0146]

[0147] In the formula, , , These represent the process noise of the zero bias error of the gyroscope in the inertial measurement unit in the x-axis, y-axis, and z-axis directions in the b-frame, respectively.

[0148] Based on the update formulas for each of the above components, merging them yields the first system state vector. Update formula:

[0149] ,

[0150] In the formula, Let represent the control vector at time k, and ; The process noise at time k represents each component. The process noise covariance matrix is ​​used to reflect the random walk intensity of each component in the first system state vector.

[0151] Therefore, based on the first system state vector at time k... and the first system state vector at time k+1 The first system state transition matrix at time k can then be obtained. It can be calculated using the following formula:

[0152]

[0153] Based on the first system state vector at time k above The first system state transition matrix This allows us to obtain the system model of the pedestrian navigation system:

[0154]

[0155] In the formula, This represents the input matrix, used to store the control vector. In and Mapping to the first system state vector middle, This represents the first system state vector at time k-1.

[0156] The first system observation matrix at current time k The determination process is as follows:

[0157] Because the information that can be observed by the pedestrian navigation system includes the heading angle error of the pedestrian navigation system. The eastward position in the second reference trajectory within the n-system and northward position Scale factor error of pedestrian odometer Therefore, the observation vector of the first system can be... Preset to:

[0158]

[0159] In the formula, This represents the observation noise vector at time k. This represents the observation noise covariance matrix, used to reflect the confidence level of each component in the first system's observation vector.

[0160] Therefore, based on the aforementioned first system state vector First system observation vector The first system observation matrix at time k can be obtained directly. :

[0161]

[0162] Based on the aforementioned first system state vector First system observation vector First system observation matrix The observation model of the pedestrian navigation system can be obtained:

[0163]

[0164] Therefore, the heading angle error of the pedestrian navigation system based on the current time k The proportional factor error of the pedestrian odometer The first system state transition matrix for the current time k can be determined. and the first system observation matrix .

[0165] For example, the steps of determining the target error using the Kalman filter algorithm at the current time k may include:

[0166] Prediction using the Kalman filter algorithm:

[0167] We can first use the first system state transition matrix at the current time k. and the first system state vector at the previous time k-1 Determine the predicted first system state vector for the current time k. It can be calculated using the following formula:

[0168]

[0169] Then, based on the first system state transition matrix at the current time k The first error covariance matrix at the previous time k-1 Determine the first error covariance matrix for the current time k. It can be calculated using the following formula:

[0170]

[0171] Among them, the first error covariance matrix With the first system state vector There is a pre-built correspondence between them, the first error covariance matrix The diagonal elements in the vector represent the state vector of the first system. The estimation error variance of each component, for example, the diagonal elements of the first error covariance matrix at time k can be: The off-diagonal elements of the first error covariance matrix represent the error correlations between the components of the first system state vector. Therefore, the first error covariance matrix at time k... It can be calculated using the following formula:

[0172]

[0173] In the formula, This represents the first system state vector at time k. Represents the first system state vector at time k. The estimated value, Indicates calculation The mathematical expectation.

[0174] Update using the Kalman filter algorithm:

[0175] We can first predict the first error covariance matrix based on the current time k. and the first system observation matrix Determine the first Kalman gain matrix for the current time k. It can be calculated using the following formula:

[0176]

[0177] Based on the first Kalman gain matrix at the current time k First system observation matrix And the prediction first error covariance matrix Determine the first error covariance matrix for the current time. It can be calculated using the following formula:

[0178]

[0179] Based on the first Kalman gain matrix at the current time k First system observation vector First system observation matrix And predict the first system state vector Determine the first system state vector for the current time k. It can be calculated using the following formula:

[0180]

[0181] Therefore, based on the first system state vector at the current time k This will allow the heading angle error of the pedestrian navigation system to be corrected. Scale factor error of pedestrian odometer Zero bias error of the gyroscope in the inertial measurement unit The second error of the pedestrian navigation system is identified and designated as the target error.

[0182] In the above-described satellite error estimation strategy, the second reference trajectory obtained based on satellite navigation information, and the Kalman filter algorithm used to estimate the target error of the original trajectory based on the second reference trajectory, can accurately estimate the heading angle error of the pedestrian navigation system, the scaling factor error of the pedestrian odometry, and the zero bias error of the gyroscope in the inertial measurement unit. This provides an accurate and reliable error correction basis for the pedestrian navigation system to generate the original trajectory for the next moment.

[0183] Method 3 uses a target error estimation strategy that is typical of scenario error estimation strategies.

[0184] In one possible implementation, step 40 may include: when the target error estimation strategy is the typical scenario error estimation strategy, determining a corresponding set of error vectors based on the original trajectory, the set of error vectors including error vectors corresponding to each moment in the original trajectory, the error vectors including at least one of the following: attitude misalignment angle of the pedestrian navigation system, velocity error of the pedestrian navigation system, position error of the pedestrian navigation system, zero bias error of the gyroscope in the inertial measurement unit, and zero bias error of the accelerometer in the inertial measurement unit; determining constraint factors for the pedestrian navigation system based on the pedestrian path characteristics and the preset typical scenario path characteristics, the constraint factors including at least one of the following: inertial measurement unit pre-integration factor, zero velocity update factor, closed-loop pose factor, barometric altitude factor, and bias soft anchor factor; obtaining the target error vector for the next moment based on the constraint factors, the set of error vectors, and the objective function; and determining the target error vector as the target error.

[0185] For example, based on the original trajectory, the error vector corresponding to time i in the original trajectory can be... Represented as:

[0186]

[0187] In the formula, This represents the attitude misalignment angle of the pedestrian navigation system at time i. This represents the velocity error of the pedestrian navigation system at time i. Let represent the position error of the pedestrian navigation system at time i. This represents the zero-bias error of the gyroscope in the inertial measurement unit at time i. This represents the zero bias error of the accelerometer in the inertial measurement unit at time i.

[0188] Therefore, the set of error vectors corresponding to the original trajectory (including the error vectors corresponding to each time step in the original trajectory) can be represented as: , These represent the error vectors corresponding to each time point in the interval from time 0 to time N.

[0189] In this implementation method, the implementation methods of each constraint factor for the pedestrian navigation system are different based on the pedestrian path characteristics and the preset typical scene path characteristics. Furthermore, the determination of the constraint factors is also different based on different typical scene path characteristics. The following uses typical scenes such as playgrounds as examples to explain the determination method of each constraint factor.

[0190] Determination of the pre-integration factor for the inertial measurement unit:

[0191] Pre-integrating the detection data of the inertial measurement unit in the interval [i, j], the error vector is... Propagation over this interval satisfies:

[0192]

[0193] In the formula, This represents the state transition matrix, used in the ideal case to... Mapped to (Error vector at time j); This represents the noise driving matrix, used to map process noise onto the error vector at each time step; This indicates process noise.

[0194] This allows us to obtain the pre-integration factor of the inertial measurement unit. It can be calculated using the following formula:

[0195] ,

[0196] In the formula, This represents the pre-integration factor of the inertial measurement unit obtained from noise propagation. The corresponding covariance matrix, i.e. .

[0197] In this way, by using the inertial measurement unit's pre-integration factor to pre-integrate the detection data of the inertial measurement unit, the changes in attitude, velocity, and position errors between the two endpoints of the corresponding interval can be obtained, thereby enabling the constraint of the integral error of the error vector. The interval can be set according to actual needs, and this disclosure does not impose any restrictions on it.

[0198] Determining the zero-rate update factor:

[0199] When a moving target is detected to be stationary, the velocity error is approximately zero, thus allowing the zero-velocity update factor to be obtained. It can be calculated using the following formula:

[0200] ,

[0201] In the formula, Represents the zero-rate update factor obtained from noise propagation. The corresponding covariance matrix, i.e. , The variance represents the speed error.

[0202] In this way, by using the zero-speed update factor, when a moving target is detected to be stationary, the speed error at the current moment is set to zero, thereby constraining the speed error of the pedestrian navigation system.

[0203] Determination of closed-loop pose factor:

[0204] If a closed loop is detected in the original trajectory within the interval [i, j], the relative pose error change estimated from time i to time j based on the original trajectory can be compared with the relative pose error reference value from time i to time j in the typical scene path features obtained through temporal signature matching to obtain the closed loop pose factor. It can be calculated using the following formula:

[0205]

[0206] In the formula, This represents the change in relative position error between time i and time j, estimated from the original trajectory. This represents the reference value of the relative positional error between time i and time j in a typical scenario path feature. This represents the change in relative attitude error between time i and time j, estimated based on the original trajectory. This represents the reference value of the relative attitude error between time i and time j in a typical scenario path feature.

[0207] In this way, by utilizing the closed-loop pose factor to form a closed loop in the original trajectory, the eastward position, northward position, and heading are constrained by the difference between the relative pose error change between the two ends of the closed loop in the original trajectory and the relative pose error reference value between the two ends of the closed loop in the typical scenario path features. The altitude change and roll / pitch direction are also weakly constrained, thereby constraining the cumulative pose error.

[0208] Determination of barometric altitude factor:

[0209] When a change in the altitude of a moving target is detected, a barometric altitude factor can be obtained. It can be calculated using the following formula:

[0210]

[0211] In the formula, This represents the change in relative altitude error between time i and time j, estimated from the original trajectory. This represents the relative altitude error reference value detected by the barometer between time i and time j.

[0212] In this way, by using the barometric altitude factor, the error in the vertical direction can be constrained based on the difference between the relative altitude error reference value detected by the barometer and the relative altitude error change of the moving target in the original trajectory. This is especially suitable for scenarios with changes in altitude, such as multi-story buildings and mountainous areas.

[0213] Determining the bias soft anchor factor:

[0214] If the original trajectory is detected to form a closed loop in the interval [time i, time j], soft constraints can be obtained for the zero bias of the gyroscope and the zero bias of the accelerometer in the inertial measurement units at both ends of the closed loop, namely the bias soft anchor factor. It can be calculated using the following formula:

[0215]

[0216] In the formula, This represents the change in zero bias of the gyroscope in the inertial measurement unit between time i and time j. This represents the change in zero bias of the accelerometer in the inertial measurement unit between time i and time j.

[0217] In this way, by using the bias soft anchor factor to constrain the zero bias of the gyroscope and the zero bias of the accelerometer in the inertial measurement units at both ends of the closed loop, unreasonable sudden changes in the zero bias are suppressed, ensuring the continuity of the navigation trajectory and improving the stability and accuracy of the pedestrian navigation system in long-term operation.

[0218] Thus, after determining the constraint factors based on actual needs, the target error vector for the next time step can be obtained based on the constraint factors, the set of error vectors, and the objective function, and the target error vector can be determined as the target error.

[0219] The objective function is used to determine the target error vector for the next time step from the set of error vectors corresponding to the original trajectory using constraint factors. The objective function can be implemented using nonlinear least squares methods, such as the Gauss-Newton algorithm or the Levenberg-Marquardt algorithm; this disclosure does not impose any limitations on this. For example, the objective function can be:

[0220]

[0221] In the formula, This represents the target error vector for the next time step k+1. Represents the set of moments with zero velocity. This represents the set of closed-loop edges. This represents a robust kernel function used to suppress erroneous loops.

[0222] Each closed-loop pose factor is equipped with a corresponding robust kernel function and switchable weight variables. When the residual is greater than the preset value, the constraint weight can be automatically reduced, thereby effectively suppressing the impact of erroneous closed loops on the calculation results.

[0223] After determining the target error vector, it can be defined as the target error. In this way, the system can correct attitude misalignment angles based on the target error vector, significantly reducing attitude drift caused by gyroscope integration accumulation, improving the stability of direction judgment, and preventing overall rotation or distortion of the entire trajectory. Correcting velocity errors based on the target error vector can significantly reduce the cumulative deviation of integral displacement, improving the accuracy of step size and displacement calculations. Correcting position errors based on the target error vector can effectively curb trajectory deviations as distance increases. Correcting the zero-bias error of the gyroscope in the inertial measurement unit based on the target error vector can significantly improve the system's robustness and reduce the divergence rate of attitude calculation results. Correcting the zero-bias error of the accelerometer in the inertial measurement unit based on the target error vector can effectively reduce the cumulative displacement error introduced by acceleration measurement deviations.

[0224] In the above-described error estimation strategy for typical scenarios, the pedestrian path characteristics determined based on the original trajectory are matched with the preset typical scenario path characteristics. By setting constraint factors and using an objective function, the target error of the original trajectory is estimated. This enables accurate estimation of the attitude misalignment angle, velocity error, position error, zero bias error of the gyroscope in the inertial measurement unit, and zero bias error of the accelerometer in the inertial measurement unit. This provides an accurate and reliable error correction basis for the pedestrian navigation system to generate the original trajectory for the next moment.

[0225] Method 4: The target error estimation strategy is a constrained error estimation strategy.

[0226] In one possible implementation, step S40 may include: when the target error estimation strategy is the constraint error estimation strategy, determining the principal direction vector of the moving target at the current moment based on the original trajectory; determining the principal direction angle of the moving target at the current moment based on the principal direction vector; determining the change in the angle between the principal direction vector at the current moment and the principal direction vector at the previous moment based on the principal direction vector at the current moment and the principal direction vector at the previous moment; if the moving target is determined to be in a straight-line state based on the change in the angle and preset motion state conditions, determining the fourth error of the pedestrian navigation system using the Kalman filter algorithm based on the principal direction angle at the current moment, and determining the fourth error as the target error; wherein, the fourth error includes the principal direction angle error.

[0227] In one possible implementation, when the target error estimation strategy is the constraint error estimation strategy, determining the main direction vector of the moving target at the current moment based on the original trajectory may include: determining the position covariance matrix for the current moment based on the original trajectory; determining a first eigenvalue and a second eigenvalue based on the position covariance matrix at the current moment, wherein the first eigenvalue is not less than the second eigenvalue; and determining the first eigenvector corresponding to the first eigenvalue as the main direction vector of the moving target at the current moment.

[0228] In this implementation method, refer to Figure 7 Based on the original trajectory, determine the principal direction vector of the moving target at the current time k. The implementation can be done in the following ways:

[0229] First, based on the original trajectory, the two-dimensional position sequence of the moving target at time i in the original trajectory can be set as follows:

[0230] ,

[0231] In the formula, and These represent the eastward and northward positions of the original trajectory at time i in the n-system, respectively.

[0232] Taking a sliding window of length n as an example, within the sliding window of the interval [k-n+1, ​​k], the above position sequence... Principal component analysis can be performed in the following ways:

[0233] First, calculate the mean of the position sequence within the sliding window, which can be done using the following formula:

[0234]

[0235] Then, the points at each position in the sliding window are decentered, which can be calculated using the following formula:

[0236]

[0237] Establish a data matrix It can be calculated using the following formula:

[0238]

[0239] This yields the position covariance matrix for the current time k. It can be calculated using the following formula:

[0240]

[0241] After determining the position covariance matrix, we can then analyze the position covariance matrix at the current time k. Perform eigenvalue decomposition to determine the first eigenvalue. Second eigenvalue ,in, , This represents the variance of the position sequence along the principal direction. This represents the variance of the position sequence along a direction orthogonal to the principal direction. For example... Figure 7 As shown, the first eigenvalue The corresponding first feature vector The main direction vector of the motion of the target at the current time k is determined.

[0242] For example, after determining the principal direction vector of the moving target k at the current time... Then, based on the principal direction vector It is possible to determine the principal direction angle of the moving target at the current moment k. The implementation method can be:

[0243] First, correct the main direction vector. This includes the displacement vector between the beginning and end points of the sliding window. It can be represented as:

[0244]

[0245] exist In this case, let The principal direction vector is corrected, thereby eliminating the uncertainty of the sign of the principal direction vector and ensuring that the principal direction vector is consistent with the actual forward direction of the moving target.

[0246] Then, based on the corrected principal direction vector, the principal direction angle of the moving target at the current moment can be determined. It can be calculated using the following formula:

[0247]

[0248] In the formula, and These represent the principal direction vectors, respectively. The eastward and northward components.

[0249] For example, based on the principal direction vector at the current time k and the principal direction vector of the previous time k-1 Determine the change in the angle between the two. It can be calculated using the following formula:

[0250]

[0251] In this embodiment, the preset motion state conditions may include: the change in included angle is less than a preset direction change threshold and the trajectory linearity index is less than a preset trajectory linearity threshold, where the trajectory linearity index is the ratio of a second feature value to a first feature value. Then, based on the change in included angle... Given that the target is moving in a straight line according to the preset motion state conditions, based on the change in the included angle... Using the Kalman filter algorithm, the fourth error of the traveler navigation system is determined, and this fourth error is set as the target error. This can be achieved as follows:

[0252] First, the first eigenvalue determined based on the above steps. Second eigenvalue It is possible to determine the trajectory linearity index for the current time k. It can be calculated using the following formula:

[0253]

[0254] Then, in the change of the included angle Less than the preset directional change threshold ,Right now And trajectory linearity index Less than the preset trajectory linearity threshold ,Right now In this case, it can be determined that the moving target at the current time k is in a straight-line state. The preset direction change threshold and trajectory linearity threshold can be set according to actual needs, and this disclosure does not impose any restrictions on them.

[0255] In one possible implementation, when the moving target is determined to be in a straight-line state based on the angle change and preset motion state conditions, the fourth error of the pedestrian navigation system is determined using a Kalman filter algorithm based on the current principal direction angle, and the fourth error is determined as the target error. This may include: determining a second system state transition matrix and a second system observation matrix for the current time based on the principal direction angle at the current time, wherein a pre-built correspondence is established between the second system state transition matrix and the second system state vector, and a pre-built correspondence is established between the second system observation matrix and the second system observation vector, wherein the second system state vector includes a heading angle error, and the second system observation vector includes a heading angle error; determining the observation noise variance for the current time based on the trajectory linearity index; determining the predicted second system state vector for the current time based on the second system state vector at the previous time; and determining the predicted second system state vector for the current time based on the current... The system predicts the second system state vector and determines the heading observation residual for the current moment. Based on the second error covariance matrix of the previous moment, it determines the predicted second error covariance matrix for the current moment, where a pre-established correspondence is established between the second error covariance matrix and the second system state vector. Based on the second system observation matrix, the predicted second error covariance matrix, and the observation noise variance for the current moment, it determines the innovation covariance matrix. Based on the predicted second error covariance matrix, the second system observation matrix, and the innovation covariance matrix for the current moment, it determines the second Kalman gain matrix. Based on the predicted second system state vector, the second Kalman gain matrix, and the heading observation residual for the current moment, it determines the second system state vector for the current moment. Based on the second system state vector for the current moment, it determines the fourth error of the pedestrian navigation system and identifies this fourth error as the target error.

[0256] The second system state vector may include at least the heading angle error of the pedestrian navigation system, and the second system state vector can be expressed as follows: In the formula, This represents the heading angle error of the pedestrian navigation system. The remaining components of the second system state vector can be added according to actual needs, and this disclosure does not impose any restrictions on this.

[0257] The second system observation vector can be represented as: Then, based on the heading angle error of the pedestrian navigation system at the current time k... and principal direction vector An observation model for the pedestrian navigation system can be obtained. .

[0258] Therefore, the principal direction angle of k at the current time can be used as a basis. The second system state transition matrix for the current time k can be determined. Second system observation matrix Among them, the second system observation matrix You can set 1 at the position corresponding to the heading angle error component and 0 at the other positions.

[0259] For example, the trajectory linearity index determined at the current time k in the above steps. The observation noise variance for the current time k can be determined. It can be calculated using the following formula:

[0260]

[0261] In the formula, and This indicates a preset parameter value, which can be set according to actual needs, and this disclosure does not impose any restrictions on it.

[0262] For example, the steps of determining the fourth error of the traveler navigation system using the Kalman filter algorithm at the current time k, and identifying the fourth error as the target error, include:

[0263] Prediction using the Kalman filter algorithm:

[0264] We can first use the second system state vector from the previous time k-1. Determine the predicted state vector of the second system for the current time k. .

[0265] The predicted state vector of the second system based on the current time k The predicted heading angle error at the current time k can be obtained. Therefore, the predicted heading angle error is determined based on the current time k. and principal direction angle This allows us to determine the heading observation residual for the current time k. It can be calculated using the following formula:

[0266]

[0267] In the formula, This means normalizing the angle value to (- Interval.

[0268] Then, based on the second error covariance matrix of the previous time k-1... The second error covariance matrix of the prediction for the current time k can be determined. Among them, the second error covariance matrix With the second system state vector There are pre-built correspondences between them, as detailed above, and this disclosure will not elaborate further here.

[0269] Update using the Kalman filter algorithm:

[0270] We can first use the second system observation matrix at the current time k. Predicting the second error covariance matrix and observation noise variance Determine the innovation covariance matrix for the current time k. It can be calculated using the following formula:

[0271]

[0272] Based on the prediction second error covariance matrix at the current time k Second system observation matrix and innovation covariance matrix The second Kalman gain matrix for the current time k can be determined. It can be calculated using the following formula:

[0273]

[0274] Based on the prediction second error covariance matrix at the current time k Second system observation matrix Second Kalman gain matrix The second error covariance matrix for the current time k can be determined. It can be calculated using the following formula:

[0275]

[0276] The predicted state vector of the second system based on the current time k Second Kalman gain matrix Heading observation residuals This allows us to determine the second system state vector for the current time k. .

[0277] Then, based on the second system state vector at the current time k The heading angle error of the pedestrian navigation system The fourth error of the pedestrian navigation system is identified and designated as the target error. By correcting the heading angle error based on this fourth error, the system can effectively eliminate the cumulative heading drift generated by the inertial measurement unit during long-term operation. This ensures that even when no navigation information source is detected and the pedestrian path characteristics do not match the preset typical scenario path characteristics, the system can still maintain the long-term accuracy of attitude recursion and the stability of motion direction calculation.

[0278] In the above-described constraint error estimation strategy, based on principal component analysis of the position sequence of the moving target and the determination that the moving target is in a straight-line state, the Kalman filter algorithm is used to estimate the target error for the original trajectory. This enables accurate estimation of the heading angle error of the pedestrian navigation system, providing an accurate and reliable error correction basis for the pedestrian navigation system to generate the original trajectory for the next moment.

[0279] This disclosure provides an error estimation method and apparatus for pedestrian navigation systems. It acquires detection data from the inertial measurement unit (IMU) of the pedestrian navigation system at the current moment and the current navigation information source. Based on the detection data, an original trajectory is generated. Then, based on the current navigation information source and the original trajectory, a corresponding target error estimation strategy is dynamically selected from at least one preset error estimation strategy. Finally, based on the original trajectory and the target error estimation strategy, a target error is determined. This target error serves as the basis for error correction when the pedestrian navigation system generates the original trajectory for the next moment. This method is not limited to relying on a single sensor or algorithm for error estimation. Instead, it comprehensively judges various situations based on different forms of external and scene information, such as the current original trajectory and current navigation information sources like digital maps and satellite navigation information. It selects a corresponding target error estimation strategy from multiple error estimation strategies (map error estimation strategy, satellite error estimation strategy, typical scene error estimation strategy, constraint error estimation strategy, etc.) for error estimation. Therefore, it can perform targeted analysis of the original trajectory based on different reference information to accurately determine various errors in the pedestrian navigation system when facing different situations. Compared to related technologies, this disclosure does not simply switch between different algorithms in different scenarios, but establishes a unified multi-source reference-assisted error estimation framework that can uniformly transform different forms of external information and scenario information into the basis for error estimation, significantly improving the positioning accuracy of pedestrian navigation systems in different scenarios, while taking into account both robustness and applicability.

[0280] This disclosure also provides an error estimation device for a pedestrian navigation system, the device comprising:

[0281] The data acquisition module is used to acquire the detection data of the inertial measurement unit in the pedestrian navigation system at the current moment and the navigation information source at the current moment, wherein the navigation information source includes digital maps and / or satellite navigation information;

[0282] A trajectory generation module is used to generate an original trajectory based on the detection data;

[0283] The strategy selection module is used to dynamically select the corresponding target error estimation strategy from at least one preset error estimation strategy based on the navigation information source and the original trajectory at the current time.

[0284] The error determination module is used to determine the target error based on the original trajectory and the target error estimation strategy. The target error is used as the basis for error correction when the pedestrian navigation system generates the original trajectory for the next moment.

[0285] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0286] It should be noted that although the above embodiments have been used as examples to illustrate the error estimation method and apparatus for pedestrian navigation systems, those skilled in the art will understand that this disclosure is not limited thereto. In fact, users can flexibly set each step and module according to their personal preferences and / or actual application scenarios, as long as it conforms to the technical solution of this disclosure.

[0287] This disclosure also provides an error estimation device for a pedestrian navigation system, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above method.

[0288] This disclosure also provides a non-volatile computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0289] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0290] Figure 8 This is a block diagram illustrating an apparatus 1900 for error estimation in a pedestrian navigation system according to an exemplary embodiment. For example, apparatus 1900 may be provided as a server or terminal device. (Refer to...) Figure 8The apparatus 1900 includes a processing component 1922, which further includes one or more processors, and memory resources represented by memory 1932 for storing instructions, such as application programs, that can be executed by the processing component 1922. The application programs stored in memory 1932 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1922 is configured to execute instructions to perform the methods described above.

[0291] Device 1900 may also include a power supply component 1926 configured to perform power management of device 1900, a wired or wireless network interface 1950 configured to connect device 1900 to a network, and an input / output interface 1958 (I / O interface). Device 1900 can operate on an operating system, such as Windows Server, stored in memory 1932. TM macOS X TM Unix TM Linux TM FreeBSD TM Or similar.

[0292] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by a processing component 1922 of the device 1900 to perform the above-described method.

[0293] Computer-readable storage media can be tangible devices capable of holding and storing programs / instructions used by instruction execution devices. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0294] The computer program (or computer-readable program instructions) described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage medium in the respective computing / processing device.

[0295] The computer program (or computer program instructions) used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing state information from the computer-readable program instructions to implement various aspects of this disclosure.

[0296] Various aspects of this disclosure are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0297] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0298] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0299] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those shown in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0300] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. An error estimation method for pedestrian navigation systems, characterized in that, The method includes: The system acquires the detection data from the inertial measurement unit in the pedestrian navigation system at the current moment and the navigation information source at the current moment, wherein the navigation information source includes digital maps and / or satellite navigation information; The original trajectory is generated based on the detection data; Based on the navigation information source and the original trajectory at the current moment, the corresponding target error estimation strategy is dynamically selected from at least one preset error estimation strategy; Based on the original trajectory and the target error estimation strategy, the target error is determined, and the target error is used as the basis for error correction when the pedestrian navigation system generates the original trajectory for the next moment.

2. The method according to claim 1, characterized in that, The step of dynamically selecting a corresponding target error estimation strategy from at least one preset error estimation strategy based on the navigation information source and the original trajectory at the current moment includes: If the pedestrian path features determined based on the original trajectory match the preset typical scene path features, the typical scene error estimation strategy is determined as the target error estimation strategy. If the pedestrian path features do not match the preset typical scene path features and the navigation information source is the digital map, the map error estimation strategy will be determined as the target error estimation strategy. If the pedestrian path features do not match the preset typical scenario path features and the navigation information source is the satellite navigation information, the satellite error estimation strategy will be determined as the target error estimation strategy. If the navigation information source is not detected and the pedestrian path features do not match the preset typical scenario path features, the constraint error estimation strategy will be determined as the target error estimation strategy.

3. The method according to claim 2, characterized in that, The step of determining the target error based on the original trajectory and the target error estimation strategy includes: When the target error estimation strategy is the map error estimation strategy, a first reference trajectory is determined for the original trajectory based on the digital map and the original trajectory; a first error of the pedestrian navigation system is determined based on the geometric relationship between the first reference trajectory and the original trajectory, and the first error is determined as the target error; wherein, the first error includes at least one of the following: the track angle error of the pedestrian navigation system, the scale coefficient error of the pedestrian odometry, and the zero bias error of the gyroscope in the inertial measurement unit; When the target error estimation strategy is the satellite error estimation strategy, a second reference trajectory is determined based on the satellite navigation information for the original trajectory; based on the second reference trajectory and the original trajectory, a first trajectory segment corresponding to the original trajectory and a second trajectory segment corresponding to the second reference trajectory are determined; based on the first trajectory segment and the second trajectory segment, the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment are determined; based on the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment, a Kalman filter algorithm is used to determine the second error of the pedestrian navigation system, and the second error is determined as the target error; wherein, the second error includes at least one of the following: the heading angle error of the pedestrian navigation system, the scaling factor error of the pedestrian odometer, and the zero bias error of the gyroscope in the inertial measurement unit; When the target error estimation strategy is the typical scenario error estimation strategy, based on the original trajectory, a corresponding set of error vectors is determined. This set includes error vectors corresponding to each moment in the original trajectory. Each error vector includes at least one of the following: the pedestrian navigation system's attitude misalignment angle, velocity error, position error, zero-bias error of the gyroscope in the inertial measurement unit (IMU), and zero-bias error of the accelerometer in the IMU. Based on the pedestrian path characteristics and the preset typical scenario path characteristics, constraint factors for the pedestrian navigation system are determined. These constraint factors include at least one of the following: IMU pre-integration factor, zero-velocity update factor, closed-loop pose factor, barometric altitude factor, and bias soft anchor factor. Based on the constraint factors, the set of error vectors, and the objective function, the target error vector for the next moment is obtained. This target error vector is then defined as the target error. When the target error estimation strategy is the constraint error estimation strategy, the main direction vector of the moving target at the current moment is determined based on the original trajectory; the main direction angle of the moving target at the current moment is determined based on the main direction vector; the change in the angle between the main direction vector at the current moment and the main direction vector at the previous moment is determined based on the main direction vector at the current moment and the main direction vector at the previous moment; if the moving target is determined to be in a straight-line state based on the change in the angle and the preset motion state conditions, the fourth error of the pedestrian navigation system is determined using the Kalman filter algorithm based on the main direction angle at the current moment, and the fourth error is determined as the target error; wherein, the fourth error includes the main direction angle error.

4. The method according to claim 3, characterized in that, The step of determining a first reference trajectory for the original trajectory based on the digital map and the original trajectory includes: Based on the digital map and the original trajectory, determine the observation probability and state transition probability of each observation point in the original trajectory relative to the road in the digital map; The first reference trajectory is determined based on the observation probability and the state transition probability corresponding to each of the aforementioned location observation points; The step of determining the first error of the pedestrian navigation system based on the geometric relationship between the first reference trajectory and the original trajectory includes: Based on the geometric relationship between the original trajectory and the first reference trajectory, determine the cross product of the total displacement of the original trajectory and the total displacement of the first reference trajectory, the first modulus of the total displacement of the original trajectory, and the second modulus of the total displacement of the first reference trajectory; The track angle error of the pedestrian navigation system is determined based on the cross product result, the first module length, and the second module length. The scale coefficient error of the pedestrian odometer is determined based on the ratio of the first module length to the second module length; If the original trajectory meets the preset conditions, the zero bias error of the gyroscope in the inertial measurement unit is determined based on the track angle error of the pedestrian navigation system.

5. The method according to claim 3, characterized in that, The step of determining the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometry at the current moment based on the first trajectory segment and the second trajectory segment includes: Based on the first trajectory segment and the second trajectory segment, determine the length ratio of the first trajectory segment to the second trajectory segment, as well as the included angle between the first trajectory segment and the second trajectory segment; The proportional factor error of the pedestrian odometer is determined based on the length ratio, and the heading angle error of the pedestrian navigation system is determined based on the angle between the first trajectory segment and the second trajectory segment. Specifically, based on the heading angle error of the pedestrian navigation system at the current moment and the scaling factor error of the pedestrian odometry, a second error of the pedestrian navigation system is determined using a Kalman filter algorithm, and this second error is defined as the target error, including: Based on the heading angle error of the pedestrian navigation system and the scaling factor error of the pedestrian odometer at the current moment, a first system state transition matrix and a first system observation matrix are determined for the current moment. A pre-established correspondence is established between the first system state transition matrix and the first system state vector, and a pre-established correspondence is established between the first system observation matrix and the first system observation vector. The first system state vector includes at least one of the following: the heading angle error of the pedestrian navigation system, the position information corresponding to the second reference trajectory, the scaling factor error of the pedestrian odometer, and the zero-bias error of the gyroscope in the inertial measurement unit. The first system observation vector includes at least one of the following: the heading angle error of the pedestrian navigation system, the position information corresponding to the second reference trajectory, and the scaling factor error of the pedestrian odometer. Based on the first system state transition matrix at the current moment and the first system state vector at the previous moment, the predicted first system state vector for the current moment is determined. Based on the first system state transition matrix at the current moment and the first error covariance matrix at the previous moment, the prediction first error covariance matrix for the current moment is determined. A pre-built correspondence is set between the first error covariance matrix and the first system state vector. Based on the prediction first error covariance matrix and the first system observation matrix at the current moment, determine the first Kalman gain matrix for the current moment; Based on the first Kalman gain matrix, the first system observation vector, the first system observation matrix, and the predicted first system state vector at the current moment, the first system state vector for the current moment is determined. Based on the first system state vector at the current moment, the second error of the pedestrian navigation system is determined, and the second error is determined as the target error.

6. The method according to claim 3, characterized in that, Determining the principal direction vector of the moving target at the current moment based on the original trajectory includes: Based on the original trajectory, determine the position covariance matrix for the current moment; Based on the position covariance matrix at the current moment, determine the first eigenvalue and the second eigenvalue, wherein the first eigenvalue is not less than the second eigenvalue; The first eigenvector corresponding to the first eigenvalue is determined as the main direction vector of the moving target at the current moment; The preset motion state conditions include: the angle change is less than a preset direction change threshold and the trajectory linearity index is less than a preset trajectory linearity threshold, wherein the trajectory linearity index is the ratio of the second feature value to the first feature value; Wherein, when the moving target is determined to be in a straight-line state based on the angle change and preset motion state conditions, the fourth error of the pedestrian navigation system is determined using a Kalman filter algorithm based on the main direction angle at the current moment, and the fourth error is determined as the target error, including: Based on the principal direction angle at the current moment, a second system state transition matrix and a second system observation matrix are determined for the current moment. A pre-built correspondence is set between the second system state transition matrix and the second system state vector, and a pre-built correspondence is set between the second system observation matrix and the second system observation vector. The second system state vector includes the heading angle error, and the second system observation vector includes the principal direction angle. Based on the trajectory linearity index, the observation noise variance for the current moment is determined; Based on the second system state vector of the previous moment, determine the predicted second system state vector for the current moment; Based on the predicted second system state vector and the principal heading angle at the current moment, the heading observation residual for the current moment is determined; Based on the second error covariance matrix of the previous time step, the prediction second error covariance matrix for the current time step is determined. A pre-built correspondence is set between the second error covariance matrix and the second system state vector. Based on the second system observation matrix at the current moment, the prediction second error covariance matrix, and the observation noise variance, the innovation covariance matrix for the current moment is determined; Based on the prediction second error covariance matrix, the second system observation matrix, and the innovation covariance matrix at the current moment, determine the second Kalman gain matrix for the current moment; Based on the predicted second system state vector, the second Kalman gain matrix, and the heading observation residual at the current moment, the second system state vector for the current moment is determined; Based on the second system state vector at the current moment, the fourth error of the pedestrian navigation system is determined, and the fourth error is identified as the target error.

7. An error estimation device for a pedestrian navigation system, characterized in that, The device includes: The data acquisition module is used to acquire the detection data of the inertial measurement unit in the pedestrian navigation system at the current moment and the navigation information source at the current moment, wherein the navigation information source includes digital maps and / or satellite navigation information; A trajectory generation module is used to generate an original trajectory based on the detection data; The strategy selection module is used to dynamically select the corresponding target error estimation strategy from at least one preset error estimation strategy based on the navigation information source and the original trajectory at the current time. The error determination module is used to determine the target error based on the original trajectory and the target error estimation strategy. The target error is used as the basis for error correction when the pedestrian navigation system generates the original trajectory for the next moment.

8. An error estimation device for a pedestrian navigation system, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A non-volatile computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images