A pedestrian positioning method and system based on building structured information assistance

Abstract

The application relates to the technical field of pedestrian navigation, in particular to a pedestrian positioning method and system based on building structured information assistance, which comprises the following steps: initializing a track angle through a dynamic window self-adaptive track angle initialization method according to pedestrian motion structured features; judging straight movement and turning according to the pedestrian motion structured features and the initial track angle; performing multi-dimensional error collaborative correction through an improved self-adaptive Kalman filter; performing multi-dimensional collaborative correction through a flexible constraint model and a state-by-state Kalman filter when turning; dynamically compensating state quantities after straight movement correction and state quantities after turning correction according to environmental interference structured features; obtaining compensated state quantities; and positioning the pedestrian through the compensated state quantities. The application can fully exploit multi-dimensional structured information, dynamically optimize correction strategies for different motion states, and realize adaptive compensation of environmental interference for pedestrian positioning.

Description

Technical Field

[0001] This invention relates to the field of pedestrian navigation technology, and in particular to a pedestrian positioning method and system based on building structure information assistance. Background Technology

[0002] Traditional foot-mounted MEMS-IMU inertial positioning systems are susceptible to heading drift, gait dynamics interference, and insufficient constraints from the building environment. In hospital settings, accurate positioning of patients and medical equipment can shorten emergency response time; in factory environments, real-time tracking of workers and materials can optimize production processes; and special scenarios such as fire fighting place stringent requirements on the continuity and anti-interference capabilities of positioning. However, indoor environments present problems such as satellite signal blockage, complex electromagnetic interference, and diverse building structures, rendering traditional positioning technologies relying on global navigation satellite systems completely ineffective, thus necessitating dedicated indoor positioning solutions.

[0003] Existing methods do not fully utilize structured information. Most schemes simply combine some environmental features without systematically extracting multi-dimensional structured features such as building geometry (e.g., corridor width, intersection angles), pedestrian movement (e.g., step length fluctuations, step frequency stability), and environmental interference (e.g., magnetic field strength, metal distribution). They also lack targeted constraint mechanisms. Error correction strategies are simplistic. Traditional Kalman filtering and improved algorithms often use fixed models and fail to dynamically adjust the correction logic based on different pedestrian movement states such as walking straight or turning. Furthermore, they often focus on single-dimensional error correction, neglecting the collaborative optimization of attitude, speed, position, and sensor drift. Environmental interference handling capabilities are weak. Soft and hard iron interference generated by indoor metal components and electrical equipment can lead to heading angle deviations. Existing compensation schemes are mostly static corrections, unable to dynamically adapt to changes in interference intensity and distribution, resulting in insufficient anti-interference stability. Some high-performance algorithms have excessively high computational costs, making it difficult to achieve real-time operation on embedded devices or wearable terminals.

[0004] Existing indoor pedestrian positioning technologies suffer from problems such as high deployment costs, poor environmental adaptability, inaccurate error correction, and weak anti-interference capabilities, and cannot simultaneously meet the practical application requirements of high precision, high stability, low cost, and easy deployment. Summary of the Invention

[0005] This invention aims to at least solve one of the technical problems existing in related technologies. To this end, this invention provides a pedestrian positioning method and system based on building structured information, which can fully mine multi-dimensional structured information, dynamically optimize correction strategies for different motion states, and achieve pedestrian positioning with adaptive compensation for environmental interference.

[0006] This invention provides a pedestrian localization method based on building structure information assistance, comprising:

[0007] S1: Extract the geometric structured features of buildings, the structured features of pedestrian movement, and the structured features of environmental disturbances through a multi-dimensional structured information extraction model;

[0008] S2: Based on the structured characteristics of pedestrian movement, the track angle is initialized using a dynamic window adaptive track angle initialization method;

[0009] S3: Determine whether to go straight or turn based on the structured characteristics of pedestrian movement and the initial track angle;

[0010] S4: When going straight, the adaptive Kalman filter is improved based on the geometric and structural features of the building. The improved adaptive Kalman filter is used to perform multi-dimensional error collaborative correction to obtain the state quantity after straight-line correction.

[0011] S5: When turning, a flexible constraint model and a state-by-state Kalman filter are constructed based on the geometric structural features of the building. Multi-dimensional collaborative correction is performed through the flexible constraint model and the state-by-state Kalman filter to obtain the state variables after turning correction.

[0012] S6: Dynamically compensate the state quantities after straight-line correction and turning correction based on the structured characteristics of environmental interference to obtain the compensated state quantities, and use the compensated state quantities to locate pedestrians.

[0013] Furthermore, the multi-dimensional structured information extraction model includes a building geometry structured layer, a pedestrian movement structured layer, and an environmental interference structured layer.

[0014] The building's geometric structure features are extracted through the building's geometric structure layer. These features include basic straight-line movement features, 90° turn features, -90° turn features, 180° turn features, corridor width, room entrance and exit coordinates, and statistical features of corridor branch angles.

[0015] By extracting the biological characteristics of pedestrian foot movement through the pedestrian movement structure layer, the structured features of pedestrian movement are obtained, including stride length fluctuation coefficient, foot landing posture angle variance, stride frequency stability when walking straight, pedestrian stride frequency and stride length stability.

[0016] Based on the distribution pattern of metal components within the building, the environmental interference structured layer establishes a spatial distribution model of the magnetic field interference area and extracts the magnetic field strength threshold and interference direction characteristics of the interference area.

[0017] Furthermore, the dynamic window adaptive track angle initialization method includes:

[0018] S21: Select valid initial steps based on step size fluctuation coefficient and attitude angle variance when the foot lands;

[0019] If the step size fluctuation coefficient of three consecutive steps is less than or equal to the step size fluctuation coefficient threshold and the landing attitude angle variance is less than or equal to the landing attitude angle variance threshold, then it is a valid initial step.

[0020] S22: Calculate the initial track angle of the effective initial step based on the mean of the sliding window and the weighting coefficient. The length of the sliding window is dynamically adjusted according to the pedestrian's step frequency. The weighting coefficient is positively correlated with the stability of the step length.

[0021] Furthermore, the straight-line judgment includes:

[0022] Straight-line determination is made based on heading angle, angular velocity variance, and step frequency stability;

[0023] If the heading rate of change is less than or equal to the heading rate of change threshold, the angular velocity variance is less than or equal to the vertical axis angular velocity variance threshold, and the step frequency change is less than or equal to the step frequency fluctuation threshold, and all three conditions are met and the change is greater than or equal to 2 steps, then the course is going straight; otherwise, it is suspected of turning.

[0024] If the conditions are not met for three consecutive steps, then it is time to turn.

[0025] Furthermore, the turning determination includes preparing to turn, turning in progress, and turning completed;

[0026] If the absolute value of the vertical axis angular velocity is greater than or equal to the first vertical axis angular velocity threshold and continues for one step, then it is time to prepare for a turn.

[0027] If the rate of change of heading is greater than or equal to the rate of change of heading threshold and the frequency of steps is reduced by greater than or equal to 10% compared to when going straight, then it is turning;

[0028] If the absolute value of the vertical axis angular velocity is less than or equal to the second vertical axis angular velocity threshold, and the heading remains stable, then the turn is complete.

[0029] The first vertical axis angular velocity threshold is greater than the second vertical axis angular velocity threshold.

[0030] Furthermore, improving the adaptive Kalman filter based on the structural features of the building geometry includes:

[0031] S41: Add building boundary constraint terms to the measurement equation of adaptive Kalman filtering, and combine them with the corridor width W to limit the lateral error of the navigation position to within [-W / 2, W / 2]. When the position solution result exceeds [-W / 2, W / 2], introduce the position error measurement value.

[0032] S42: Dynamically adjust the filter gain based on the pedestrian heading error covariance;

[0033] When the pedestrian heading error covariance is greater than the first stability threshold, increase the gain to speed up the correction.

[0034] When the pedestrian heading error covariance is less than the second stability threshold, the gain is reduced to suppress noise.

[0035] The first stability threshold is greater than the second stability threshold.

[0036] Furthermore, the flexible constraint model includes:

[0037] S51: Based on building topology information, pre-store the standard turning angle of the current area;

[0038] S52: The turning angle is calculated using angular velocity integral plus step size weighting;

[0039] S53: Calculate the heading correction value based on the turning angle, standard turning angle, and flexible compensation amount. The flexible compensation amount is dynamically adjusted according to the attitude stability during the turning phase.

[0040] Furthermore, the state-specific Kalman filtering includes zero-speed stage Kalman filtering and non-zero-speed stage Kalman filtering;

[0041] The Kalman filter measurement equation at zero speed includes speed error, heading error, and a turn angle consistency constraint. The turn angle consistency constraint is the difference between the turn angle and the standard turn angle, and its calculation expression is:

[0042]

[0043] in, Measurement status at zero speed. For speed error, For heading error, For the turning angle, For the floor function, The standard step size for heading;

[0044] In the non-zero velocity phase, Kalman filtering introduces angular velocity error constraints into the measurement equations, and uses the difference between the gyroscope constant drift and the observed angular velocity value for auxiliary correction. The calculation expression is as follows:

[0045]

[0046] in, Measurement state during the non-zero speed phase. Angular velocity, The measured angular velocity is from the micro inertial measurement unit. This represents the constant drift of the gyroscope.

[0047] Furthermore, step S6 includes:

[0048] When the magnetic field sensor of the micro inertial measurement unit detects that the magnetic field strength is greater than or equal to the first magnetic field strength threshold, and the current position is in the pre-stored metal component area, it is determined to be a magnetic field interference state.

[0049] Based on historical undisturbed heading and magnetic field direction, dynamic compensation is performed, and the calculation expression is as follows:

[0050]

[0051] in, For the compensated course, To maintain a stable course before interference, This is the current magnetic field direction angle. This is the reference magnetic field direction angle under undisturbed conditions. This is the compensation coefficient;

[0052] The magnetic field strength is dynamically adjusted; the greater the magnetic field strength, the better. The larger;

[0053] Under interference conditions, the compensated heading angle is used as the measurement input for Kalman filtering; when the magnetic field strength is less than the second magnetic field strength threshold for two consecutive steps, the compensation mode is exited and normal heading calculation is resumed.

[0054] The first magnetic field strength threshold is greater than the second magnetic field strength threshold.

[0055] This invention also provides a pedestrian positioning system based on building structure information assistance, for executing the aforementioned pedestrian positioning method based on building structure information assistance, comprising:

[0056] The feature extraction module extracts architectural geometric structured features, pedestrian movement structured features, and environmental interference structured features through a multi-dimensional structured information extraction model.

[0057] An initialization module initializes the track angle based on the structured characteristics of pedestrian movement using a dynamic window adaptive track angle initialization method.

[0058] The determination module determines whether a pedestrian is going straight or turning based on the pedestrian's structured movement characteristics and initial trajectory angle.

[0059] A straight-line correction module, which improves the adaptive Kalman filter based on the geometric structural features of the building when moving straight, and performs multi-dimensional error collaborative correction through the improved adaptive Kalman filter to obtain the state quantity after straight-line correction;

[0060] A turning correction module is provided. When turning, the turning correction module constructs a flexible constraint model and a state-by-state Kalman filter based on the building's geometric structural features. Through the flexible constraint model and the state-by-state Kalman filter, multi-dimensional collaborative correction is performed to obtain the state variables after turning correction.

[0061] The compensation positioning module dynamically compensates the state quantities after straight-line correction and the state quantities after turning correction based on the structured characteristics of environmental interference, and obtains the compensated state quantities, and uses the compensated state quantities to locate pedestrians.

[0062] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects:

[0063] This invention, based on the classic strapdown inertial navigation system (SINS) solution and zero-velocity correction (ZUPT), breaks through the limitations of traditional single-constraint approach relying solely on fixed-angle turns in straight lines. It enriches the constraint dimensions by constructing a three-layer structured information extraction framework encompassing building geometry, motion characteristics, and environmental disturbance resistance. Through dynamic adaptive track angle initialization, multi-parameter straight-line and turning determination, and flexible constraints, it improves model adaptability. Furthermore, it enhances anti-interference capabilities and error correction accuracy by utilizing structured features of environmental disturbances. Finally, through an improved adaptive Kalman filter algorithm, it achieves coordinated correction of heading, velocity, position, and sensor drift errors. Without the need for additional sensors, it significantly improves positioning accuracy and heading stability in complex building scenarios, providing more reliable theoretical and technical support for the engineering application of micro-inertial navigation systems in indoor navigation.

[0064] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0065] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0066] Figure 1 This is a flowchart illustrating a pedestrian positioning method based on building structure information provided by the present invention.

[0067] Figure 2 This is a schematic diagram of a pedestrian positioning system based on building structure information provided by the present invention.

[0068] Figure label:

[0069] 101. Feature extraction module; 102. Initialization module; 103. Judgment module; 104. Straight-line correction module; 105. Turning correction module; 106. Compensation positioning module. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but cannot be used to limit the scope of this invention.

[0071] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0072] The following is combined with Figures 1 to 2 This invention describes a pedestrian positioning method and system based on building structure information assistance.

[0073] like Figure 1 As shown, a pedestrian localization method based on building structure information includes:

[0074] S1: Extract the geometric structured features of buildings, the structured features of pedestrian movement, and the structured features of environmental disturbances through a multi-dimensional structured information extraction model;

[0075] The multi-dimensional structured information extraction model includes a building geometry structured layer, a pedestrian movement structured layer, and an environmental disturbance structured layer.

[0076] The building's geometric structure features are extracted through the building's geometric structure layer. These features include basic straight-line movement features, 90° turn features, -90° turn features, 180° turn features, corridor width, room entrance and exit coordinates, statistical features of corridor branch angles, pedestrian walking frequency, and step length stability, forming macroscopic geometric boundary constraints.

[0077] By extracting the biological characteristics of foot movement through the pedestrian movement structure layer, the structured features of pedestrian movement are obtained, including the stride length fluctuation coefficient, the variance of the posture angle when the foot lands, and the stability of the stride frequency when walking straight, thus forming micro-motion state constraints.

[0078] The environmental interference structured layer establishes a spatial distribution model of the magnetic field interference area based on the distribution pattern of metal components within the building, extracts the magnetic field strength threshold and interference direction characteristics of the interference area, and forms an anti-interference compensation constraint.

[0079] The building geometry structure layer, pedestrian movement structure layer, and environmental interference structure layer are synchronized in real time through data fusion, solving the problems of single constraints and weak anti-interference ability of traditional solutions.

[0080] S2: Based on the structured characteristics of pedestrian movement, the track angle is initialized using a dynamic window adaptive track angle initialization method;

[0081] Existing technologies correct the initial heading solely based on the average of the first N track angles, making them susceptible to initial gait instability. This invention initializes the track angle using a dynamic window adaptive track angle initialization method, including:

[0082] S21: Select valid initial steps based on step size fluctuation coefficient and landing attitude angle variance;

[0083] If the step size fluctuation coefficient of three consecutive steps is less than or equal to the step size fluctuation coefficient threshold and the landing attitude angle variance is less than or equal to the landing attitude angle variance threshold, then it is a valid initial step.

[0084] By using stride consistency and posture stability as dual criteria, abnormal gaits such as starting and sudden stops are eliminated to obtain effective initial steps;

[0085] In some specific embodiments of the present invention, the step size fluctuation coefficient threshold is 5%, and the landing attitude angle variance threshold is 3°.

[0086] S22: Calculate the initial track angle of the effective initial step based on the mean of the sliding window and the weighting coefficient. The length of the sliding window is dynamically adjusted according to the pedestrian's step frequency. The weighting coefficient is positively correlated with the stability of the step length.

[0087] The expression for calculating the initial track angle is:

[0088]

[0089] in, The initial track angle for the pedestrian. For pedestrians Walking track angle, For the first Step weight coefficient, This represents the total number of steps.

[0090] The formula for calculating the weighting coefficient is:

[0091]

[0092] in, For the first Step and average step size The difference.

[0093] In some specific embodiments of the present invention, the window length is dynamically adjusted according to the pedestrian's walking frequency. When the walking frequency is 1.5-2.5Hz, the window length is 8-12 steps.

[0094] In navigation calculation, heading error For the heading angle of the micro inertial measurement unit (MIMU) Angle of pedestrian track The difference, ;

[0095] like Real-time correction is performed, among which... This is the error accumulation threshold.

[0096] In some specific embodiments of the present invention .

[0097] when Greater than When the zero-speed phase is reached, real-time correction is triggered, rather than relying solely on the zero-speed phase, thus achieving a dual correction mechanism of zero-speed correction plus real-time threshold triggering.

[0098] S3: Determine whether to go straight or turn based on the structured characteristics of pedestrian movement and the initial track angle;

[0099] Straight-line judgment includes:

[0100] Straight-line determination is made based on heading angle, angular velocity variance, and step frequency stability;

[0101] If the heading rate of change is less than or equal to the heading rate of change threshold, the angular velocity variance is less than or equal to the vertical axis angular velocity variance threshold, and the step frequency change is less than or equal to the step frequency fluctuation threshold, and all three conditions are met and continuously greater than or equal to 2 steps, then it is a suspected turn.

[0102] If the condition is not met for 3 consecutive steps, then it is time to turn.

[0103] Existing technologies determine straight-line movement solely based on the amplitude of heading changes, which is prone to misjudgment; for example, slight swaying may be mistaken for a turn. This invention integrates three parameters—heading change rate, angular velocity variance, and stride frequency stability—for straight-line determination. The calculation expression is as follows:

[0104]

[0105] in, For the first The change in heading during the step, The threshold for the magnitude of the heading change. Let be the variance of angular velocity. The vertical axis angular velocity variance threshold. For the first The change in step frequency. This is the step frequency fluctuation threshold.

[0106] When all three conditions are met simultaneously and for two or more consecutive steps, the vehicle is determined to go straight; if any one condition is not met, the vehicle enters a suspected turning state. If the condition is not met for three consecutive steps, the vehicle is determined to turn, which significantly improves the robustness of the straight-going determination.

[0107] In some specific embodiments of the present invention , , .

[0108] Turn detection includes preparing to turn, turning in progress, and turning completed;

[0109] If the absolute value of the vertical axis angular velocity is greater than or equal to the first vertical axis angular velocity threshold and continues for one step, then it is time to prepare for a turn.

[0110] If the rate of change of heading is greater than or equal to the rate of change of heading threshold and the frequency of steps is reduced by greater than or equal to 10% compared to when going straight, then it is turning;

[0111] If the absolute value of the vertical axis angular velocity is less than or equal to the second vertical axis angular velocity threshold, and the heading remains stable, then the turn is complete.

[0112] The first vertical axis angular velocity threshold is greater than the second vertical axis angular velocity threshold.

[0113] In some specific embodiments of the present invention, the first vertical axis angular velocity threshold is 0.3, the heading change rate threshold is 0.5, and the second vertical axis angular velocity threshold is 0.1.

[0114] Existing technologies determine turns solely based on changes in heading, failing to differentiate between the stages of preparing to turn, turning in progress, and turn completion, which can easily lead to errors in correction timing. This invention divides the stages based on three parameters: angular velocity, rate of change of heading, and stride frequency.

[0115] Preparing to turn: Vertical axis angular velocity The absolute value is greater than or equal to 0.3 rad / s and lasts for 1 step;

[0116] Turning: Rate of change of heading Furthermore, the step frequency decreased by ≥10% compared to walking in a straight line;

[0117] Turn complete: The absolute value is ≤0.1rad / s, and the heading remains stable;

[0118] That is, the amplitude of two consecutive heading changes is ≤1°, and the difference between the building branch angle matching turning angle and the building branch angle is ≤5°.

[0119] S4: When going straight, the adaptive Kalman filter is improved based on the geometric and structural features of the building. The improved adaptive Kalman filter is used to perform multi-dimensional error collaborative correction to obtain the state quantity after straight-line correction.

[0120] To address the problem of insufficient dynamic response caused by the fixed gain of traditional Kalman filters, this invention employs an adaptive Kalman filter (AKF) that preserves the state variables, including three-dimensional attitude error, three-dimensional velocity error, three-dimensional position error, gyroscope constant drift, and accelerometer constant drift.

[0121] The adaptive Kalman filter is improved by incorporating architectural geometric structural features. A building boundary constraint term is added to the measurement equation, specifically, by combining the corridor width W, the lateral error of the navigation position (perpendicular to the direction of travel) is limited to the range [-W / 2, W / 2]. When the position calculation result exceeds this range, the position error measurement value is automatically introduced. The calculation expression is as follows:

[0122]

[0123]

[0124] in, In straight-line measurement mode, For speed error, For heading error, The positional error is obtained from the building boundary constraints. For location;

[0125] When the pedestrian heading error covariance is greater than the first stability threshold, increase the gain to speed up the correction.

[0126] When the pedestrian heading error covariance is less than the second stability threshold, the gain is reduced to suppress noise.

[0127] The first stability threshold is greater than the second stability threshold.

[0128] In some specific embodiments of the present invention, the first stability threshold is: The second stability threshold is ;

[0129] Based on the heading error covariance Dynamically adjust filter gain ;

[0130] when At that time, increase To speed up the calibration process;

[0131] when When, decrease To suppress noise, the expression for calculating the filter gain is:

[0132]

[0133]

[0134] in, For filter gain, Here is the state error covariance matrix. For the observation matrix, This is the transpose of the matrix. To measure the noise variance, The initial measurement noise variance, This is used as a reference covariance threshold.

[0135] S5: When turning, a flexible constraint model and a state-by-state Kalman filter are constructed based on the geometric structural features of the building. Multi-dimensional collaborative correction is performed through the flexible constraint model and the state-by-state Kalman filter to obtain the state variables after turning correction.

[0136] Flexible constraint models include:

[0137] S51: Based on building topology information, pre-store the standard turning angle of the current area;

[0138] Existing technologies use fixed 90°, -90° and 180° corrections, without considering non-standard building turns and gait interference, such as 85° and 95°.

[0139] Based on building topology information, the standard turning angle of the current area is pre-stored. Such as the angle between corridor branches, or the turning angle corresponding to the opening direction of a room door;

[0140] S52: The turning angle is calculated using angular velocity integral plus step size weighting. The calculation expression is as follows:

[0141]

[0142] in, For the turning angle, for The vertical axis angular velocity at time t. The start time of the turn. The end time of the turn. For the duration of the turn, This represents the average step length during the turning phase. The average step size is used to correct the integral error caused by gait shortening by using step size weights;

[0143] S53: Calculate the heading correction value based on the turning angle, standard turning angle, and flexible compensation amount. The flexible compensation amount is dynamically adjusted according to the attitude stability during the turning phase.

[0144] The formula for calculating the heading correction value is:

[0145]

[0146] in, This is the heading correction value. For the heading before the turn, For the floor function, For the turning angle, The standard step size for heading. This is a flexible compensation amount;

[0147] The dynamic adjustment is based on the posture stability during the turning phase. The larger the posture variance, the greater the compensation amount, with a maximum of no more than 3°, to solve the errors caused by non-standard turns and gait interference.

[0148] State-based Kalman filtering includes zero-speed phase Kalman filtering and non-zero-speed phase Kalman filtering;

[0149] The Kalman filter measurement equation at zero speed includes speed error, heading error, and a turn angle consistency constraint. The turn angle consistency constraint is the difference between the turn angle and the standard turn angle, and its calculation expression is:

[0150]

[0151] in, Measurement status at zero speed. For speed error, For heading error, For the turning angle, For the floor function, The standard step size for heading;

[0152] During the non-zero speed phase of the turn, the Kalman filter introduces an angular velocity error constraint into the measurement equation. This error is corrected using the difference between the gyroscope constant drift and the observed angular velocity value. The calculation expression is as follows:

[0153]

[0154] in, Measurement state during the non-zero speed phase. Angular velocity, The measured angular velocity is from the micro inertial measurement unit. This represents the constant drift of the gyroscope.

[0155] Improve the real-time ability to suppress heading errors during turns.

[0156] S6: Dynamically compensate the state quantities after straight-line correction and turning correction based on the structured characteristics of environmental interference to obtain the compensated state quantities, and use the compensated state quantities to locate pedestrians.

[0157] To address magnetic field interference caused by metal components within buildings, such as beams, columns, and pipes, this invention incorporates building structural information (the distribution locations of metal components are pre-stored in the system) to add adaptive compensation for magnetic field interference.

[0158] When the magnetic field sensor of the micro inertial measurement unit detects that the magnetic field strength is greater than or equal to the first magnetic field strength threshold, and the current position is in the pre-stored metal component area, it is determined to be a magnetic field interference state.

[0159] Based on historical undisturbed heading and magnetic field direction, dynamic compensation is performed, and the calculation expression is as follows:

[0160]

[0161] in, For the compensated course, To maintain a stable course before interference, This is the current magnetic field direction angle. This is the reference magnetic field direction angle under undisturbed conditions. For compensation coefficient,

[0162] The magnetic field strength is dynamically adjusted; the greater the magnetic field strength, the better. The larger, The value range is [0.8-1];

[0163] Under interference conditions, the compensated heading angle is used as the measurement input for Kalman filtering; when the magnetic field strength is less than the second magnetic field strength threshold for two consecutive steps, the compensation mode is exited and normal heading calculation is resumed.

[0164] The first magnetic field strength threshold is greater than the second magnetic field strength threshold.

[0165] In some specific embodiments of the present invention, the first magnetic field strength threshold is 0.5 Gauss and the second magnetic field strength threshold is 0.3 Gauss.

[0166] like Figure 2 As shown, a pedestrian positioning system based on building structure information is used to execute the aforementioned pedestrian positioning method based on building structure information, including:

[0167] The feature extraction module 101 extracts the geometric structured features of buildings, the structured features of pedestrian movement, and the structured features of environmental interference through a multi-dimensional structured information extraction model;

[0168] Initialization module 102 initializes the track angle based on the structured characteristics of pedestrian movement using a dynamic window adaptive track angle initialization method;

[0169] The determination module 103 makes a straight-ahead or turning decision based on the pedestrian's structured movement characteristics and initial track angle;

[0170] When the straight-line correction module 104 is straight, it improves the adaptive Kalman filter based on the geometric and structural features of the building. The improved adaptive Kalman filter is used to perform multi-dimensional error collaborative correction to obtain the state quantity after straight-line correction.

[0171] When turning, the turning correction module 105 constructs a flexible constraint model and a state-by-state Kalman filter based on the building's geometric structural features. Through the flexible constraint model and the state-by-state Kalman filter, multi-dimensional collaborative correction is performed to obtain the state variables after turning correction.

[0172] The compensation positioning module 106 dynamically compensates the state quantities after straight-line correction and the state quantities after turning correction based on the structured characteristics of environmental interference, obtains the compensated state quantities, and uses the compensated state quantities to locate pedestrians.

[0173] Through the collaborative work of the above modules, this invention enriches the constraint dimensions by constructing a three-layer structured information extraction framework of building geometry, motion features, and environmental disturbance resistance; it improves model adaptability through dynamic adaptive track angle initialization, multi-parameter straight-line and turning determination, and flexible constraints; it enhances anti-interference capability and error correction accuracy through structured features of environmental disturbance; and it achieves coordinated correction of heading error, velocity error, position error, and sensor drift through an improved adaptive Kalman filter algorithm. Without the need for additional sensor assistance, it can significantly improve positioning accuracy and heading stability in complex building scenarios, providing more reliable theoretical and technical support for the engineering application of micro inertial navigation systems in indoor navigation.

[0174] Compared with traditional methods, this invention improves the heading error suppression rate by more than 40% and reduces the positioning error by 35%. It can still maintain stable navigation performance in complex scenarios such as irregular turns, magnetic field interference, and unstable gait. It provides a more engineering-value technical solution for pedestrian positioning in buildings and expands the application boundaries of inertial navigation systems in indoor scenarios.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A pedestrian positioning method based on building structured information assistance, characterized in that, Comprise: S1: extract building geometric structured features, pedestrian motion structured features and environmental interference structured features through multi-dimensional structured information extraction model; S2: initialize the track angle according to the pedestrian motion structured features through the dynamic window adaptive track angle initialization method; S3: determine straight and turning according to the pedestrian motion structured features and the initial track angle; S4: improve adaptive Kalman filtering according to building geometric structured features when straight, and correct multi-dimensional errors through improved adaptive Kalman filtering to obtain corrected state quantities when straight; S5: construct a flexible constraint model and a state-dependent Kalman filter according to building geometric structured features when turning, and correct multi-dimensional errors through the flexible constraint model and the state-dependent Kalman filter to obtain corrected state quantities when turning; S6: dynamically compensate the corrected state quantities when straight and the corrected state quantities when turning according to the environmental interference structured features to obtain compensated state quantities, and perform pedestrian positioning through the compensated state quantities.

2. The method of claim 1, wherein, The multi-dimensional structured information extraction model comprises a building geometric structured layer, a pedestrian motion structured layer and an environmental interference structured layer; The building geometric structured features are extracted through the building geometric structured layer, including basic straight line travel features, 90° turning features, -90° turning features, 180° turning features, corridor width, room entrance and exit position coordinates, and corridor fork angle statistical features; The pedestrian motion structured features are obtained by extracting the biological feature rules of pedestrian foot motion through the pedestrian motion structured layer, including step length fluctuation coefficient, foot landing posture angle variance, step frequency stability when straight, pedestrian step frequency and step length stability; The environmental interference structured layer establishes a spatial distribution model of the magnetic field interference area based on the distribution rules of the metal components in the building, and extracts the magnetic field intensity threshold and interference direction features of the interference area.

3. The pedestrian positioning method based on building structure information assistance according to claim 2, characterized in that, The dynamic window adaptive track angle initialization method comprises: S21: screen effective initial steps according to the step length fluctuation coefficient and the foot landing posture angle variance; If the step length fluctuation coefficients of the last 3 steps are less than or equal to the step length fluctuation coefficient threshold and the foot landing posture angle variances are less than or equal to the foot landing posture angle variance threshold, they are effective initial steps; S22: calculate the initial track angle of the effective initial steps according to the sliding window mean and the weight coefficient, the sliding window length is dynamically adjusted according to the pedestrian step frequency, and the weight coefficient is positively correlated with the step length stability.

4. The pedestrian positioning method based on building structure information assistance according to claim 2, characterized in that, The straight determination comprises: Determine straight according to the heading angle, angular velocity variance and step frequency stability; If the heading change rate is less than or equal to the heading change rate threshold, the angular velocity variance is less than or equal to the vertical axis angular velocity variance threshold, and the step frequency variation is less than or equal to the step frequency fluctuation threshold, and all of them are satisfied and last for more than or equal to 2 steps, it is straight, otherwise it is suspected to be turning; If it does not satisfy for 3 steps, it is turning.

5. A pedestrian positioning method based on building structure information assistance according to claim 4, characterized in that, The turning determination comprises preparing to turn, turning and turning completion; If the absolute value of the vertical axis angular velocity is greater than or equal to the first vertical axis angular velocity threshold and lasts for 1 step, it is preparing to turn; If the heading change rate is greater than or equal to the heading change rate threshold and the step frequency is reduced by more than or equal to 10% compared with that when straight, it is turning. If the absolute value of the vertical axis angular velocity is less than or equal to the second vertical axis angular velocity threshold value, and the heading is stable, the turn is completed. The first vertical axis angular velocity threshold value is greater than the second vertical axis angular velocity threshold value.

6. The method of claim 1, wherein the method further comprises: Improving adaptive Kalman filtering according to building geometric structured features includes: S41: Adding a building boundary constraint term to the measurement equation of adaptive Kalman filtering, combining the corridor width W, limiting the lateral error of the navigation position within [-W / 2, W / 2], and introducing a position error measurement value when the position solution result exceeds [-W / 2, W / 2]; S42: Dynamically adjusting the filter gain according to the pedestrian heading error covariance; When the pedestrian heading error covariance is greater than a first stability threshold value, the gain is increased to speed up the correction; When the pedestrian heading error covariance is less than a second stability threshold value, the gain is reduced to suppress noise; The first stability threshold value is greater than the second stability threshold value.

7. The method of claim 1, wherein the method further comprises: The flexible constraint model includes: S51: Based on the building topology information, prestore the standard turn angle of the current area; S52: Calculate the turn angle by integrating the angular velocity and weighting the step size; S53: Calculate the heading correction value according to the turn angle, the standard turn angle, and the flexible compensation amount, which is dynamically adjusted according to the attitude stability of the turn stage.

8. The method of claim 1, wherein the method further comprises: The state-dependent Kalman filter includes a zero-speed stage Kalman filter and a non-zero-speed stage Kalman filter; The measurement equation of the zero-speed stage Kalman filter contains a speed error, a heading error, and a turn angle consistency constraint term, and the turn angle consistency constraint term is the difference between the turn angle and the standard turn angle, and the calculation expression is: wherein, is a zero speed phase measurement state, is a speed error, is a heading error, is a turn angle, is a rounding function, is a standard step size for heading. The non-zero-speed stage Kalman filter introduces an angular velocity error constraint in the measurement equation, which is assisted by the difference between the constant drift of the gyroscope and the angular velocity observation value, and the calculation expression is: wherein, is a non-zero velocity phase measurement state, is an angular velocity, is a micro inertial measurement unit measured angular velocity, is a gyroscope constant drift.

9. The method of claim 1, wherein the method further comprises: S6 steps include: When the magnetic field sensor of the micro inertial measurement unit detects that the magnetic field intensity is greater than or equal to a first magnetic field intensity threshold value, and the current position is in a pre-stored metal component area, it is determined that it is in a magnetic field interference state; Based on the historical non-interference heading and the magnetic field direction, dynamic compensation is performed, and the calculation expression is: wherein, is the compensated heading, is the stable heading before disturbance, is the current magnetic field direction angle, is the magnetic field reference direction angle without disturbance, is the compensation coefficient; According to the dynamic adjustment of the magnetic field intensity, the greater the magnetic field intensity the greater; In the interference state, the compensated heading angle is taken as the measurement input of the Kalman filter; when the magnetic field intensity is less than a second magnetic field intensity threshold value and lasts for 2 steps, the compensation mode is exited, and the normal heading solution is restored; The first magnetic field intensity threshold value is greater than the second magnetic field intensity threshold value.

10. A pedestrian positioning system based on building structured information assistance, characterized in that, To perform a building structured information assisted pedestrian positioning method as claimed in any one of claims 1 to 9, comprising: a feature extraction module, which extracts building geometric structured features, pedestrian motion structured features, and environmental interference structured features through a multi-dimensional structured information extraction model; an initialization module, which initializes the track angle through a dynamic window adaptive track angle initialization method according to the pedestrian motion structured features; a determination module, which determines straight running and turning according to the pedestrian motion structured features and the initial track angle; The straight correction module improves adaptive Kalman filtering according to building geometric structured features when straightening, and performs multi-dimensional error collaborative correction through the improved adaptive Kalman filtering to obtain the state quantity after straight correction; The turning correction module constructs a flexible constraint model and a sub-state Kalman filter according to building geometric structured features when turning, and performs multi-dimensional collaborative correction through the flexible constraint model and the sub-state Kalman filter to obtain the state quantity after turning correction; The compensation positioning module dynamically compensates the state quantity after straight correction and the state quantity after turning correction according to environmental interference structured features to obtain the state quantity after compensation, and performs pedestrian positioning through the state quantity after compensation.

Images