High-precision positioning system in complex environment based on multi-signal source cooperation

Abstract

The application discloses a complex environment high-precision positioning system based on multi-signal source cooperation, belongs to the technical field of cooperative positioning and specifically comprises the following steps: a multi-source signal extraction module receives satellite, cellular base station and wireless access point signals, extracts carrier phase, propagation time delay and multipath reflection intensity distribution; a reflection weight generation module identifies environment characteristics according to the multipath reflection intensity and generates a fusion weight set containing a satellite phase dominant factor and the like; a quantity constraint conversion module converts the above measurement values into three types of distance constraint quantities; a constraint fusion field construction module nonlinearly superimposes the three types of constraint quantities according to weights, and generates a three-dimensional probability density field; and an elevation coordinate solving module fuses air pressure data to compress a solution space, and iteratively converges to obtain a positioning coordinate; the application effectively improves positioning precision and reliability in a complex environment through multi-source signal cooperation.

Description

Technical Field

[0001] This invention relates to the field of collaborative positioning technology, and more specifically to a high-precision positioning system for complex environments based on multi-signal source collaboration. Background Technology

[0002] With the rapid development of information technology and intelligent equipment, high-precision positioning technology has become a core supporting technology in fields such as intelligent transportation, drone management and control, IoT sensing, and emergency rescue. In complex environments such as urban canyons, indoor spaces, and underground facilities, the demand for positioning is experiencing explosive growth, and its accuracy and reliability directly affect the implementation effectiveness of key applications such as autonomous driving decision-making, indoor navigation services, and industrial IoT collaboration. Currently, multi-source signal fusion positioning, due to its ability to overcome the environmental limitations of a single signal source, has become an important research direction for improving positioning performance in complex scenarios, attracting widespread attention from academia and industry.

[0003] In existing technologies, satellite navigation system-based positioning methods can provide high accuracy in open environments, but are susceptible to building obstruction and multipath effects, resulting in significant performance degradation in densely populated urban areas or indoor environments. Cellular base station positioning leverages wide-area coverage to achieve regional-level positioning, but its accuracy is limited by base station deployment density and signal propagation models, making it difficult to meet centimeter-level or sub-meter-level requirements. Wireless access point positioning is suitable for indoor scenarios, achieving positioning through signal strength or time difference of arrival, but faces stability issues caused by multipath interference and dynamic environmental changes. To overcome the shortcomings of single technologies, existing research has attempted to fuse signals from multiple sources, including satellites, base stations, and wireless access points, and optimize positioning results through algorithms such as Kalman filtering and particle filtering, which has improved positioning robustness in complex environments to some extent.

[0004] However, existing multi-source fusion positioning technologies still have significant limitations. They lack the ability to dynamically adapt to environmental characteristics and often employ fixed-weight fusion strategies, failing to adjust the contribution of each signal source in real time based on spatial reflection characteristics. Especially in complex building environments, reflected and refracted signals can easily lead to distance measurement deviations, and existing methods struggle to effectively separate direct paths from interference paths. Furthermore, in terms of elevation information fusion, they often rely on single barometric pressure sensors or satellite elevation data, lacking a probability density field optimization mechanism under multiple constraints, resulting in a significant decrease in positioning accuracy in scenarios such as multi-story buildings and underground spaces. Summary of the Invention

[0005] The purpose of this invention is to provide a high-precision positioning system for complex environments based on multi-signal source collaboration, and to solve the following technical problems:

[0006] Existing multi-source fusion positioning technology is insufficient for dynamic adaptation to environmental characteristics. Using fixed weights makes it impossible to adjust the contribution of signal sources in real time. In complex building environments, it is difficult to separate direct sunlight and interference paths, leading to distance measurement errors. Furthermore, elevation fusion relies on a single data source.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] A high-precision positioning system for complex environments based on multi-signal source collaboration includes:

[0009] The multi-source signal extraction module is used to receive carrier signals transmitted by the satellite navigation system, downlink pilot signals of cellular base stations, and broadcast signals of wireless access points, and extract carrier phase measurement values ​​of satellite signals, propagation delay measurement values ​​of cellular base station signals, and multipath reflection intensity distribution of wireless access point signals from the physical layer raw data;

[0010] The reflection weight generation module is used to identify the spatial reflection characteristics of the current environment based on the multipath reflection intensity distribution, and dynamically generate a signal feature fusion weight set that matches the spatial reflection characteristics. The weight set includes a satellite phase dominance factor, a base station delay compensation factor, and a wireless signal path suppression factor.

[0011] The quantity constraint conversion module is used to convert satellite carrier phase measurement values ​​into first-type distance constraint values, cellular base station propagation delay measurement values ​​into second-type distance constraint values, and wireless access point signals into third-type distance constraint values ​​through path suppression factors.

[0012] The constraint fusion field construction module is used to nonlinearly superimpose three types of distance constraint quantities based on the signal feature fusion weight set to generate a three-dimensional spatial probability density field.

[0013] The elevation coordinate solution module integrates barometric elevation data to compress the probability density field solution space and iteratively converges to the positioning coordinate solution along the direction of field strength variation.

[0014] As a further aspect of the present invention: in the reflection weight generation module, the process of identifying the spatial reflection characteristics of the current environment based on the multipath reflection intensity distribution is as follows:

[0015] Calculate the proportion of multipath components of satellite signals within a unit time window, and determine a high reflection environment when the proportion exceeds a preset threshold;

[0016] Analyze the histogram of the angle of arrival distribution of the downlink pilot signal of the cellular base station. When the kurtosis of the histogram is lower than the critical value and there are continuous strong reflection pulses, it is determined to be an open indoor environment.

[0017] By monitoring the intensity abrupt change event sequence of the wireless access point broadcast signal and combining it with the attitude angle change rate output by the inertial navigation device, the transition area where the moving body crosses the boundary of the building structure can be identified.

[0018] The signal feature fusion weight set enhances the role of the base station delay compensation factor in high-reflection environments and activates the wireless signal path suppression factor calculation module in open indoor environments.

[0019] As a further aspect of the present invention: the generation process of the path inhibition factor is as follows:

[0020] A pre-defined database of building material reflection characteristics stores phase reversal shift modes caused by metal surfaces, signal attenuation slope parameters caused by glass media, and multipath scattering distribution templates generated by concrete walls.

[0021] The phase jump characteristics of the received signal are compared with the database recording pattern in real time. When a phase reverse offset matching the metal reflection pattern is detected, a negative phase compensation operator is generated and injected into the third type of distance constraint quantity reconstruction process.

[0022] When the signal strength attenuation slope corresponding to the glass attenuation characteristics is identified, the amplitude scaling ratio of the distance constraint is dynamically adjusted according to the number of electromagnetic wave penetration layers and the attenuation slope parameter.

[0023] For template matching scenarios involving multipath scattering in concrete, path separation technology is used to extract the phase value of the main reflection component to replace the original measurement value.

[0024] As a further aspect of the present invention: in the quantity constraint conversion module, the specific conversion process of the distance constraint quantity is as follows:

[0025] A dual-frequency ionospheric delay correction is applied to the satellite carrier phase measurement, and a tropospheric refraction compensation is calculated based on the satellite elevation angle to generate a first-type range constraint.

[0026] Decode the orthogonal frequency division multiplexing symbols in the downlink pilot signal of the cellular base station, infer the signal propagation delay by the phase difference between adjacent subcarriers, and generate the second type of distance constraint by combining the base station's geographical location information;

[0027] The preamble waveform of the wireless access point broadcast signal is captured, the phase rotation components of the direct path and the reflected path are separated, the path suppression factor is applied to reduce the contribution of the reflected path, and the third type of distance constraint is reconstructed.

[0028] As a further aspect of the present invention: the process of generating a three-dimensional spatial probability density field in the constrained fusion field construction module is as follows:

[0029] A three-dimensional cubic grid coordinate system is established based on the position prediction points output by the inertial navigation device; the first type of distance constraint is mapped to the direction of the satellite line-of-sight vector to form a cone-shaped constraint domain, and the cone opening angle is dynamically adjusted by the satellite elevation confidence.

[0030] The second type of distance constraint constructs an annular constraint band along the azimuth of the cellular base station, and the width of the annular band is determined by the accuracy of the base station signal time measurement.

[0031] The third type of distance constraint quantity generates a spherical constraint body with the physical location of the wireless access point as the center. The radius of the sphere is determined by the distance value after path suppression factor calibration.

[0032] Based on the signal feature fusion weight set, the three sets of constraint structures are spatially weighted and fused. The weight of the cone constraint domain is positively correlated with the satellite phase dominance factor, the weight of the ring constraint band is adjusted by the base station delay compensation factor, and the scaling factor of the spherical constraint body is controlled by the wireless signal path suppression factor, thus forming a probability density field isosurface distribution model.

[0033] As a further aspect of the present invention: the specific process of iteratively converging to the positioning coordinate solution along the direction of field intensity change in the elevation coordinate calculation module is as follows:

[0034] Read the altitude data collected by the barometric pressure sensor and compress the vertical dimension of the probability density field solution space to the horizontal slice layer corresponding to the altitude;

[0035] Perform field strength driven position search within the constrained solution space, initialize the search point as the inertial navigation predicted position, calculate the comprehensive matching error between the point and the three types of distance constraints, the comprehensive matching error is a weighted synthesis of the projection residual of the conical constraint domain, the distance deviation of the annular constraint zone and the radius error of the spherical constraint body;

[0036] The search point is moved along the negative gradient direction of the matching error. The step size is dynamically adjusted according to the historical positioning stability. After each move, the elevation constraint is updated synchronously, and the vertical deviation compensation between the current position and the horizontal slice layer is recalculated. When the angle between the trajectory vectors formed by three consecutive moves is less than the direction change threshold, it is determined that a closed loop has been formed. The iteration is terminated and the coordinates of the minimum matching error point within the loop are output.

[0037] As a further aspect of the present invention: the elevation coordinate calculation module also includes feedback optimization of spatial reflection characteristics after each successful positioning, specifically as follows:

[0038] Each time a positioning is successful, a multipath reflection intensity distribution map and corresponding weight set parameters are recorded. The map includes a histogram of the proportion of satellite signal multipath components, a cellular base station angle of arrival spread curve, and a distribution of wireless access point intensity abrupt events.

[0039] An environmental feature index library is established, and the feature vectors of the multipath reflection intensity distribution map are associated and stored with the weight set parameters. When the similarity between the newly acquired multipath reflection intensity distribution map and the historical feature vector exceeds the matching threshold, the weight calculation process is skipped and the associated historical weight set is directly loaded.

[0040] Regularly perform spatial overlay analysis on the positioning trajectory point cloud and building information model to identify signal reflection anomalies in areas where the trajectory crosses walls; correct the parameters of the building material reflection feature database based on the material distribution around the anomalies, focusing on optimizing the glass medium attenuation slope parameter and the concrete multipath scattering distribution template.

[0041] As a further aspect of the present invention: the calling rule for the historical weight set is as follows:

[0042] When the environment recognition module triggers historical pattern matching, the fast positioning channel is activated and historical weight set parameters are loaded; the grid size control strategy in the probability density field construction stage is adjusted, the side length of the three-dimensional cube grid is set as a monotonically increasing function of historical positioning accuracy, and sparse grids are used in areas with small historical positioning errors;

[0043] In the position search phase, the heading angle output by the inertial navigation device is introduced as the guidance vector, and a sector search interval is constructed with the current position as the origin. The sector angle is adaptively adjusted according to the rate of change of the heading angle, and the movement direction of the search point is restricted within the sector interval to reduce computational complexity.

[0044] As a further aspect of the present invention: when the satellite signal fails, the joint positioning mode of the cellular base station signal and the wireless access point signal is activated, and the propagation delay measurement value of the cellular base station and the reconstructed distance constraint of the wireless access point are aligned with the clock reference. The alignment process uses the frame timing deviation of the cellular base station to compensate for the receiver clock drift.

[0045] Based on displacement data from an inertial navigation device, a smoothing constraint equation for the motion trajectory is constructed. The equation parameters include velocity continuity and acceleration boundary conditions. The smoothing constraint equation is then transformed into a probability density field with additional boundary conditions to restrict the position search path to within the kinematically feasible solution space.

[0046] When satellite signal recovery is detected, the weight of the satellite phase dominance factor is gradually increased until the system is fully switched to multi-source fusion mode.

[0047] The beneficial effects of this invention are:

[0048] This invention addresses existing technological bottlenecks through a multi-source signal coordination mechanism, significantly improving positioning performance in complex environments. Based on multipath reflection intensity distribution, it identifies spatial reflection characteristics and dynamically generates a fusion weight set including satellite phase dominance factors, base station delay compensation factors, and wireless signal path suppression factors. This enhances base station compensation in high-reflection environments and activates wireless suppression in open indoor environments, replacing fixed-weight strategies to achieve real-time adaptation of signal source contributions and significantly improving dynamic response capabilities in complex environments. Utilizing a building material reflection feature database, it implements phase compensation, attenuation adjustment, and principal component extraction for materials such as metal, glass, and concrete, accurately separating direct and interference paths and effectively eliminating distance measurement deviations in complex building environments. Through dual-frequency correction and subcarrier phase difference back-calculation, three types of distance constraints are generated, weighted and fused into a three-dimensional probability density field. This field is then combined with barometric pressure data to compress the solution space and iteratively converge, forming a multi-constraint positioning solution mechanism that significantly improves positioning accuracy in multi-story buildings and underground spaces. Simultaneously, a feedback optimization mechanism that uses historical data matching to quickly call weights and correct the material database further enhances positioning stability and efficiency, fully meeting the high-precision positioning requirements in complex scenarios. Attached Figure Description

[0049] The invention will now be further described with reference to the accompanying drawings.

[0050] Figure 1 This is a schematic diagram of the modules of the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please see Figure 1 As shown, this invention is a high-precision positioning system for complex environments based on multi-signal source collaboration, comprising:

[0053] The multi-source signal extraction module is responsible for simultaneously receiving signals transmitted from multiple signal sources and extracting key features from the raw physical layer data. This module receives carrier signals transmitted by the satellite navigation system, downlink pilot signals from cellular base stations, and broadcast signals from wireless access points. Through signal acquisition and analysis, it extracts carrier phase measurements of satellite signals, propagation delay measurements of cellular base station signals, and multipath reflection intensity distributions of wireless access point signals from the raw physical layer data. These extracted features provide fundamental data support for subsequent modules, with the multipath reflection intensity distribution directly reflecting the intensity variation of the signal after reflection from the environment during propagation.

[0054] The reflection weight generation module identifies the spatial reflection characteristics of the current environment based on the multipath reflection intensity distribution output by the multi-source signal extraction module. By analyzing the intensity distribution pattern of multipath signals, this module can determine whether the environment belongs to a high-reflection area, an open indoor area, or a transition area between building structures, and dynamically generates a set of signal feature fusion weights accordingly. This set includes a satellite phase dominance factor, a base station delay compensation factor, and a wireless signal path suppression factor. The values ​​of each factor are adjusted in real time according to environmental characteristics to ensure that the signal fusion strategy is adapted to the current environment.

[0055] The quantity constraint conversion module transforms measurements from different signal sources into a unified range constraint form. Specifically, this module converts satellite carrier phase measurements into Type I range constraints, cellular base station propagation delay measurements into Type II range constraints, and, combined with the path suppression factor output by the reflection weight generation module, processes and reconstructs the wireless access point signal into Type III range constraints. These three types of range constraints together constitute the basic spatial constraints for positioning calculation.

[0056] The constraint fusion field construction module uses the signal feature fusion weight set output by the reflection weight generation module to nonlinearly superimpose the three types of distance constraint quantities obtained by the magnitude constraint conversion module. By weighted integration of spatial constraint information from different signal sources, this module generates a three-dimensional spatial probability density field, which reflects the probability of each location as a positioning result in a spatial distribution.

[0057] The elevation coordinate calculation module integrates elevation data collected by barometric pressure sensors, using this data to compress the solution space of the three-dimensional probability density field, thus narrowing the positioning search range. Subsequently, the module performs iterative calculations along the direction of field strength variation in the probability density field, gradually converging to the optimal positioning coordinates, and finally outputting accurate positioning results.

[0058] In the reflection weight generation module, the process of identifying the spatial reflection characteristics of the current environment based on the multipath reflection intensity distribution includes three-dimensional judgment logic, and the accurate classification of environment types is achieved through cross-validation of features from multiple signal sources.

[0059] To identify high-reflectivity environments, the module first sets a fixed time window. The duration of this window is determined based on a combination of the signal sampling rate and the dynamic rate of environmental change, balancing real-time performance and identification stability. Within each time window, the module separates and calculates the multipath and direct signal components of the satellite signal, and calculates the proportion of multipath components in the total signal energy. When this proportion exceeds a preset threshold, the current environment is determined to be a high-reflectivity environment. This threshold is set based on measured data from typical high-reflectivity scenarios (such as dense building complexes and metal structure factories), effectively distinguishing the strength of multipath interference.

[0060] For determining an open indoor environment, the module focuses on the angle of arrival (AHA) distribution characteristics of the downlink pilot signal from the cellular base station. First, signal processing techniques are used to analyze the AHA distribution histogram, which reflects the energy distribution of the signal arriving at the receiver from different directions. When the kurtosis of the histogram is below a critical value, it indicates a relatively dispersed AHA distribution, meaning the signal originates from multiple directions rather than a single direct path. Simultaneously, if consecutive strong reflected pulses are detected, and the time intervals and intensity variations of these pulses conform to indoor reflection patterns, the environment is comprehensively determined to be an open indoor environment. In such environments, there are typically few obstructions, but reflections from walls and ceilings create multi-directional signal paths.

[0061] In identifying transition zones where mobile objects cross building structural boundaries, the module uses abrupt changes in the intensity of the wireless access point broadcast signal as the core clue. By continuously monitoring sudden increases or decreases in signal strength, a sequence of intensity abrupt events is generated. These abrupt changes are often related to crossing structures such as walls, doors, and windows. Simultaneously, the module retrieves the rate of change of attitude angles output by the inertial navigation system. When the rate of change of attitude angles exceeds the normal movement threshold and is synchronized with the intensity abrupt event in time, it can be confirmed that the mobile object is in the transition zone crossing the building structural boundary. The signal propagation characteristics in this zone change rapidly with position, requiring dynamic adjustment of the fusion strategy.

[0062] Based on the above environmental identification results, the signal feature fusion weight set will be adjusted accordingly. In high-reflection environments, due to severe multipath interference with satellite signals, the module will actively enhance the weight of the base station delay compensation factor, allowing the cellular base station signal to play a more dominant role in the positioning calculation. In open indoor environments, the multipath effect of wireless access point signals is relatively controllable, and the module will activate the wireless signal path suppression factor calculation module to improve the positioning contribution of the wireless signal by suppressing interference from reflection paths.

[0063] The generation process of the path suppression factor relies on a pre-set database of building material reflection characteristics. This database contains three types of core data: the phase reversal mode unique to metal surfaces, which is the typical feature that the phase of the signal will be reversed by 180 degrees after being reflected by metal; the signal attenuation slope parameter corresponding to glass medium, which quantifies the rate at which the signal intensity changes with the thickness when it penetrates the glass; and the multipath scattering distribution template generated by concrete walls, which records the distribution law of the multipath signal intensity and time delay formed after the signal is reflected by concrete.

[0064] In real-time processing, the module compares the phase transition characteristics of the received signal with the metal reflection patterns in the database. When a phase reversal that perfectly matches the metal reflection pattern is detected, a negative phase compensation operator is immediately generated and injected into the reconstruction process of the third type of distance constraint to compensate for the phase error caused by metal reflection.

[0065] When the signal strength attenuation slope is identified as matching the characteristics of the glass medium, the module first estimates the number of glass layers through which the electromagnetic wave penetrates, and then, in conjunction with the attenuation slope parameters in the database, dynamically adjusts the amplitude scaling ratio of the third type of distance constraint to ensure that the distance calculation is not affected by the glass penetration loss.

[0066] For scenarios involving concrete multipath scattering distribution templates, the module activates path separation technology. This technology distinguishes between the primary reflection component and stray reflection components based on signal time delay and phase characteristics. The phase value of the primary reflection component is extracted to replace the original measurement value, thereby reducing the interference of concrete multipath scattering on distance constraints. Through this material-specific suppression strategy, the path suppression factor effectively improves the positioning reliability of wireless access point signals in complex building environments.

[0067] In the quantity constraint conversion module, the conversion process of distance constraints employs a differentiated processing mechanism for different signal source characteristics. For satellite carrier phase measurements, a dual-frequency ionospheric delay correction is first applied to offset the influence of the ionosphere on signal transmission through the propagation differences of signals at different frequencies. Simultaneously, a tropospheric refraction compensation is calculated based on the satellite elevation angle. Because the path length of the signal through the troposphere varies with different satellite elevation angles, the compensation more accurately reflects the actual propagation distance, ultimately generating the first type of distance constraint.

[0068] For cellular base station signals, the module first decodes the orthogonal frequency division multiplexing (OFDM) symbols in the downlink pilot signal. These symbols have specific frequency distribution characteristics, making it easy to extract phase information. By analyzing the phase difference between adjacent subcarriers, the propagation delay of the signal from the base station to the receiver can be deduced. Then, combined with the known geographical location information of the cellular base station, the time-dimensional delay is converted into a spatial-dimensional distance, thereby generating a second type of distance constraint.

[0069] For wireless access point signals, the module first captures the preamble waveform of the broadcast signal. The preamble has fixed structural characteristics and is a key identifier for distinguishing signals from different paths. Signal processing techniques are used to separate the phase rotation components of the direct and reflected paths. Due to the different propagation environments of the direct and reflected paths, their phase rotation characteristics differ, allowing for effective differentiation. Subsequently, a path suppression factor is applied, which specifically reduces the contribution of the reflected path to the overall signal, minimizing multipath interference. Finally, the third type of distance constraint is reconstructed.

[0070] In the constraint fusion field construction module, the process of generating a three-dimensional spatial probability density field integrates multi-source constraint information in a hierarchical manner. First, a three-dimensional cubic mesh coordinate system is established based on the position prediction points output by the inertial navigation device. This coordinate system provides a spatial reference framework for the subsequent construction of the constraint domain.

[0071] The first type of distance constraint is mapped to the direction of the satellite line-of-sight vector to form a cone-shaped constraint domain. The satellite line-of-sight vector is the direction from the receiver to the satellite. The opening angle of the cone is dynamically adjusted by the confidence level of the satellite elevation angle. The higher the satellite elevation angle, the less interference the signal propagation is affected by the environment. The higher the confidence level, the smaller the opening angle of the cone and the stricter the constraint.

[0072] The second type of distance constraint constructs a ring-shaped constraint band along the azimuth angle of the cellular base station. The azimuth angle determines the horizontal direction of the base station relative to the receiver. The width of the ring band is determined by the accuracy of the base station signal time measurement. The higher the measurement accuracy, the narrower the ring band width and the more precise the constraint.

[0073] The third type of distance constraint generates a spherical constraint body with the physical location of the wireless access point as the center. The radius of the sphere is determined by the distance value after path suppression factor calibration. The calibrated distance value is closer to the actual distance of the direct path, ensuring that the sphere range can accurately surround the possible positioning area.

[0074] Finally, the three sets of constraint structures are spatially weighted and fused based on the signal feature fusion weight set. The weight of the conical constraint domain is positively correlated with the satellite phase dominance factor; when the satellite signal reliability is high, its corresponding constraint domain weight increases. The weight of the annular constraint band is adjusted by the base station delay compensation factor; when the base station signal is more suitable for the current environment, the weight of this constraint band increases. The scaling factor of the spherical constraint is controlled by the wireless signal path suppression factor; when reflection path interference is effectively suppressed, the size of the sphere more closely matches the actual situation. Through this weighted fusion, a probability density field isosurface distribution model reflecting the positioning probability of each spatial location is ultimately formed.

[0075] In the elevation coordinate calculation module, the process of iteratively converging to the positioning coordinate solution along the direction of field intensity change achieves precise positioning through multi-layer constraints and dynamic adjustments. First, the module reads real-time altitude data collected by a barometric pressure sensor. Using this data, it performs vertical dimensional compression on the solution space of the three-dimensional probability density field, focusing the original positioning range covering three-dimensional space onto a horizontal slice layer corresponding to the altitude. This operation significantly reduces the spatial range of the positioning search, decreases unnecessary computation, and ensures that the solution process always revolves around the actual elevation.

[0076] Within the compressed, constrained solution space, the module initiates a field-driven position search process. The initial search point is set to the position prediction value output by the inertial navigation device. This initial value integrates historical information about the motion trajectory, providing a realistic starting point for the search. Subsequently, the module calculates the comprehensive matching error between this initial point and three types of distance constraints, including the projection residual of the conical constraint domain, the distance deviation of the annular constraint zone, and the radius error of the spherical constraint body. These error components are combined according to preset weights into a comprehensive index to quantify the degree of deviation between the current point and the true position.

[0077] During the search, the module moves the search point along the negative gradient direction of the integrated matching error. This negative gradient direction, where the error decreases most rapidly, guides the search point towards a better position. The step size is not fixed but dynamically adjusted based on historical positioning stability: a larger step size is used to accelerate convergence when past positioning results show little fluctuation; when historical data shows significant positioning jitter, the step size is automatically reduced to improve search accuracy. After each move, the module synchronously updates the elevation constraints and recalculates the vertical deviation compensation between the current position and the horizontal slice layer, ensuring the search always adheres to the horizontal layer defined by the air pressure elevation. The iteration terminates based on the angle between the trajectory vectors formed by three consecutive moves. When this angle is less than a preset directional change threshold, it indicates that the search point has entered a stable local region, forming a closed loop. At this point, the iteration terminates, and the coordinates of the minimum integrated matching error point within the loop are output as the final positioning result.

[0078] The elevation coordinate calculation module also integrates a spatial reflection characteristic feedback optimization mechanism after successful positioning, continuously learning to improve the system's adaptability to the environment. Each time positioning is successful, the module automatically records the multipath reflection intensity distribution map of the current environment and the corresponding weight set parameters. The map includes a histogram of the proportion of satellite signal multipath components, the cellular base station angle of arrival spread curve, and the distribution of wireless access point strength abrupt events. These data comprehensively characterize the signal propagation characteristics in the current environment.

[0079] Based on these records, the module constructs an environmental feature index library, associating and storing the feature vectors of the multipath reflection intensity distribution map with the corresponding weight set parameters. When a multipath reflection intensity distribution map is collected in a new positioning scene, the system compares its feature vectors with historical data in the index library for similarity. If the similarity exceeds the matching threshold, it indicates that the current environment is highly similar to a certain type of historical scene. At this point, the complex weight calculation process can be skipped, and the associated historical weight set can be directly loaded, significantly improving the positioning response speed.

[0080] To further optimize environmental adaptability, the module periodically performs spatial overlay analysis on the accumulated positioning trajectory point cloud and building information model. By comparing the spatial relationship between the trajectory and the building structure, it identifies signal reflection anomalies when the trajectory passes through areas of building structures such as walls and floors. Based on the distribution of building materials around these anomalies, the module corrects the parameters in the building material reflection feature database, focusing on optimizing the attenuation slope parameter of glass and the multipath scattering distribution template of concrete, so that the database can more accurately reflect the signal propagation patterns in the actual environment.

[0081] Regarding the rules for calling the historical weight set, when the environment recognition module determines that the current scene matches the historical pattern, the system immediately activates the fast positioning channel and loads the corresponding historical weight set parameters. At this time, the grid size control strategy in the probability density field construction stage is adjusted accordingly. The side length of the three-dimensional cube grid is set as a monotonically increasing function of the historical positioning accuracy. Sparse grids are used in areas with small historical positioning errors to reduce computational load, while denser grids are used in areas with large errors to ensure positioning accuracy.

[0082] During the location search phase, the module uses the heading angle output by the inertial navigation system as a guidance vector, constructing a fan-shaped search interval with the current position as the origin. The angle of the fan is adaptively adjusted according to the rate of change of the heading angle: when the moving body's heading is stable, the angle decreases, focusing the search area near the direction of travel; when the heading changes drastically, the angle increases to avoid missing possible positioning points. By restricting the movement direction of the search point to within the fan-shaped interval, invalid search paths are effectively reduced, computational complexity is significantly lowered, while ensuring the real-time performance and accuracy of positioning.

[0083] This invention also includes an automatic activation of a joint positioning mode using cellular base station signals and wireless access point signals when satellite signals fail, to maintain positioning continuity in complex environments. In this case, the primary task is to align the clock references of the two types of signals. This is achieved by extracting the frame timing deviation from the cellular base station to compensate for clock errors in the receiver caused by crystal oscillator drift, ensuring that the cellular base station propagation delay measurement and the wireless access point reconstructed distance constraint are on the same time reference, thus avoiding positioning errors caused by time asynchrony.

[0084] Based on this, the system constructs motion trajectory smoothing constraint equations using displacement data output from the inertial navigation device. These equations incorporate velocity continuity parameters to ensure that the velocity changes of the positioning point conform to physical laws, avoiding abrupt changes. Simultaneously, acceleration boundary conditions are set to limit the acceleration range of the moving body within reasonable physical thresholds, preventing positioning results from exceeding actual motion capabilities. These constraint equations are further transformed into additional boundary conditions for a probability density field. By compressing the solution space, the position search path is strictly restricted to the kinematically feasible region, effectively filtering out unreasonable positioning solutions.

[0085] When satellite signal recovery is detected, the system does not immediately switch to multi-source fusion mode. Instead, it employs a gradual weight adjustment strategy: progressively increasing the weight of the satellite phase dominance factor while correspondingly decreasing the weights of the base station delay compensation factor and the wireless signal path suppression factor. This allows the positioning process to smoothly transition from dual-source joint mode to multi-source fusion mode. This gradual switching mechanism avoids the impact of signal abrupt changes on positioning stability, ensuring the continuity and reliability of positioning results throughout the transition process.

[0086] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A high-precision positioning system for complex environments based on multi-signal source collaboration, characterized in that, include: The multi-source signal extraction module is used to receive carrier signals transmitted by the satellite navigation system, downlink pilot signals of cellular base stations, and broadcast signals of wireless access points, and extract carrier phase measurement values ​​of satellite signals, propagation delay measurement values ​​of cellular base station signals, and multipath reflection intensity distribution of wireless access point signals from the physical layer raw data; The reflection weight generation module is used to identify the spatial reflection characteristics of the current environment based on the multipath reflection intensity distribution, and dynamically generate a signal feature fusion weight set that matches the spatial reflection characteristics. The weight set includes a satellite phase dominance factor, a base station delay compensation factor, and a wireless signal path suppression factor. The quantity constraint conversion module is used to convert satellite carrier phase measurement values ​​into first-type distance constraint values, cellular base station propagation delay measurement values ​​into second-type distance constraint values, and wireless access point signals into third-type distance constraint values ​​through path suppression factors. The constraint fusion field construction module is used to nonlinearly superimpose three types of distance constraint quantities based on the signal feature fusion weight set to generate a three-dimensional spatial probability density field. The elevation coordinate solution module integrates barometric elevation data to compress the probability density field solution space and iteratively converges to the positioning coordinate solution along the direction of field strength variation. In the reflection weight generation module, the process of identifying the spatial reflection characteristics of the current environment based on the multipath reflection intensity distribution is as follows: Calculate the proportion of multipath components of satellite signals within a unit time window, and determine a high reflection environment when the proportion exceeds a preset threshold; Analyze the histogram of the angle of arrival distribution of the downlink pilot signal of the cellular base station. When the kurtosis of the histogram is lower than the critical value and there are continuous strong reflection pulses, it is determined to be an open indoor environment. By monitoring the intensity abrupt change event sequence of the wireless access point broadcast signal and combining it with the attitude angle change rate output by the inertial navigation device, the transition zone where the moving body crosses the boundary of the building structure can be identified. The signal feature fusion weight set enhances the role of the base station delay compensation factor in high reflection environments and activates the wireless signal path suppression factor calculation module in open indoor environments. The generation process of the path inhibition factor is as follows: A pre-defined database of building material reflection characteristics stores phase reversal shift modes caused by metal surfaces, signal attenuation slope parameters caused by glass media, and multipath scattering distribution templates generated by concrete walls. The phase jump characteristics of the received signal are compared with the database recording pattern in real time. When a phase reverse offset matching the metal reflection pattern is detected, a negative phase compensation operator is generated and injected into the third type of distance constraint quantity reconstruction process. When the signal strength attenuation slope corresponding to the glass attenuation characteristics is identified, the amplitude scaling ratio of the distance constraint is dynamically adjusted according to the number of electromagnetic wave penetration layers and the attenuation slope parameter. For template matching scenarios involving multipath scattering in concrete, path separation technology is used to extract the phase value of the main reflection component to replace the original measurement value.

2. The high-precision positioning system for complex environments based on multi-signal source collaboration according to claim 1, characterized in that, In the quantity constraint conversion module, the specific conversion process of the distance constraint quantity is as follows: A dual-frequency ionospheric delay correction is applied to the satellite carrier phase measurement, and a tropospheric refraction compensation is calculated based on the satellite elevation angle to generate a first-type range constraint. Decode the orthogonal frequency division multiplexing symbols in the downlink pilot signal of the cellular base station, infer the signal propagation delay by the phase difference between adjacent subcarriers, and generate the second type of distance constraint by combining the base station's geographical location information; The preamble waveform of the wireless access point broadcast signal is captured, the phase rotation components of the direct path and the reflected path are separated, the path suppression factor is applied to reduce the contribution of the reflected path, and the third type of distance constraint is reconstructed.

3. The high-precision positioning system for complex environments based on multi-signal source collaboration according to claim 1, characterized in that, In the constrained fusion field construction module, the process of generating the three-dimensional spatial probability density field is as follows: A three-dimensional cubic grid coordinate system is established based on the position prediction points output by the inertial navigation device; the first type of distance constraint is mapped to the direction of the satellite line-of-sight vector to form a cone-shaped constraint domain, and the cone opening angle is dynamically adjusted by the satellite elevation confidence. The second type of distance constraint constructs an annular constraint band along the azimuth of the cellular base station, and the width of the annular band is determined by the accuracy of the base station signal time measurement. The third type of distance constraint quantity generates a spherical constraint body with the physical location of the wireless access point as the center. The radius of the sphere is determined by the distance value after path suppression factor calibration. Based on the signal feature fusion weight set, the three sets of constraint structures are spatially weighted and fused. The weight of the cone constraint domain is positively correlated with the satellite phase dominance factor, the weight of the ring constraint band is adjusted by the base station delay compensation factor, and the scaling factor of the spherical constraint body is controlled by the wireless signal path suppression factor, thus forming a probability density field isosurface distribution model.

4. The high-precision positioning system for complex environments based on multi-signal source collaboration according to claim 1, characterized in that, In the elevation coordinate calculation module, the specific process of iteratively converging to the positioning coordinate solution along the direction of field intensity change is as follows: Read the altitude data collected by the barometric pressure sensor and compress the vertical dimension of the probability density field solution space to the horizontal slice layer corresponding to the altitude; Perform field strength driven position search within the constrained solution space, initialize the search point as the inertial navigation predicted position, calculate the comprehensive matching error between the point and the three types of distance constraints, the comprehensive matching error is a weighted synthesis of the projection residual of the conical constraint domain, the distance deviation of the annular constraint zone and the radius error of the spherical constraint body; The search point is moved along the negative gradient direction of the matching error. The step size is dynamically adjusted according to the historical positioning stability. After each move, the elevation constraint is updated synchronously, and the vertical deviation compensation between the current position and the horizontal slice layer is recalculated. When the angle between the trajectory vectors formed by three consecutive moves is less than the direction change threshold, it is determined that a closed loop has been formed. The iteration is terminated and the coordinates of the minimum matching error point within the loop are output.

5. The high-precision positioning system for complex environments based on multi-signal source collaboration according to claim 1, characterized in that, The elevation coordinate calculation module also includes feedback optimization of spatial reflection characteristics after each successful positioning, specifically: Each time a positioning is successful, a multipath reflection intensity distribution map and corresponding weight set parameters are recorded. The map includes a histogram of the proportion of satellite signal multipath components, a cellular base station angle of arrival spread curve, and a distribution of wireless access point intensity abrupt events. An environmental feature index library is established, and the feature vectors of the multipath reflection intensity distribution map are associated and stored with the weight set parameters. When the similarity between the newly acquired multipath reflection intensity distribution map and the historical feature vector exceeds the matching threshold, the weight calculation process is skipped and the associated historical weight set is directly loaded. Regularly perform spatial overlay analysis on the positioning trajectory point cloud and building information model to identify signal reflection anomalies in areas where the trajectory crosses walls; correct the parameters of the building material reflection feature database based on the material distribution around the anomalies, focusing on optimizing the glass medium attenuation slope parameter and the concrete multipath scattering distribution template.

6. The high-precision positioning system for complex environments based on multi-signal source collaboration according to claim 5, characterized in that, The rules for accessing the historical weight set are as follows: When the environment recognition module triggers historical pattern matching, the fast positioning channel is activated and historical weight set parameters are loaded; the grid size control strategy in the probability density field construction stage is adjusted, the side length of the three-dimensional cube grid is set as a monotonically increasing function of historical positioning accuracy, and sparse grids are used in areas with small historical positioning errors; In the position search phase, the heading angle output by the inertial navigation device is introduced as the guidance vector, and a sector search interval is constructed with the current position as the origin. The sector angle is adaptively adjusted according to the rate of change of the heading angle, and the movement direction of the search point is restricted within the sector interval to reduce computational complexity.

7. The high-precision positioning system for complex environments based on multi-signal source collaboration according to claim 1, characterized in that, When satellite signals fail, the joint positioning mode of cellular base station signals and wireless access point signals is activated. The propagation delay measurement of the cellular base station and the reconstructed distance constraint of the wireless access point are aligned with the clock reference. The alignment process uses the frame timing deviation of the cellular base station to compensate for the receiver clock drift. Based on displacement data from an inertial navigation device, a smoothing constraint equation for the motion trajectory is constructed. The equation parameters include velocity continuity and acceleration boundary conditions. The smoothing constraint equation is then transformed into a probability density field with additional boundary conditions to restrict the position search path to within the kinematically feasible solution space. When satellite signal recovery is detected, the weight of the satellite phase dominance factor is gradually increased until the system is fully switched to multi-source fusion mode.

Images