A big data-based pedestrian floor tile navigation data analysis method and system

Abstract

This application relates to a method and system for pedestrian pavement navigation data analysis based on big data. The method includes: acquiring pressure signals from a pedestrian pavement area; adjusting the processing parameters of the pressure signals according to the local environmental conditions of the area to extract pressure events; classifying pressure events into footstep signals, guide cane touch signals, and wheel signals based on pressure characteristics; correlating the footstep signals and guide cane touch signals in time and space to form a pedestrian movement trajectory; and determining whether the wheel signals have a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent. This invention achieves an accurate and complete depiction of the travel trajectory of visually impaired individuals, providing them with a more complete and accurate travel navigation service.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of digital city construction, in particular to a pedestrian tile navigation data analysis method and system based on big data. BACKGROUND

[0002] In order to improve the travel convenience and safety of visually impaired people, the intelligent pedestrian tile system integrated with pressure sensors is deployed in the barrier-free design of digital city construction. The intelligent pedestrian tile system generally captures pressure signals through sensors, transmits them to the regional data processing center through underground communication lines, and analyzes them to depict the travel route of visually impaired people, identify potential obstacles, and evaluate the status of tile facilities, thereby providing data support for urban planning and barrier-free facility maintenance. The system can theoretically significantly improve the travel experience of visually impaired groups.

[0003] However, after the system is put into operation, some problems are gradually exposed. First, shared bicycles, electric scooters, and vendor trolleys, etc. continuously produce small pressure changes on the tiles due to parking movement, wind blowing, or road vibration, which are misjudged as pedestrian signals by the sensors, resulting in high-frequency signals in unpopulated alleys at night, which pollute the data analysis results. Second, although the signal discrimination program is developed by engineers based on the waveform difference between "pedestrian footsteps / cane touching the ground as short-time strong pulse and wheel interference as continuous low-amplitude vibration" during design, in actual application, it mistakenly filters the wheel movement signals of the luggage of visually impaired people (similar to wheel interference characteristics), and rainwater (water film buffer reduces the peak value of the footstep signal and increases the pulse width) and sandy soil (causing irregular signal jumps) further exacerbate signal distortion, making it difficult to distinguish between the four main signals: pedestrian footsteps, cane touching the ground, user luggage wheel movement, and wheel interference.

[0004] In related technologies, the intelligent pedestrian tile system cannot avoid the pollution of data by wheel interference, and is prone to losing effective user data, seriously damaging the accuracy and reliability of navigation data analysis, making it difficult to achieve the goal of improving the travel friendliness of visually impaired people, and improvement is urgently needed. SUMMARY

[0005] The present application provides a pedestrian tile navigation data analysis method and system based on big data to at least solve the problem that the intelligent pedestrian tile system is difficult to distinguish between the interference signals generated by various wheeled objects and the effective signals generated by visually impaired people themselves, especially when environmental changes cause signal distortion, the existing identification method lacks accuracy, seriously affecting the reliability of navigation data analysis, and the system cannot obtain complete travel trajectories.

[0006] In a first aspect, the present application provides a pedestrian tile navigation data analysis method based on big data, which comprises:

[0007] Acquiring a pressure signal of a pedestrian floor tile area, and adjusting a processing parameter of the pressure signal according to a local environment state of the pedestrian floor tile area to extract a pressure event from the pressure signal, wherein the pressure event contains time, position and pressure characteristics;

[0008] Preliminarily classifying the pressure event into a footstep signal, a walking stick touch ground signal and a wheel signal based on the pressure characteristics;

[0009] Correlating the footstep signal and the walking stick touch ground signal in time and space to form a pedestrian moving track, and judging whether the wheel signal has a spatio-temporal correlation with the pedestrian moving track and a moving mode consistency;

[0010] If the wheel signal has a spatio-temporal correlation with the preliminary pedestrian moving track or a moving mode consistency, the wheel signal is confirmed as a user luggage wheel signal and fused into the preliminary pedestrian moving track to construct a human behavior chain;

[0011] If the wheel signal has no spatio-temporal correlation with the preliminary pedestrian moving track or a moving mode consistency, the wheel signal is excluded;

[0012] Based on context information of the human behavior chain, missing track segments in the human behavior chain are inferred and completed to obtain a complete travel track.

[0013] Optionally, the correlating the footstep signal and the walking stick touch ground signal in time and space to form a pedestrian moving track, and judging whether the wheel signal has a spatio-temporal correlation with the pedestrian moving track and a moving mode consistency, comprises:

[0014] Identifying a user adaptive behavior mode, and evaluating a momentum of the pedestrian moving track;

[0015] Combining the local environment state and the user adaptive behavior mode, dynamically adjusting a time tolerance window and a space tolerance window of the wheel signal in association with the pedestrian moving track;

[0016] According to the momentum of the pedestrian moving track, predicting a path of the pedestrian moving track, and judging whether the wheel signal has a spatio-temporal correlation with the pedestrian moving track within the time tolerance window, whether a moving mode of the wheel signal has a consistency with a moving mode of the preliminary pedestrian moving track, and whether the wheel signal appears a local deviation on the predicted path;

[0017] If all the judged conditions are met, the wheel signal is confirmed as a user luggage wheel signal and fused into the preliminary pedestrian moving track to form the human behavior chain;

[0018] If there is no spatio-temporal correlation or inconsistent movement pattern between the wheeled signal and the pedestrian movement trajectory, the duration, micro-movement pattern and periodic characteristics of the wheeled signal are identified to exclude the wheeled signal.

[0019] Optionally, based on the context information of the human behavior chain, missing trajectory segments in the human behavior chain are inferred and completed to obtain a complete travel trajectory, including:

[0020] The continuity of the human behavior chain is monitored, and if the human behavior chain is broken, an inference mechanism is started;

[0021] It is checked whether there is a wheeled signal that has been classified as a non-human behavior pattern in the broken area, and if so, the wheeled signal is excluded;

[0022] According to the movement trend, speed and direction before and after the break, combined with the adaptive gait state of the user before and after the break and the local environmental state of the broken area, the path of the user during the break is predicted, including the sequence of tiles that may be passed through and the expected time;

[0023] The behavior semantic consistency of the predicted path is evaluated;

[0024] If the behavior semantic consistency of the predicted path is high and consistent with the actual break time, a missing footstep signal is generated in the broken area and inserted into the human behavior chain to complete the missing trajectory segments in the human behavior chain to obtain a complete travel trajectory.

[0025] Optionally, the evaluation of the behavior semantic consistency of the predicted path includes:

[0026] The micro-behavior patterns on the predicted path are monitored and marked as non-navigation related behavior deviations, wherein the micro-behavior patterns include a stress event sequence with a duration shorter than a first preset threshold or a footstep signal with a slight and short deviation from the average step frequency and step length of the user;

[0027] The predicted path is compared with the non-navigation related behavior deviations to determine whether there is a local deviation between the predicted path and the actual travel trajectory of the user, and whether the local deviation is consistent with the non-navigation related behavior deviations in time and space;

[0028] If the local deviation is consistent with the non-navigation related behavior deviations in time and space, it is determined that the local deviation is caused by user behavior, the semantic consistency evaluation result of the predicted path is maintained, and the predicted path is adjusted according to the non-navigation related behavior deviations to complete the missing trajectory segments in the human behavior chain.

[0029] Optionally, the monitoring of the continuity of the human behavior chain, if the human behavior chain appears to be broken, starts an inference mechanism, comprising:

[0030] Monitoring the continuity of the human behavior chain, and situational assessment of the continuity change of the human behavior chain;

[0031] Judging whether the continuity change is caused by signal loss with a duration shorter than a second preset threshold or step signal with slight and temporary deviation from the average step frequency and step length of the user;

[0032] Judging whether the continuity change is consistent with the stability of the local environment;

[0033] If the continuity change is not caused by signal loss with a duration shorter than a preset threshold or step signal with slight and temporary deviation from the average step frequency and step length of the user, and the continuity change is not consistent with the stability of the local environment, it is judged that the human behavior chain appears to be broken, and an inference mechanism is started.

[0034] Optionally, the situational assessment of the continuity change of the human behavior chain comprises:

[0035] Continuously acquiring real-time environmental parameters of the floor tile area, and performing multi-time scale analysis on the real-time environmental parameters to identify instantaneous fluctuations, short-term trends and long-term baselines of the real-time environmental parameters;

[0036] When the real-time environmental parameters appear to have a large instantaneous fluctuation, a local environment state rapid sampling mode is started, the sampling frequency is increased, and a multi-sensor cross-validation is started to obtain a fine-grained local environment state;

[0037] According to the short-term trend and the long-term baseline of the real-time environmental parameters, the sensitivity threshold of the continuity evaluation of the human behavior chain is dynamically adjusted;

[0038] When the instantaneous fluctuation of the real-time environmental parameters is consistent with the continuity change of the human behavior chain in time and space, the influence of the instantaneous fluctuation on the continuity change is given priority, and the situational assessment result is corrected.

[0039] Optionally, the situational assessment of the continuity change of the human behavior chain further comprises:

[0040] Continuously acquiring real-time environmental parameters of the floor tile area, and monitoring abnormal states of the real-time environmental parameters, wherein the abnormal states include data loss, data anomaly or data delay;

[0041] When the abnormal state occurs, a multi-source cross-validation mechanism of environmental data is started, and the multi-source cross-validation mechanism comprises:

[0042] acquiring environment parameters from adjacent tile areas and comparing with the abnormal environment parameters;

[0043] acquiring average environment parameters of the same time period and the same area from historical environment data and comparing with the abnormal environment parameters;

[0044] correcting or completing the abnormal environment parameters according to the comparison results, and comprehensively scoring the continuity of the human behavior chain according to the real-time environment parameters after correction or completion, combined with the dynamic characteristics such as moving speed, direction, step frequency and step length of the human behavior chain.

[0045] Optionally, the acquiring environment parameters from adjacent tile areas and comparing with the abnormal environment parameters comprises:

[0046] identifying the fluctuation degree of each adjacent tile area environment parameter, and calculating the reliability weight of each adjacent tile area environment parameter according to the fluctuation degree of each adjacent tile area environment parameter;

[0047] selecting adjacent tile area environment parameters with reliability weight higher than a preset threshold, and weightedly averaging the selected adjacent tile area environment parameters to obtain fused environment parameters;

[0048] comparing the fused environment parameters with the abnormal environment parameters.

[0049] Optionally, the correcting or completing the abnormal environment parameters according to the comparison results comprises:

[0050] identifying the deviation dimension and the deviation degree of the abnormal environment parameters according to the comparison results;

[0051] identifying the dominant abnormal factor according to the deviation dimension and the deviation degree, combined with the physical characteristics of the tile area and the historical environment event records;

[0052] selecting a corresponding correction strategy from a preset correction strategy library according to the type of the dominant abnormal factor, and correcting or completing the abnormal environment parameters in a hierarchical manner based on the correction strategy, wherein the hierarchical correction or completion comprises:

[0053] correcting the environment parameter dimension most affected by the dominant abnormal factor in priority;

[0054] auxiliary correction of the environment parameter dimension affected by the secondary abnormal factor;

[0055] consistency check of the corrected environment parameters to avoid introducing new errors in the correction process.

[0056] Secondly, this application provides a pedestrian paving stone navigation data analysis system based on big data, the system comprising:

[0057] The signal acquisition module acquires the pressure signal of the pedestrian paving area and adjusts the processing parameters of the pressure signal according to the local environmental state of the pedestrian paving area to extract pressure events from the pressure signal, wherein the pressure events include time, location and pressure characteristics;

[0058] The preliminary classification module, based on the pressure characteristics, preliminarily classifies the pressure events into footstep signals, guide cane ground contact signals, and wheel signals;

[0059] The human-centered behavior chain construction module correlates the footstep signals and the guide cane touch-the-ground signals in time and space to form a pedestrian movement trajectory, and determines whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent:

[0060] The trajectory forming unit is used to correlate the preliminarily classified footstep signals and the preliminarily classified guide cane ground-touching signals in time and space to form a preliminary pedestrian movement trajectory.

[0061] The wheel signal determination unit, if the wheel signal has a spatiotemporal correlation or the same movement pattern as the preliminary pedestrian movement trajectory, then confirms the wheel signal as the user's suitcase wheel signal and integrates it into the pedestrian movement trajectory to construct a human-centered behavior chain;

[0062] The wheel signal exclusion unit excludes the wheel signal if the wheel signal does not have a spatiotemporal correlation with the preliminary pedestrian movement trajectory or the movement pattern is inconsistent with it.

[0063] The trajectory completion module, based on the contextual information of the human behavior chain, infers and completes the missing trajectory segments in the human behavior chain to obtain a complete travel trajectory.

[0064] Compared with related technologies, the pedestrian paving data analysis method and system based on big data provided in this application have at least the following technical advantages:

[0065] By receiving pressure signals and adjusting signal processing parameters according to environmental conditions, pressure events containing time, location, and pressure characteristics are accurately extracted from the pressure signals. Based on these characteristics, pressure events are initially classified into footstep signals, guide cane touch-the-ground signals, or wheel signals, effectively distinguishing signals from different sources. Subsequently, by constructing a human-centered behavior chain, the initially classified footstep signals and guide cane touch-the-ground signals are spatiotemporally correlated to form a preliminary pedestrian movement trajectory. Further analysis is then conducted to determine whether wheel signals exhibit spatiotemporal correlation with this trajectory and consistency in movement pattern.

[0066] If present, the wheel signal is identified as the user's suitcase wheel signal and fused into the trajectory, thus solving the problem in the existing technology of not being able to accurately distinguish between the user's suitcase and interference objects, and avoiding the loss of real user data;

[0067] If the signal is not present, wheeled signals are excluded, effectively filtering out interfering signals such as those from shared bicycles. Finally, based on the contextual information of the human behavior chain, this application infers and completes the missing trajectory segments to obtain a complete travel trajectory, overcoming the shortcomings of existing technologies that cannot obtain a complete trajectory due to signal distortion caused by environmental changes.

[0068] In summary, this application constructs an intelligent processing flow from raw pressure signals to complete travel trajectories. Through a series of steps, including adaptive signal processing, multi-dimensional signal classification, intelligent wheel signal recognition, and context-aware trajectory completion, a complete and efficient navigation data analysis method is formed. The various technical features work together to solve problems such as signal distortion, difficulty in distinguishing wheel interference, and missing trajectory data in existing technologies. Ultimately, it achieves an accurate and complete depiction of the travel trajectories of visually impaired individuals, providing them with more complete and accurate travel navigation services.

[0069] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0070] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0071] Figure 1 This is a flowchart illustrating a big data-based pedestrian paving navigation data analysis method according to an exemplary embodiment.

[0072] Figure 2 This is the process of step S3 as shown in another exemplary embodiment.

[0073] Figure 3 This is a flowchart illustrating step S4 according to an exemplary embodiment.

[0074] Figure 4 This is a flowchart illustrating step S44 according to an exemplary embodiment.

[0075] Figure 5 This is a flowchart illustrating step S41 according to an exemplary embodiment.

[0076] Figure 6This is a block diagram illustrating a big data-based pedestrian paving navigation data analysis system according to an exemplary embodiment. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0078] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any creative effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0079] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0080] In related technologies, after the intelligent pedestrian paving system was put into operation, some problems gradually emerged: First, wheeled objects such as shared bicycles, electric scooters, and merchant carts, due to parking, movement, wind, or road vibration, cause continuous and minute pressure changes on the paving stones, which are misinterpreted by sensors as pedestrian signals, resulting in high-frequency signals in deserted alleys late at night, contaminating the data analysis results; Second, although the engineers developed a signal discrimination program based on the waveform difference that "pedestrian footsteps / guide cane contact with the ground is a short-term strong pulse, while wheel interference is a continuous low-amplitude vibration," in actual application, it mistakenly filtered out the wheel movement signal of visually impaired people dragging suitcases (which has similar characteristics to wheel interference). Moreover, rainwater (water film buffering reduces the peak value of footstep signals and increases the pulse width) and sand (causing irregular signal jumps) further aggravate signal distortion, making it difficult to distinguish the four main types of signals: pedestrian footsteps, guide cane contact with the ground, user suitcase wheel movement, and wheel interference. This seriously undermines the accuracy and reliability of navigation data analysis, making it difficult to achieve the system's goal of improving the travel friendliness of visually impaired people.

[0081] Based on the above, embodiments of the present invention provide a method and system for pedestrian paving navigation data analysis based on big data, which will be described in detail below with reference to specific embodiments and accompanying drawings.

[0082] Example 1

[0083] This invention provides a method for analyzing pedestrian paving data for navigation based on big data. Figure 1 This is a flowchart illustrating a pedestrian paving data analysis method based on big data, according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0084] S1. Acquire pressure signals from the pedestrian paving area and adjust the processing parameters of the pressure signals according to the local environmental conditions of the pedestrian paving area to extract pressure events from the pressure signals, wherein the pressure events include time, location and pressure characteristics.

[0085] In this embodiment, the pressure signal refers to the raw electrical signal collected by the pressure sensor inside the pavement tile. Its intensity and waveform reflect the pressure changes on the tile surface. Pressure events are discrete events extracted from the raw pressure signal. Each event includes the time of occurrence, its precise location within the tile system, and pressure characteristics (such as peak value, duration, and waveform). To cope with complex and changing environmental conditions, such as rain or debris on the road surface, the system dynamically adjusts the signal processing parameters according to real-time environmental conditions. For example, when rainwater is detected, the signal threshold can be lowered to compensate for the attenuation effect of the water film on the pressure signal, or the filter parameters can be adjusted to reduce noise generated by raindrop impact. Through this adaptive parameter adjustment, pressure events can be extracted more accurately from the raw pressure signal. Each pressure event includes the time of occurrence, its location within the tile system, and its unique pressure characteristics.

[0086] S2. Based on the pressure characteristics, the pressure events are initially classified into footstep signals, guide cane ground contact signals, and wheel signals;

[0087] In this embodiment, footstep signals specifically refer to pressure events generated by pedestrians stepping on paving stones, typically characterized by specific impact force and duration. Guide cane touch signals refer to pressure events generated when visually impaired individuals use their guide canes to touch paving stones; their characteristics may differ from footstep signals, but they are also human-related behaviors. Wheel signals broadly refer to pressure events generated when wheeled objects (such as shared bicycles, trolleys, suitcases, etc.) roll or stop on paving stones, often characterized by strong continuity and low amplitude. Based on the pressure characteristics of the pressure events extracted in step S1, they are initially classified into footstep signals, guide cane touch signals, or wheel signals. For example, they can be distinguished by analyzing features such as the peak value, duration, and the steepness of the rising and falling edges of the pressure events. Footstep signals typically manifest as short, powerful pulses, guide cane touch signals may have sharper peak values, while wheel signals often exhibit continuous vibrations with longer durations and relatively lower amplitudes.

[0088] S3. Correlate the footstep signal and the guide cane touch the ground signal in time and space to form a pedestrian movement trajectory, and determine whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent:

[0089] S3-1. If the wheel signal has a spatiotemporal correlation or the same movement pattern as the preliminary pedestrian movement trajectory, then the wheel signal is confirmed as the user's suitcase wheel signal and fused into the pedestrian movement trajectory to construct a human-centered behavior chain.

[0090] In this embodiment, the human-centered behavior chain is the core of this application. It represents the complete travel trajectory of a visually impaired person, including not only footsteps and cane touch signals, but also possibly signals from the user's suitcase wheels.

[0091] S3-2. If the wheel signal does not have a spatiotemporal correlation with the preliminary pedestrian movement trajectory or the movement pattern is inconsistent, then the wheel signal is excluded.

[0092] In this embodiment, the initially classified footstep signals and initially classified guide cane touch signals are first correlated in time and space to form a preliminary pedestrian movement trajectory. For example, if footstep signals or guide cane touch signals are continuously detected in adjacent paving areas within a short period of time, and these signals show a reasonable direction of movement in space, they can be connected to form a trajectory.

[0093] Secondly, the system determines whether the initially classified wheeled signals have a spatiotemporal correlation and consistency in movement pattern with the initial pedestrian movement trajectory. For example, if a wheeled signal is almost synchronous in time with a segment of the initial pedestrian movement trajectory and closely follows it in space, and the speed and direction of the wheeled signal are highly consistent with the movement pattern of the pedestrian trajectory, then the wheeled signal is likely from a suitcase being dragged by the user. In this case, the system will identify the wheeled signal as a user's suitcase wheel signal and integrate it into the initial pedestrian movement trajectory, thereby forming a more complete human-centered behavior chain.

[0094] Finally, if a wheeled signal does not have a spatiotemporal correlation with the initial pedestrian movement trajectory or its movement pattern is inconsistent, the wheeled signal will be excluded. For example, if a wheeled signal does not match any pedestrian trajectory in time, or its movement pattern (such as excessive speed or irregular direction) is significantly inconsistent with pedestrian behavior, the signal is likely to originate from interfering objects such as shared bicycles. In this case, the system will exclude it to avoid contaminating the human behavior chain.

[0095] S4. Based on the contextual information of the human behavior chain, infer and complete the missing trajectory segments in the human behavior chain to obtain the complete travel trajectory;

[0096] In this embodiment, contextual information refers to auxiliary data such as environmental data, user historical behavior patterns, and geographic information related to the human behavior chain. When the human behavior chain experiences a brief break, the system does not simply abandon this trajectory. Instead, it intelligently infers and completes the missing trajectory segment by utilizing the movement trend, speed, direction, environmental information, historical behavior patterns, and other contextual data before and after the break. This significantly improves the integrity and continuity of the travel trajectory, effectively compensating for trajectory interruptions caused by brief signal loss or environmental interference, and providing more reliable and continuous navigation data for visually impaired individuals. For example, if the human behavior chain experiences a brief break in a certain area, but the trajectory before and after the break shows that the user is still moving, and combined with historical pedestrian flow data and environmental information for that area, the system can intelligently infer the path the user may have taken and generate a virtual trajectory segment to complete it.

[0097] The implementation environment of the above embodiments is typically an urban public area where an intelligent pedestrian paving system has been deployed, such as sidewalks, squares, and stations.

[0098] The technical solution of the above embodiments constructs an intelligent processing flow from raw pressure signals to complete travel trajectories. Through a series of steps such as adaptive signal processing, multi-dimensional signal classification, intelligent wheel signal recognition, and context-aware trajectory completion, a complete and efficient navigation data analysis method is formed. The various technical features cooperate with each other to solve the problems of signal distortion, difficulty in distinguishing wheel interference, and missing trajectory data in the prior art, and finally realize the accurate and complete depiction of the travel trajectory of visually impaired people.

[0099] In one possible design, Figure 2 This is the process of step S3 as illustrated in another exemplary embodiment. (Refer to the appendix.) Figure 2 Step S3 includes:

[0100] S31. Identify user adaptive behavior patterns and assess the momentum of the pedestrian movement trajectory;

[0101] In this embodiment, identifying user adaptive behavior patterns involves analyzing changes in the user's gait characteristics in specific environments. For example, when a user walks on a flat surface versus a slope, a slippery surface, or a congested area, their stride frequency, stride length, and ground pressure will undergo adaptive adjustments. Evaluating the momentum of the preliminary pedestrian movement trajectory refers to calculating or estimating the trends in the speed and direction of the pedestrian's movement trajectory, which helps in understanding the pedestrian's motion inertia.

[0102] S32. Combining the local environmental state and the user's adaptive behavior pattern, dynamically adjust the time tolerance window and spatial tolerance window associated with the wheel signal and the pedestrian movement trajectory;

[0103] In this embodiment, adjusting the time tolerance window and the spatial tolerance window makes the association judgment more flexible and accurate. For example, when the user's gait is stable and the environment is stable, the tolerance window can be set to be smaller; while when the user's gait undergoes adaptive changes or the environment is complex, the tolerance window can be appropriately enlarged to adapt to the actual situation.

[0104] S33. Based on the momentum of the pedestrian's movement trajectory, predict the path of the pedestrian's movement trajectory and determine:

[0105] Whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory within the time tolerance window;

[0106] Whether the movement pattern of the wheeled signal is consistent with the movement pattern of the preliminary pedestrian movement trajectory;

[0107] Does the wheel signal exhibit local deviation on the predicted path?

[0108] In this embodiment, the path of the preliminary pedestrian movement trajectory is predicted based on the trajectory momentum, providing an expected range of motion for the association of wheeled signals. By analyzing the current speed and direction of the pedestrian trajectory, its possible travel path in a short period of time can be inferred, thereby providing a reference for the location of the wheeled signal.

[0109] Among them, spatiotemporal correlation ensures that wheel signals and pedestrian trajectories are synchronized in time and space; movement pattern consistency compares the coordination between wheel signals (such as periodicity and pressure distribution) and pedestrian gait; and the judgment of local deviation takes into account that when users are carrying suitcases, the wheels may deviate slightly from the pedestrian's footstep trajectory due to uneven ground or turning, rather than completely overlapping.

[0110] S34. If all the conditions are met, the wheel signal is confirmed as the user's suitcase wheel signal and integrated into the preliminary pedestrian movement trajectory to form the human behavior chain.

[0111] S35. If the wheel signal has no spatiotemporal correlation with the pedestrian movement trajectory or the movement pattern is inconsistent, the duration, minute movement pattern, and periodicity of the wheel signal are identified to exclude the wheel signal. For example, the wheel signal of a non-user's suitcase may exhibit a duration completely unrelated to the pedestrian trajectory (such as a shopping cart that has been stationary for a long time), a unique minute movement pattern (such as the swaying of a stroller), or irregular periodicity. These features can effectively distinguish it from the wheel signal of the user's suitcase and exclude it.

[0112] The technical solution of the above embodiments dynamically and multidimensionally determines the association between wheel signals and pedestrian movement trajectories by introducing perception of local environmental states, recognition of user adaptive behavior patterns, and evaluation of trajectory momentum. The introduction of dynamically adjusted time and space tolerance windows enables the system to better adapt to the complex and ever-changing real-world environment and user behavior, avoiding misjudgments or omissions caused by fixed thresholds. Simultaneously, trajectory momentum prediction provides a more accurate expected range for the occurrence of wheel signals, further narrowing the search space and improving judgment efficiency. Furthermore, the inclusion of local deviations in the predicted path for wheel signals in the judgment conditions allows the system to tolerate small, reasonable deviations that suitcase wheels may make during actual movement, thus more accurately capturing the user's actual behavior of carrying a suitcase. For wheel signals that do not meet the association conditions, their duration, minute movement patterns, and periodic characteristics can be identified to distinguish them from the user's suitcase wheel signals, thereby avoiding the incorrect integration of irrelevant signals into the human behavior chain and ensuring the accuracy of the human behavior chain.

[0113] In one example, suppose a user is walking with a suitcase in a crowded shopping mall. The mall floor may be slightly uneven, and there are other wheeled objects such as trolleys and strollers passing by, while the ambient noise is relatively high.

[0114] If a basic approach is adopted, relying solely on fixed spatiotemporal correlations and consistency of movement patterns for judgment, it's possible that the spatiotemporal correlation between suitcase wheel signals and footstep signals might momentarily deviate due to pedestrian obstruction, uneven ground, or slight differences in movement patterns from the preset ideal pattern, leading to incorrect exclusion. Simultaneously, signals emitted by other wheeled objects may also exhibit occasional spatiotemporal correlations with the user's trajectory within a short period, resulting in misjudgments.

[0115] The proposed solution first involves the system sensing the local environmental conditions of a shopping mall, such as high pedestrian traffic and strong noise, and adjusting the extraction parameters of the stress signal accordingly to ensure the accuracy of stress events. Second, the system identifies the user's adaptive gait patterns in crowded environments (e.g., slight changes in cadence and shortened stride length) and assesses the momentum of their initial pedestrian movement trajectory. Based on this information, the system dynamically adjusts the temporal and spatial tolerance windows for associating wheel signals with pedestrian trajectories, making them more flexible. For example, when a user turns or avoids pedestrians, the tolerance window may be appropriately widened to allow for a brief, reasonable spatiotemporal deviation between the suitcase wheel signal and the footstep signal. Simultaneously, the system predicts the user's path at the next moment based on trajectory momentum. When judging wheel signals, it not only checks the spatiotemporal correlation and movement pattern consistency but also determines whether the wheel signals exhibit local deviations on the predicted path. For example, a suitcase wheel might make a small, brief deviation when passing over a seam in the ground, but as long as it remains within a reasonable range of the predicted path and the movement pattern is coordinated with the user's gait, it will be recognized as a user's suitcase wheel signal. For other wheeled signals such as those from strollers or baby carriages, even if they happen to have a spatiotemporal correlation with the user's trajectory, the system will identify and exclude them because their duration, minute movement patterns (such as the swaying frequency of a baby carriage), or periodic characteristics are significantly different from the user's suitcase wheel signal.

[0116] In one possible design, Figure 3 This is a flowchart illustrating step S4 according to an exemplary embodiment. (Refer to the attached document.) Figure 3 Step S4 includes:

[0117] S41. Monitor the continuity of the human-centered behavior chain. If the human-centered behavior chain is broken, initiate the inference mechanism.

[0118] In this embodiment, monitoring the continuity of the human-centered behavior chain refers to performing real-time or near-real-time analysis of the continuity of the human-centered behavior chain in time and space. If the human-centered behavior chain is broken, that is, if no continuous effective pressure events are detected within a preset time or space threshold, resulting in trajectory interruption, then the inference mechanism is activated.

[0119] S42. Check whether there are wheel-like signals that have been classified as non-human behavior patterns in the fractured area. If so, exclude the wheel-like signals.

[0120] In this embodiment, wheel-like signals representing non-human behavior patterns can be understood as wheel-like signals from non-user luggage such as shopping carts or cleaning carts. Excluding these signals is to avoid misjudging signals unrelated to the user's primary behavior as part of the user's trajectory, thereby improving the accuracy of trajectory completion.

[0121] S43. Based on the movement trend, speed and direction before and after the break, combined with the user's adaptive gait state before and after the break and the local environmental state of the broken area, predict the user's path during the break. The predicted path includes the possible sequence of paving stones and the expected time.

[0122] In this embodiment, movement trend, speed, and direction are inferred by analyzing trajectory data before and after the break to determine the user's approximate direction of movement and speed variation. Adaptive gait state refers to changes in gait characteristics such as stride frequency and stride length under different environments or situations (e.g., encountering obstacles, changing direction, accelerating or decelerating). Local environmental conditions include the crowding level of the paving area, obstacle distribution, and lighting conditions, all of which influence the user's movement pattern. By combining the above information, machine learning models or rule-based inference engines can be used to generate the user's most likely movement path and required time during the break.

[0123] S44. Evaluate the behavioral semantic consistency of the predicted path;

[0124] In this embodiment, behavioral semantic consistency refers to whether the predicted path conforms to the natural movement logic and behavioral patterns of humans in a specific scenario. For example, the predicted path should not pass through walls, nor should it make unreasonable sharp turns or sudden speed changes within a short period of time.

[0125] S45. If the predicted path has high behavioral semantic consistency and matches the actual break time, then generate missing footstep signals in the broken area and insert them into the human behavior chain to complete the missing trajectory segments in the human behavior chain, so as to obtain a complete travel trajectory.

[0126] In this embodiment, high behavioral semantic consistency indicates that the predicted path is logically reasonable and conforms to user behavior habits. Matching actual break times means that the time required for the predicted path basically matches the actual observed break duration. Generating missing footstep signals involves synthesizing virtual footstep signals based on the inferred step frequency and step length on the predicted path, and seamlessly integrating these generated signals into the human behavior chain to achieve trajectory completion.

[0127] The technical solution described above monitors the continuity of the human behavior chain in real time. Once a trajectory break is detected, a refined inference mechanism is immediately initiated. First, non-human-related wheel signals unrelated to user behavior are excluded to ensure the purity of subsequent inferences. Then, by combining user movement characteristics, gait information, and environmental factors before and after the break, the possible paths of the user during the break are intelligently predicted. By evaluating the behavioral semantic consistency of the predicted paths, the rationality and authenticity of the completed trajectory are ensured. Finally, under the premise of satisfying consistency conditions and time matching, the missing footstep signals are generated and inserted, thereby achieving accurate completion of trajectory segments in the human behavior chain. This solves the problem of signal loss or interruption that may occur during the data collection process of the human behavior chain, significantly improving the completeness and accuracy of travel trajectories.

[0128] In one possible design, Figure 4 This is a flowchart illustrating step S44 according to an exemplary embodiment. (Refer to the attached diagram.) Figure 4 Step S44 includes:

[0129] S441. Monitor micro-behavioral patterns on the predicted path and mark the micro-behavioral patterns as non-navigation-related behavioral deviations, wherein the micro-behavioral patterns include stress event sequences with a duration shorter than a first preset threshold or footstep signals that have a small and brief deviation from the user's average step frequency and step length.

[0130] In this embodiment, monitoring micro-behavioral patterns on the predicted path refers to the system continuously analyzing stress event data around the predicted path to identify signal sequences that do not constitute regular gait or cane-touching signals, but may reflect subtle user movements.

[0131] Micro-behavioral patterns are pressure signals exerted on the paving stones by users during movement for non-navigational purposes (such as observing the environment, adjusting posture, or briefly pausing). Specifically, these patterns include sequences of pressure events with durations shorter than a preset threshold, such as a user briefly pausing or lightly stepping on a point, the duration of which may be much shorter than the duration of a normal step; or footstep signals with small, transient deviations from the user's average cadence and stride length, such as a sudden increase or decrease in cadence or a slight change in stride length, which quickly returns to normal. These deviations are usually transient and localized and should not be considered substantial breaks or errors in the trajectory.

[0132] S442. Compare the predicted path with the deviation of the non-navigation related behavior, and determine whether there is a local deviation between the predicted path and the user's actual travel trajectory, and whether the local deviation matches the deviation of the non-navigation related behavior in time and space.

[0133] In this embodiment, signals caused by subtle user behaviors can be distinguished from actual navigation trajectory breaks or anomalies through comparison. For example, the system can assign a specific label or attribute to these identified micro-behavioral patterns, indicating that they are "non-navigation-related." In practical applications, the system matches the predicted user's travel path with these labeled micro-behavioral patterns in time and space. For example, the system checks whether a segment of the predicted path occurs within the same time period and the same tile area as a non-navigation-related behavioral deviation. This determines whether there is a local deviation between the predicted path and the user's actual travel trajectory, and whether the local deviation coincides with the non-navigation-related behavioral deviation in time and space. A local deviation refers to an inconsistency between the predicted path and the actual perceived sequence of stress events; this inconsistency may manifest as a missing point, offset, or additional stop on the predicted path. If this local deviation highly coincides with a non-navigation-related behavioral deviation in time and space, it indicates that the deviation is likely caused by that micro-behavioral behavior.

[0134] S443. If the local deviation matches the non-navigation-related behavior deviation in time and space, it is determined that the local deviation is caused by user behavior, the semantic consistency evaluation result of the predicted path is maintained, and the predicted path is adjusted according to the non-navigation-related behavior deviation to complete the missing trajectory segments in the human behavior chain.

[0135] In this embodiment, the system confirms that the local deviation is not a navigation error, but rather a reflection of normal user behavior, and therefore will not affect the overall semantic consistency judgment of the predicted path. Based on this, the system will fine-tune the predicted path according to these non-navigation-related behavioral deviations, such as inserting a brief stop in the predicted path or adjusting the path's slight curves, so that the completed trajectory is closer to the user's actual behavior, thereby obtaining a complete travel trajectory.

[0136] The technical solution described above, by introducing a mechanism for monitoring and identifying micro-behavioral patterns, can precisely distinguish between brief, non-navigation-related behavioral deviations and actual trajectory breaks that may occur during normal walking. First, by monitoring stress event sequences with durations shorter than a preset threshold or footstep signals that deviate slightly and briefly from the user's average step frequency and stride length, the system can capture subtle user movements such as brief pauses, posture adjustments, or looking around. These micro-behavioral patterns are labeled as non-navigation-related behavioral deviations, which can then be correlated with local deviations on the predicted path during subsequent path comparison. Subsequently, when a local deviation on the predicted path is found to coincide with these non-navigation-related behavioral deviations in time and space, the system can accurately determine that these deviations are not caused by navigation errors but are manifestations of the user's own behavior. This avoids misjudging these normal user behaviors as trajectory inconsistencies, thus maintaining the semantic consistency evaluation results of the predicted path, and allowing for fine-tuning of the predicted path only based on these micro-deviations, ensuring the accuracy and naturalness of trajectory completion.

[0137] In one example, suppose a user experiences a brief break in their behavioral chain while traversing a paved area. When evaluating the behavioral semantic consistency of the predicted path, the system first detects a series of extremely short-duration stress events near the break, such as the user pausing for less than a second at a point, or experiencing minute, irregular changes in their stride frequency and stride length within a very short time. These signals are identified and labeled as non-navigation-related behavioral deviations; for example, the user might simply stop briefly to check their phone or adjust their backpack. The system then compares the predicted user path with these labeled non-navigation-related behavioral deviations. If a local deviation is found on the predicted path that highly matches these micro-behaviors in time and space—for example, the predicted path indicates the user should proceed in a straight line, but the actual stress signals show a brief pause at that point—the system determines that this local deviation is caused by the user's own brief behavior, rather than an actual navigation error. Therefore, the system maintains the overall semantic consistency evaluation result of the predicted path and makes fine adjustments to the predicted path based on this brief pause, such as inserting a brief stop in the predicted path, thereby more accurately and naturally completing the missing trajectory segments in the human behavior chain.

[0138] In one possible design, Figure 5 This is a flowchart illustrating step S41 according to an exemplary embodiment. (Refer to the attached document.) Figure 5 Step S41 includes:

[0139] S411. Monitor the coherence of the human-centered behavior chain and contextually assess changes in the coherence of the human-centered behavior chain.

[0140] In this embodiment, monitoring the continuity of the human behavior chain refers to continuously tracking the user's movement trajectory in the smart tile area and recording the temporal and spatial continuity between trajectory points. Contextualizing the changes in the continuity of the human behavior chain means that when assessing changes in continuity, the current environmental background and the user's behavioral characteristics are comprehensively considered, rather than simply based on the presence or absence of a signal or a simple time interval.

[0141] S412. Determine whether the change in continuity is caused by a signal missing with a duration shorter than the second preset threshold or by a foot signal that has a small, brief deviation from the user's average step frequency and step length.

[0142] In this embodiment, step S412 filters out minor disturbances that do not represent actual trajectory breaks. For example, a user briefly lifting their foot or shifting their center of gravity while walking, or a sensor experiencing data loss for a very short time, may cause a brief loss of signal or a minor deviation in gait. The preset threshold can be set according to the actual application scenario and user behavior habits to distinguish between normal behavioral fluctuations and abnormal breaks.

[0143] S413. Determine whether the continuous change is consistent with the stability of the local environment;

[0144] In this embodiment, the consistent changes in the human behavior chain are compared with the real-time environmental state of the tile area. For example, if the consistent changes occur when environmental parameters (such as light, temperature, humidity, noise, etc.) fluctuate drastically, it may indicate that the change is caused by environmental factors rather than a break in user behavior. The stability of the local environment can be obtained by continuously monitoring environmental sensor data.

[0145] S414. If the change in continuity is not caused by the absence of a signal with a duration shorter than a preset threshold or by a foot signal that has a small and brief deviation from the user's average step frequency and step length, and the change in continuity does not match the stability of the local environment, then it is determined that the human behavior chain has been broken, and the inference mechanism is activated.

[0146] In this embodiment, a valid break in the human behavior chain is only determined when the change in continuity is not caused by the aforementioned transient signal loss or minor gait deviation, and when the change is inconsistent with the stability of the local environment. This means that false breaks caused by non-critical factors are excluded, ensuring that only genuine trajectory interruptions trigger subsequent inference mechanisms.

[0147] The technical solution described above, by introducing a contextualized evaluation mechanism and combining it with multi-dimensional judgment conditions, effectively solves the problem of false alarms or missed alarms that may exist in the trajectory breakage judgment in the basic solution. First, contextualized evaluation of the continuity changes in the human behavior chain allows the system to more comprehensively understand the background of the continuity changes, avoiding over-reliance on a single indicator. Second, by judging whether the continuity change is caused by signal loss with a duration shorter than a preset threshold or foot signals with slight, transient deviations from the user's average step frequency and step length, it can effectively identify and eliminate trajectory interruption illusions caused by non-substantial factors such as sensor momentary malfunctions, network fluctuations, or normal gait adjustments by the user. For example, when a user briefly pauses in place or adjusts their posture, there may be short-term signal loss or step frequency / step length deviations, but these are not true trajectory breaks. Third, by judging whether the continuity change matches the stability of the local environment, the accuracy of the judgment is further enhanced. If the continuity change occurs simultaneously with drastic environmental fluctuations, it may indicate that environmental factors, rather than user behavior, caused the signal anomaly. Only when these non-substantive factors are excluded, and the consistent change does not match the environmental stability, is it confirmed as a valid break, thus initiating the subsequent trajectory inference mechanism. The multiple, contextualized judgments in the technical solution of this embodiment make the identification of trajectory breaks more accurate and robust.

[0148] In one example, suppose a user is walking in a smart tiled area, and their human behavior chain is being continuously monitored.

[0149] At a certain moment, the system detected a brief absence of pressure signal, lasting 0.5 seconds, which was shorter than the preset 1-second threshold. Simultaneously, the user's cadence and stride length also showed slight, transient deviations, but these deviations remained within the normal fluctuation range of the user's average gait. Furthermore, environmental sensors in the tiled area indicated that the local environmental conditions in that area were stable at this time, with no abnormal fluctuations.

[0150] According to the technical solution of this application, the system first detects changes in the continuity of the human behavior chain. Then, the system determines whether the change in continuity is caused by a signal absence lasting less than a preset threshold or by footstep signals that deviate slightly and briefly from the user's average step frequency and step length. In this example, since the signal absence time is less than the threshold and the gait deviation is slight, the system considers these changes to be caused by non-substantial factors. Simultaneously, the system also determines whether the change in continuity is consistent with the stability of the local environment. Since the local environment is stable, the system considers the change in continuity to be consistent with environmental stability.

[0151] Therefore, according to the condition that "if the change in coherence is not caused by a signal deficiency with a duration shorter than a preset threshold or a footstep signal with a small, transient deviation from the user's average step frequency and step length, and the change in coherence does not match the stability of the local environment," the situation in this example does not meet the two conditions of "not caused by..." and "does not match." That is, the change in coherence is caused by a short-term signal deficiency / small gait deviation, and it matches the stability of the local environment. Therefore, the system will not determine that there is a valid break in the human behavior chain, nor will it activate the inference mechanism.

[0152] Conversely, if the system detects a prolonged (e.g., 3 seconds) signal loss in the human behavior chain, and the user's gait data is completely interrupted, while the local environment remains stable, the system will determine that this change in continuity is not caused by a signal loss shorter than a preset threshold or a minor gait deviation, and that this change in continuity does not match the stability of the local environment (because the environment is stable but the trajectory is broken). In this case, the system will determine that the human behavior chain has been effectively broken and initiate an inference mechanism to infer and complete the missing trajectory segment.

[0153] In this way, the proposed solution can effectively distinguish between real trajectory breaks and non-critical disturbances, thereby improving the accuracy and efficiency of trajectory completion.

[0154] In one possible design, the contextual assessment of the coherence changes in the human-centered behavioral chain includes:

[0155] S4111. Continuously acquire real-time environmental parameters of the paving area, and perform multi-timescale analysis on the real-time environmental parameters to identify and obtain the instantaneous fluctuations, short-term trends and long-term baselines of the real-time environmental parameters.

[0156] In this embodiment, environmental data is continuously collected by various sensors deployed within the paved area, such as temperature sensors, humidity sensors, light sensors, wind speed sensors, and noise sensors. These environmental parameters reflect the external conditions of the paved area, aiming to provide comprehensive environmental background information for subsequent consistency assessments. The collected environmental data is then processed and analyzed at different time granularities. For example, instantaneous fluctuations refer to drastic changes occurring within a very short time (e.g., seconds or milliseconds), which may be caused by sudden events; short-term trends refer to patterns of change observed over minutes to hours, such as weather changes or changes in pedestrian traffic; and long-term baselines refer to relatively stable average states maintained over days, weeks, or even months, reflecting the inherent environmental characteristics of the area.

[0157] S4112. When the real-time environmental parameters experience instantaneous large fluctuations, the local environmental state fast sampling mode is activated, the sampling frequency is increased, and multi-sensor cross-validation is activated to obtain fine-grained local environmental state.

[0158] In this embodiment, when the real-time environmental parameters experience sudden and significant fluctuations, such as sudden strong winds, rain, snow, or external impacts, the system will activate a rapid sampling mode for the local environmental state, increasing the sampling frequency of relevant sensors to capture more densely packed real-time environmental parameters. Simultaneously, multi-sensor cross-validation is initiated, comparing real-time environmental parameters from sensors of different types or locations for mutual verification and calibration to ensure data accuracy and reliability. Finally, fine-grained local environmental states are obtained to avoid misjudgments due to insufficient sampling or data errors.

[0159] S4113. Based on the short-term trend and long-term baseline of the real-time environmental parameters, dynamically adjust the sensitivity threshold of the human behavior chain coherence assessment.

[0160] In this embodiment, when real-time environmental parameters are stable in the long term or have a flat short-term trend, the sensitivity threshold can be appropriately increased to reduce false alarms; conversely, when the short-term trend of real-time environmental parameters shows signs of instability, the sensitivity threshold can be appropriately decreased to detect potential consistency issues earlier. This makes the consistency assessment more adaptable to the current environmental context, improving the accuracy and robustness of the assessment.

[0161] S4114. When the instantaneous fluctuation of the real-time environmental parameters coincides with the coherent change of the human behavior chain in time and space, the impact of the instantaneous fluctuation on the coherent change shall be given priority, and the contextualized evaluation result shall be corrected accordingly.

[0162] In this embodiment, when the instantaneous fluctuations of the real-time environmental parameters coincide with the coherence changes of the human-centered behavior chain in time and space—for example, when a brief break in the human-centered behavior chain is detected in a certain tile area, and the environmental sensor in that area also records a momentary large vibration—the system will prioritize the impact of the instantaneous fluctuations on the coherence changes. That is, if the environmental fluctuations can reasonably explain the coherence changes in the behavior chain, then the change may not be caused by an interruption in user behavior, but rather by environmental interference. Based on this, the contextualized evaluation results are corrected, for example, by marking the coherence change as environmental interference rather than a valid break, thereby avoiding misjudging environmental interference as a break in the human-centered behavior chain and improving the accuracy of trajectory inference.

[0163] The technical solution of the above embodiments, firstly, by continuously acquiring real-time environmental parameters of the paving area and performing multi-timescale analysis, comprehensively grasps the dynamic changes of the environment. Subsequently, when the environment experiences sudden and significant fluctuations, the system activates a rapid sampling mode for local environmental states and multi-sensor cross-validation to obtain more detailed and accurate local environmental state information. Based on this, by combining the short-term trends and long-term baselines of environmental parameters, the sensitivity threshold for assessing the coherence of the human behavior chain is dynamically adjusted, enabling the assessment process to adapt to different environmental backgrounds and avoiding misjudgments that may arise from fixed thresholds. Furthermore, when the instantaneous fluctuations of real-time environmental parameters coincide with changes in the coherence of the human behavior chain in time and space, the system can prioritize the impact of environmental fluctuations and correct the assessment results, thereby effectively distinguishing between signal changes caused by environmental interference and actual breaks in the human behavior chain, significantly improving the accuracy of contextualized assessment.

[0164] In one example, suppose a pedestrian paving navigation system is monitoring the movement of a pedestrian in a busy commercial area. At some point, the pedestrian passes through an area undergoing minor construction, where a small cart suddenly passes by, causing the paving sensor to receive a brief, large fluctuation in pressure signal, accompanied by a momentary high reading from the ambient noise sensor.

[0165] According to the technical solution of this application, the system continuously acquires real-time environmental parameters of the paving area and identifies instantaneous large fluctuations in the construction area. At this time, a rapid sampling mode for the local environmental state is activated, increasing the sampling frequency of the sensors in that area and performing multi-sensor cross-validation to confirm that this is an instantaneous event caused by external environmental interference. Simultaneously, the system dynamically adjusts the sensitivity threshold for assessing the coherence of the human behavior chain based on the historical environmental data and short-term trends of the area.

[0166] When the human behavior chain continuity assessment module detects a brief signal interruption in the construction area, the system prioritizes the impact of environmental fluctuations because the instantaneous fluctuations of real-time environmental parameters closely match the signal interruption in time and space. After correction, the signal interruption is determined to be caused by external environmental interference, rather than an actual interruption of pedestrian movement. Therefore, the system will not misjudge this as a valid interruption and initiate unnecessary inference mechanisms, but will maintain the continuity of the human behavior chain, ensuring the accuracy and integrity of the pedestrian's movement trajectory.

[0167] In another possible design, the contextual assessment of the coherence changes in the human-centered behavioral chain further includes:

[0168] S4115. Continuously acquire real-time environmental parameters of the pedestrian paving area and monitor abnormal states of the real-time environmental parameters, wherein the abnormal states include data loss, data anomaly, or data delay.

[0169] In this embodiment, the system continuously receives environmental data streams, i.e., real-time environmental parameters, from various sensors (e.g., temperature sensors, humidity sensors, light sensors, noise sensors, etc.) deployed within the tiled area. These parameters reflect the current physical state of the tiled area and provide contextual information for assessing the coherence of human behavior chains. Monitoring abnormal states of these real-time environmental parameters involves performing real-time quality checks on the received environmental data to identify any abnormalities in the data stream. Specific abnormal states may include: missing data (no data received within the expected timeframe or incomplete data packets); data anomalies (data values ​​exceeding a preset reasonable range or exhibiting sudden changes); and data delays (data arrival times later than its generation time, resulting in insufficient data timeliness).

[0170] S4116. When the aforementioned abnormal state occurs, the environmental data multi-source cross-validation mechanism is activated, the multi-source cross-validation mechanism including:

[0171] Environmental parameters are obtained from adjacent tile areas and compared with the abnormal environmental parameters;

[0172] Obtain the average environmental parameters for the same time period and region from historical environmental data, and compare them with the abnormal environmental parameters;

[0173] In this embodiment, the multi-source cross-validation mechanism introduces multiple data sources for mutual verification and supplementation to improve the accuracy and completeness of abnormal environmental parameters. Specifically, the system queries environmental data collected from other paving areas physically adjacent to the abnormal paving area. By comparing the normal environmental parameters of these adjacent areas with those of the abnormal area, the locality or universality of the anomaly can be determined, providing a reference for correction. Simultaneously, the average environmental parameters for the same time period and area are obtained from historical environmental data and compared with the abnormal environmental parameters. Leveraging the regularity of historical data accumulation, by comparing with the average or typical environmental parameters of the same time period and area in the past, the degree of deviation of the current anomaly can be assessed, providing a benchmark for data correction.

[0174] S4117. Based on the comparison results, the abnormal environmental parameters are corrected or supplemented, and based on the corrected or supplemented real-time environmental parameters, combined with the dynamic characteristics of the human-centered behavior chain such as movement speed, direction, step frequency, and step length, the coherence of the human-centered behavior chain is comprehensively scored.

[0175] In this embodiment, correction refers to adjusting erroneous or unreasonable data values ​​to restore them to a reasonable range; completion refers to estimating and filling in missing data to eliminate the impact of abnormal environmental parameters on subsequent evaluations. Subsequently, after the environmental parameters undergo reliability processing through correction or completion, the system combines them with the user's own behavioral characteristics (such as movement speed, direction, cadence, and stride length) for a more comprehensive and accurate consistency assessment. For example, if environmental parameters show a sudden dimming of the current area's light, and the user's cadence and stride length also change accordingly, this may indicate that the user is adapting to environmental changes rather than a break in the trajectory. Through comprehensive scoring, it is possible to more accurately determine whether the consistency of the human behavior chain is reasonably affected by the environment or the user's own behavior, thereby avoiding misjudgments.

[0176] The technical solution of the above embodiments, by introducing a multi-source cross-validation mechanism for environmental data, effectively assesses the coherence changes of the human behavior chain in a contextualized manner, thereby handling abnormal states of real-time environmental parameters and improving the accuracy of coherence assessment.

[0177] When real-time environmental parameters suffer from missing, abnormal, or delayed data, traditional assessment methods may directly use this unreliable data, leading to distorted assessment results. This application's technical solution employs a multi-source cross-validation mechanism. First, environmental parameters are compared from adjacent tile areas, helping to determine the locality of anomalies and allowing for correction based on spatial correlation. Second, average environmental parameters from the same time period and area in historical environmental data are compared, leveraging temporal regularity to provide a temporal reference benchmark for correcting or supplementing abnormal data. These two comparison methods effectively correct or supplement abnormal environmental parameters, ensuring the accuracy and reliability of the environmental data used for assessment. Finally, based on the corrected or supplemented real-time environmental parameters, a comprehensive score is calculated by combining dynamic characteristics of human behavior chains such as movement speed, direction, step frequency, and step length. This ensures that the consistency assessment not only considers environmental factors but also fully integrates the user's own behavioral patterns, enabling consistent judgments even when environmental data is uncertain.

[0178] In one possible design, obtaining environmental parameters from adjacent tile areas and comparing them with the abnormal environmental parameters includes:

[0179] S41161. Identify the degree of fluctuation of environmental parameters in each adjacent paving area, and calculate the reliability weight of environmental parameters in each adjacent paving area based on the degree of fluctuation of environmental parameters in each adjacent paving area.

[0180] In this embodiment, identifying the degree of fluctuation of environmental parameters in each adjacent paving tile area refers to performing statistical analysis on the environmental parameter data of each adjacent paving tile area within a certain time window, such as calculating its standard deviation, variance, or coefficient of variation. A greater degree of fluctuation generally means that the environmental parameters in that area are more unstable or more likely to be disturbed. These environmental parameters can include various physical quantities such as temperature, humidity, light intensity, and noise. In the calculation of reliability weights, the smaller the degree of fluctuation, the higher the reliability weight; conversely, the greater the degree of fluctuation, the lower the reliability weight. This weight can be calculated using an inverse proportional function, a Gaussian function, or other nonlinear functions to quantify the contribution of each adjacent area's data to the correction of abnormal parameters.

[0181] S41162. Select environmental parameters of adjacent tile areas with reliability weights higher than a preset threshold, and perform a weighted average of the selected environmental parameters of adjacent tile areas to obtain fused environmental parameters.

[0182] In this embodiment, the preset threshold can be set according to the actual application scenario and the requirements for data reliability. For example, it can be set to 0.6 or 0.7. This step aims to filter out adjacent area data with poor quality, excessive fluctuations, and low reliability, so as to avoid their negative impact on the overall correction process. Subsequently, the purpose of weighted averaging is to combine the filtered adjacent area environmental parameters with different reliability weights to form a more representative and accurate environmental parameter value. For example, the fused environmental parameter can be obtained by multiplying each selected adjacent area environmental parameter by its corresponding reliability weight, then summing all the products, and then dividing by the sum of all reliability weights.

[0183] S41163. Compare the fused environment parameters with the abnormal environment parameters;

[0184] In this embodiment, the comparison results will be used for subsequent correction or supplementation of abnormal environmental parameters, ensuring that the correction process is based on more reliable and accurate data from neighboring regions, thereby improving the accuracy and effectiveness of the correction.

[0185] The technical solution of the above embodiments quantifies the data quality and stability by identifying the fluctuation level of environmental parameters in each adjacent tile area. A reliability weight is calculated based on this fluctuation level, giving greater influence to adjacent areas with higher data quality in the subsequent fusion process. Furthermore, by setting a preset threshold to select environmental parameters of adjacent tile areas with higher reliability, data sources that may have significant errors or interference can be effectively excluded. Finally, a weighted average of these filtered and weighted adjacent area environmental parameters generates a more stable and accurate fusion environmental parameter, which more realistically reflects the actual environmental conditions of abnormal tile areas. Therefore, comparing the fusion environmental parameter with the abnormal environmental parameter provides a more precise reference.

[0186] In one possible design, the step of correcting or supplementing the abnormal environmental parameters based on the comparison results includes:

[0187] S41171. Based on the comparison results, identify the deviation dimension and degree of deviation of the abnormal environmental parameters;

[0188] In this embodiment, the system will accurately analyze the environmental parameters (such as temperature, humidity, light intensity, air pressure, etc.) that are abnormal based on the comparison results of multi-source cross-validation, and the extent of the abnormality, such as slight deviation or serious deviation from the preset threshold or normal range.

[0189] S41172. Based on the deviation dimension and the deviation degree, and in conjunction with the physical characteristics of the pedestrian paving area and historical environmental event records, identify the dominant abnormal factors;

[0190] In this embodiment, the physical characteristics of the paving area include its material, laying method, geographical location, sensor type and deployment, etc., while historical environmental event records may include information such as past extreme weather events, equipment failure records, construction activities, and peak periods of pedestrian traffic. By comprehensively analyzing this information, the main causes of abnormal environmental parameters can be determined, such as sensor failure, partial obstruction, sudden weather changes, network transmission delays, or other external interference.

[0191] S41173. Based on the type of the dominant abnormal factor, select a corresponding correction strategy from a preset correction strategy library, and perform hierarchical correction or completion on the abnormal environment parameters based on the correction strategy. The hierarchical correction or completion includes:

[0192] Prioritize the correction of environmental parameter dimensions that are most affected by the aforementioned dominant anomaly factors;

[0193] Auxiliary correction of environmental parameter dimensions affected by minor anomalies;

[0194] The corrected environmental parameters are checked for consistency to ensure that the correction process does not introduce new errors.

[0195] In this embodiment, the preset correction strategy library stores correction algorithms or rules designed for different anomalies and deviation types. For example, for sensor data drift, moving average, Kalman filtering, or machine learning-based prediction models may be used for correction; for missing data, interpolation methods (such as linear interpolation, spline interpolation), prediction models based on historical data, or data fusion of adjacent paving areas may be used for completion; for sudden environmental changes, strategies such as dynamic threshold adjustment or multi-sensor cross-calibration may be used.

[0196] Subsequently, according to the aforementioned correction strategy, the abnormal environmental parameters are corrected or supplemented in a hierarchical manner. Specifically, environmental parameter dimensions most affected by the dominant anomaly are corrected first. For example, if the dominant anomaly is a sensor malfunction causing a significant deviation in temperature data, the temperature parameter will be processed first, employing the correction strategy most suitable for that type of malfunction. Simultaneously, environmental parameter dimensions affected by secondary anomalies are corrected secondarily. For example, if humidity data is also slightly affected, after the temperature data is corrected, secondary corrections will be made based on the degree and type of impact to ensure the accuracy of all relevant parameters.

[0197] Finally, a consistency check is performed on the corrected environmental parameters to ensure that no new errors were introduced during the correction process. This check aims to verify that the corrected environmental parameters are logically reasonable and consistent with the overall environmental conditions of the paved area. For example, it checks whether the corrected parameters such as temperature and humidity are within reasonable ranges, whether they match the parameter trends of adjacent areas, and whether there are significant conflicts with historical data patterns. This check effectively avoids introducing new errors or causing data distortion during the correction process, thereby ensuring the overall quality of the environmental data.

[0198] The technical solutions described above, through refined correction or completion of abnormal environmental parameters, effectively improve the accuracy and reliability of environmental data processing. First, by identifying the deviation dimensions and degrees of abnormal environmental parameters, the problem can be precisely located, laying the foundation for subsequent correction work. Second, by combining the physical characteristics of the paving area and historical environmental event records to identify the dominant abnormal factors, the correction process can address the root cause of the problem, rather than merely treating the symptoms. Therefore, selecting the corresponding correction strategy from a pre-set correction strategy library ensures the professionalism and effectiveness of the correction method. Furthermore, by adopting a hierarchical correction or completion approach, priority is given to processing the parameter dimensions most affected by the dominant abnormal factors, while assisting in correcting dimensions affected by secondary factors, making the correction process more efficient and accurate. Finally, by performing a consistency check on the corrected environmental parameters, new errors introduced during the correction process can be effectively avoided, thereby ensuring the overall quality of the environmental data.

[0199] In summary, the pedestrian paving data analysis method based on big data provided in Embodiment 1 of this invention constructs an intelligent processing flow from raw pressure signals to complete travel trajectories. Through a series of steps such as adaptive signal processing, multi-dimensional signal classification, intelligent wheel signal recognition, and context-aware trajectory completion, a complete and efficient navigation data analysis method is formed. The various technical features work together to solve problems such as signal distortion, difficulty in distinguishing wheel interference, and missing trajectory data in existing technologies. Ultimately, it achieves an accurate and complete depiction of the travel trajectory of visually impaired individuals, providing them with more complete and accurate travel navigation services.

[0200] Example 2

[0201] Embodiment 2 of the present invention provides a pedestrian paving tile navigation data analysis system based on big data. Figure 6 This is a block diagram illustrating a pedestrian paving stone navigation data analysis system based on big data, according to an exemplary embodiment. (See attached diagram.) Figure 6 The system includes:

[0202] Signal acquisition module 1 acquires pressure signals from the pedestrian paving area and adjusts the processing parameters of the pressure signals according to the local environmental conditions of the pedestrian paving area to extract pressure events from the pressure signals, wherein the pressure events include time, location and pressure characteristics;

[0203] Preliminary classification module 2, based on the pressure characteristics, preliminarily classifies the pressure events into footstep signals, guide cane ground contact signals, and wheel signals;

[0204] The human-centered behavior chain construction module 3 correlates the footstep signal and the guide cane touch-the-ground signal in time and space to form a pedestrian movement trajectory, and determines whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent:

[0205] The trajectory forming unit 3-1 is used to correlate the preliminarily classified footstep signals and the preliminarily classified guide cane ground contact signals in time and space to form a preliminary pedestrian movement trajectory;

[0206] Wheel signal judgment unit 3-2: If the wheel signal has a spatiotemporal correlation or the same movement pattern as the preliminary pedestrian movement trajectory, then the wheel signal is confirmed as the user's suitcase wheel signal and fused into the pedestrian movement trajectory to construct a human-centered behavior chain;

[0207] The wheel signal exclusion unit 4 excludes the wheel signal if the wheel signal does not have a spatiotemporal correlation with the preliminary pedestrian movement trajectory or the movement pattern is inconsistent with it.

[0208] The trajectory completion module 5, based on the contextual information of the human behavior chain, infers and completes the missing trajectory segments in the human behavior chain to obtain a complete travel trajectory.

[0209] In summary, the pedestrian paving visual interactive image generation system provided in Embodiment 2 of this invention constructs an intelligent processing flow from raw pressure signals to complete travel trajectories. Through a series of steps such as adaptive signal processing, multi-dimensional signal classification, intelligent wheel signal recognition, and context-aware trajectory completion, a complete and efficient navigation data analysis method is formed. The various technical features cooperate with each other to solve the problems of signal distortion, difficulty in distinguishing wheel interference, and missing trajectory data in the prior art. Ultimately, it realizes an accurate and complete depiction of the travel trajectory of visually impaired people, and can provide visually impaired people with more complete and accurate travel navigation services.

[0210] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0211] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for analyzing pedestrian paving stone navigation data based on big data, characterized in that, The method includes: The pressure signal of the pedestrian paving area is acquired, and the processing parameters of the pressure signal are adjusted according to the local environmental state of the pedestrian paving area to extract pressure events from the pressure signal, wherein the pressure events include time, location and pressure characteristics; Based on the pressure characteristics, the pressure events are initially classified into footstep signals, guide cane ground-touching signals, and wheel signals; The footstep signal and the guide cane touch the ground signal are correlated in time and space to form a pedestrian movement trajectory, and it is determined whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent: If the wheel signal has a spatiotemporal correlation or the same movement pattern as the preliminary pedestrian movement trajectory, then the wheel signal is identified as the user's suitcase wheel signal and fused into the pedestrian movement trajectory to construct a human-centered behavior chain; If the wheel signal does not have a spatiotemporal correlation with the preliminary pedestrian movement trajectory or the movement pattern is inconsistent, then the wheel signal is excluded. Based on the contextual information of the human-centered behavior chain, the missing trajectory segments in the human-centered behavior chain are inferred and completed to obtain the complete travel trajectory; Specifically, the step of inferring and completing missing trajectory segments in the human-centered behavior chain based on contextual information to obtain a complete travel trajectory includes: Monitor the continuity of the human-centered behavior chain; if the human-centered behavior chain is broken, initiate an inference mechanism. Check whether there are wheel-like signals that have been classified as non-human behavior patterns within the fractured area; if so, exclude the wheel-like signals. Based on the movement trend, speed, and direction before and after the break, combined with the user's adaptive gait state before and after the break and the local environmental state of the broken area, the user's path during the break is predicted. The predicted path includes the possible sequence of paving stones and the expected time. Evaluate the behavioral semantic consistency of the predicted path; If the predicted path has high semantic consistency and matches the actual break time, then missing footstep signals are generated in the broken area and inserted into the human behavior chain to complete the missing trajectory segments in the human behavior chain, so as to obtain a complete travel trajectory.

2. The pedestrian paving stone navigation data analysis method according to claim 1, characterized in that, The step of correlating the footstep signal and the guide cane touch the ground signal in time and space to form a pedestrian movement trajectory, and determining whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent, includes: Identify user adaptive behavior patterns and assess the momentum of the pedestrian movement trajectory; By combining the local environmental conditions and the user's adaptive behavior patterns, the time tolerance window and spatial tolerance window associated with the wheel signal and the pedestrian movement trajectory are dynamically adjusted. Based on the momentum of the pedestrian's movement trajectory, predict the path of the pedestrian's movement trajectory, and determine: whether the wheeled signal has a spatiotemporal correlation with the pedestrian's movement trajectory within the time tolerance window; whether the movement pattern of the wheeled signal is consistent with the movement pattern of the preliminary pedestrian movement trajectory; and whether the wheeled signal exhibits local deviations on the predicted path. If all the conditions are met, the wheel signal is confirmed as the user's suitcase wheel signal and integrated into the preliminary pedestrian movement trajectory to form the human behavior chain; If the wheel signal and the pedestrian movement trajectory do not have a spatiotemporal correlation or the movement pattern is inconsistent, the duration, minute movement pattern and periodic characteristics of the wheel signal are identified to exclude the wheel signal.

3. The pedestrian paving stone navigation data analysis method according to claim 1, characterized in that, The evaluation of the behavioral semantic consistency of the predicted path includes: Monitor micro-behavioral patterns on the predicted path and mark the micro-behavioral patterns as non-navigation-related behavioral deviations. The micro-behavioral patterns include stress event sequences with a duration shorter than a first preset threshold or footstep signals that deviate slightly and briefly from the user's average step frequency and step length. By comparing the predicted path with the deviation of the non-navigation-related behavior, it is determined whether there is a local deviation between the predicted path and the user's actual travel trajectory, and whether the local deviation matches the deviation of the non-navigation-related behavior in time and space; If the local deviation matches the non-navigation-related behavior deviation in time and space, it is determined that the local deviation is caused by user behavior. The semantic consistency evaluation result of the predicted path is maintained, and the predicted path is adjusted according to the non-navigation-related behavior deviation to complete the missing trajectory segments in the human behavior chain.

4. The pedestrian paving stone navigation data analysis method according to claim 1, characterized in that, The monitoring of the continuity of the human-centered behavior chain, and the inference mechanism initiated if the human-centered behavior chain is broken, includes: Monitor the coherence of the human-centered behavior chain and contextually assess changes in the coherence of the human-centered behavior chain; Determine whether the change in continuity is caused by a signal missing with a duration shorter than the second preset threshold or by a foot signal that has a small, brief deviation from the user's average step frequency and step length. Determine whether the continuity change is consistent with the stability of the local environment; If the change in continuity is not caused by the absence of signals with a duration shorter than a preset threshold or by footstep signals that have a small and brief deviation from the user's average step frequency and step length, and the change in continuity does not match the stability of the local environment, then it is determined that the human behavior chain has been broken, and the inference mechanism is activated.

5. The pedestrian paving stone navigation data analysis method according to claim 4, characterized in that, The contextualized assessment of the coherence changes in the human-centered behavioral chain includes: Continuously acquire real-time environmental parameters of the paving area, and perform multi-timescale analysis on the real-time environmental parameters to identify the instantaneous fluctuations, short-term trends and long-term baselines of the real-time environmental parameters; When the real-time environmental parameters experience sudden and significant fluctuations, a rapid sampling mode for the local environmental state is activated, the sampling frequency is increased, and multi-sensor cross-validation is initiated to obtain fine-grained local environmental states. Based on the short-term trend and long-term baseline of the real-time environmental parameters, the sensitivity threshold for the human behavior chain coherence assessment is dynamically adjusted. When the instantaneous fluctuations of the real-time environmental parameters coincide with the coherent changes of the human behavior chain in time and space, the impact of the instantaneous fluctuations on the coherent changes is given priority, and the contextualized assessment results are corrected accordingly.

6. The pedestrian paving stone navigation data analysis method according to claim 4, characterized in that, The contextualized assessment of the coherence changes in the human-centered behavioral chain also includes: Continuously acquire real-time environmental parameters of the pedestrian paving area and monitor abnormal states of the real-time environmental parameters, wherein the abnormal states include data missing, data anomaly, or data delay. When the aforementioned abnormal state occurs, a multi-source cross-validation mechanism for environmental data is activated. This multi-source cross-validation mechanism includes: Environmental parameters are obtained from adjacent tile areas and compared with the abnormal environmental parameters; Obtain the average environmental parameters for the same time period and region from historical environmental data, and compare them with the abnormal environmental parameters; Based on the comparison results, the abnormal environmental parameters are corrected or supplemented, and based on the corrected or supplemented real-time environmental parameters, combined with the dynamic characteristics of the human-centered behavior chain such as movement speed, direction, step frequency, and step length, the coherence of the human-centered behavior chain is comprehensively scored.

7. The pedestrian paving stone navigation data analysis method according to claim 6, characterized in that, The step of obtaining environmental parameters from adjacent tile areas and comparing them with the abnormal environmental parameters includes: Identify the degree of fluctuation of environmental parameters in each adjacent paving tile area, and calculate the reliability weight of environmental parameters in each adjacent paving tile area based on the degree of fluctuation of environmental parameters in each adjacent paving tile area. Environmental parameters of adjacent paving tile areas with reliability weights higher than a preset threshold are selected, and the selected environmental parameters of adjacent paving tile areas are weighted and averaged to obtain fused environmental parameters. The fused environment parameters are compared with the abnormal environment parameters.

8. The pedestrian paving stone navigation data analysis method according to claim 6, characterized in that, The step of correcting or supplementing the abnormal environmental parameters based on the comparison results includes: Based on the comparison results, the deviation dimension and degree of the abnormal environmental parameters are identified; Based on the deviation dimension and the deviation degree, combined with the physical characteristics of the pedestrian paving area and historical environmental event records, the dominant anomaly factors are identified. Based on the type of the dominant anomaly factor, a corresponding correction strategy is selected from a preset correction strategy library, and the anomaly environment parameters are corrected or completed hierarchically based on the correction strategy. The hierarchical correction or completion includes: Prioritize the correction of environmental parameter dimensions that are most affected by the aforementioned dominant anomaly factors; Auxiliary correction of environmental parameter dimensions affected by minor anomalies; The corrected environmental parameters are checked for consistency to ensure that the correction process does not introduce new errors.

9. A pedestrian paving stone navigation data analysis system based on big data, characterized in that, The system includes: The signal acquisition module acquires the pressure signal of the pedestrian paving area and adjusts the processing parameters of the pressure signal according to the local environmental state of the pedestrian paving area to extract pressure events from the pressure signal, wherein the pressure events include time, location and pressure characteristics; The preliminary classification module, based on the pressure characteristics, preliminarily classifies the pressure events into footstep signals, guide cane ground contact signals, and wheel signals; The human-centered behavior chain construction module correlates the footstep signals and the guide cane touch-the-ground signals in time and space to form a pedestrian movement trajectory, and determines whether the wheel signal has a spatiotemporal correlation with the pedestrian movement trajectory and whether the movement pattern is consistent: The trajectory forming unit is used to correlate the preliminarily classified footstep signals and the preliminarily classified guide cane ground-touching signals in time and space to form a preliminary pedestrian movement trajectory. The wheel signal determination unit, if the wheel signal has a spatiotemporal correlation or the same movement pattern as the preliminary pedestrian movement trajectory, then confirms the wheel signal as the user's suitcase wheel signal and integrates it into the pedestrian movement trajectory to construct a human-centered behavior chain; The wheel signal exclusion unit excludes the wheel signal if the wheel signal does not have a spatiotemporal correlation with the preliminary pedestrian movement trajectory or the movement pattern is inconsistent with it. The trajectory completion module, based on the contextual information of the human behavior chain, infers and completes the missing trajectory segments in the human behavior chain to obtain a complete travel trajectory; Specifically, the step of inferring and completing missing trajectory segments in the human-centered behavior chain based on contextual information to obtain a complete travel trajectory includes: Monitor the continuity of the human-centered behavior chain; if the human-centered behavior chain is broken, initiate an inference mechanism. Check whether there are wheel-like signals that have been classified as non-human behavior patterns within the fractured area; if so, exclude the wheel-like signals. Based on the movement trend, speed, and direction before and after the break, combined with the user's adaptive gait state before and after the break and the local environmental state of the broken area, the user's path during the break is predicted. The predicted path includes the possible sequence of paving stones and the expected time. Evaluate the behavioral semantic consistency of the predicted path; If the predicted path has high semantic consistency and matches the actual break time, then missing footstep signals are generated in the broken area and inserted into the human behavior chain to complete the missing trajectory segments in the human behavior chain, so as to obtain a complete travel trajectory.

Images