Position navigation device and method based on mobile cloud passenger flow data

Abstract

The invention relates to the technical field of intelligent navigation, in particular to a position guide device and method based on mobile cloud passenger flow data, and the device comprises a cloud server which is used for collecting and processing anonymous passenger flow data from a mobile terminal in real time and generating a dynamic thermodynamic diagram; the indoor and outdoor positioning module determines the current position of the user through a multi-source fusion positioning technology; the navigation terminal is used for generating an optimal navigation route through a path planning algorithm according to the user preference and the real-time passenger flow density and outputting navigation guidance; according to the method, the real-time passenger flow analysis and dynamic path planning technology is adopted, the congestion area can be accurately predicted and avoided, the navigation efficiency is greatly improved, and special requirements such as barrier-free navigation are supported by combining an AR (Augmented Reality) navigation interface with a personalized recommendation algorithm, so that different user groups can obtain customized navigation experience.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent navigation, and particularly to a location navigation device and method based on mobile cloud passenger flow data. Background Art

[0002] With the continuous expansion of the scale of public places such as large commercial complexes, transportation hubs, and tourist attractions, traditional navigation methods (such as static signboards and paper maps) have been difficult to meet the navigation needs of users. In the prior art, indoor positioning systems based on Wi-Fi or Bluetooth beacons, such as Apple iBeacon, can provide basic indoor location services, while outdoor navigation based on GPS, such as Gaode Map and Baidu Map, is suitable for path planning in open environments. In addition, some intelligent navigation devices, such as airport navigation robots, attempt to combine simple path algorithms to achieve automatic guidance, but the adaptability of these solutions in a dynamic passenger flow environment is still limited.

[0003] The prior art has the following obvious deficiencies. First, traditional navigation systems rely on static map data and cannot perceive passenger flow changes in real time, resulting in possible congestion on the planned route. For example, users are navigated to a high-density area where a promotion is being held. Also, the existing centralized data processing architecture has a risk of privacy leakage and a high response latency, making it difficult to meet the real-time navigation requirements. Existing navigation terminals lack intelligent interaction capabilities, such as being unable to dynamically adjust the route according to user characteristics, resulting in a poor user experience for special groups such as the elderly and children.

[0004] Therefore, in view of the above problems, the present invention proposes a location navigation device and method based on mobile cloud passenger flow data, which realizes dynamic path optimization while protecting user privacy through anonymized data collection, real-time passenger flow analysis, and federated learning technology. Summary of the Invention

[0005] In order to overcome the problem that traditional navigation systems rely on static map data and cannot perceive passenger flow changes in real time, resulting in possible congestion on the planned route, the present invention proposes a location navigation device and method based on mobile cloud passenger flow data.

[0006] The technical solution of the present invention is as follows: A location navigation device based on mobile cloud passenger flow data includes:

[0007] A cloud server for collecting and processing anonymized passenger flow data from mobile terminals in real time to generate a dynamic heat map;

[0008] An indoor and outdoor positioning module for determining the user's current location through multi-source fusion positioning technology;

[0009] The navigation terminal generates the optimal navigation route according to the user preferences and the real-time passenger flow density through the path planning algorithm and outputs the navigation guidance.

[0010] Preferably, the cloud server further includes:

[0011] The data cleaning unit is used to filter abnormal location data;

[0012] The clustering analysis unit uses the DBSCAN algorithm to identify high-density passenger flow areas;

[0013] The prediction module predicts the passenger flow distribution in the future period based on historical data and the LSTM model.

[0014] Preferably, the multi-source fusion positioning technology adopts a loose coupling architecture, including Wi-Fi positioning and Bluetooth beacons. Wi-Fi positioning is based on RSSI fingerprint library matching, and the deployment density of Bluetooth beacons is 4 per 100 ㎡.

[0015] Preferably, the navigation terminal integrates the AR display function, captures the user's surrounding environment in real time through the camera, and superimposes virtual navigation information on the screen. The virtual navigation information includes dynamic arrows, hotspot labels, and dynamic obstacle avoidance prompts. The dynamic arrows display the direction guidance in the real scene according to the path planning result, and the user only needs to move following the arrows. At the same time, floating labels are superimposed at key positions such as shops, elevators, and restrooms, and clicking on them can view the details. When detecting high-density passenger flow or temporarily closed areas ahead, the AR interface displays warning signs.

[0016] Preferably, the path planning algorithm is an improved A* algorithm, and its cost function integrates the passenger flow density weight, the user's walking speed preference, and the facility queuing duration. On the basis of the classical A* algorithm, the passenger flow density weight (D(n)) and the facility waiting time (W(n)) are introduced, and the cost function is adjusted to:

[0017] f(n) = g(n) + h(n) + α·D(n) + β·W(n)

[0018] In the formula: g(n): The actual movement cost from the starting point to the current node;

[0019] h(n): The heuristic estimated value;

[0020] D(n): The passenger flow density penalty value of node n;

[0021] W(n): The estimated waiting time of the facility associated with node n;

[0022] α, β: Adjustable parameters, by default α = 0.3, β = 0.2, and dynamically optimized through machine learning.

[0023] Preferably, the method for location navigation based on mobile cloud passenger flow data includes the following steps:

[0024] S1. Use a mobile phone or smart watch to obtain the real-time location through GPS, Wi-Fi, or Bluetooth beacon, remove sensitive information such as the device's MAC address and IMEI, retain the device's timestamp, longitude and latitude coordinates, and moving speed, and upload the data to the cloud using the MQTT protocol;

[0025] S2. Analyze the passenger flow, use the DBSCAN algorithm, set the neighborhood radius to 5 meters, and the minimum number of cluster points to 10, output the boundary of the high-density area, and at the same time calculate the dynamic congestion index, and calculate the number of people per square meter based on kernel density estimation;

[0026] S3. Generate a personalized route by combining user tags and the real-time congestion index.

[0027] Preferably, the dynamic congestion index calculation method includes: First, calculate the regional pedestrian flow density based on kernel density estimation, use the Gaussian kernel function to perform spatial smoothing on discrete positioning points, and the bandwidth parameter h is dynamically adjusted according to the venue type. Further introduce the moving speed variance as a correction factor. When the standard deviation of the crowd moving speed σ < 0.2m / s is detected, the system automatically increases the congestion index by 20% to distinguish between static queuing and slow moving states, and finally outputs a standardized congestion level of 0-100, and this calculation is updated every 30 seconds.

[0028] Preferably, when it is detected that the user deviates from the route or there is a sudden peak in passenger flow, the system triggers a dynamic replanning mechanism, quickly updates the path through the edge computing node, and the replanning process adopts a hierarchical response strategy. When there is a slight deviation, it is quickly corrected locally by the terminal. When there is a major change, it requests the edge node to calculate. When it is an extreme situation, it switches to the manual service channel.

[0029] Preferably, the dynamic replanning mechanism includes three triggering conditions. First, the user deviates from the planned route by more than 5 meters for 30 seconds. Second, the congestion index of the forward path suddenly increases by more than 20%. Third, the status of the target POI changes.

[0030] Preferably, during the navigation process, the waiting time prediction and preferential information of surrounding facilities are pushed in real time. The push frequency is negatively correlated with the user's moving speed, and the types of push content include facility waiting time, time-limited offers, and safety tips.

[0031] The beneficial effects of the present invention:

[0032] The system adopts real-time passenger flow analysis and dynamic path planning technologies, which can accurately predict and avoid congested areas, greatly improving the navigation efficiency. Through the collaborative design of the federated learning framework and edge computing, while ensuring the privacy and security of user location data, the system response delay is controlled within 500 ms, with a low response delay. Finally, the AR augmented reality navigation interface, combined with the personalized recommendation algorithm, supports special needs such as barrier-free navigation, enabling different user groups to obtain a customized navigation experience. Brief Description of the Drawings

[0033] Figure 1 It shows a schematic diagram of the system framework of the present invention;

[0034] Figure 2 It shows a schematic diagram of the system method flow of the present invention. Detailed Description of the Invention

[0035] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0036] Please refer to Figure 1 , the present invention provides an embodiment: a location navigation device based on mobile cloud passenger flow data, including:

[0037] A cloud server for collecting and processing anonymous passenger flow data from mobile terminals in real time to generate a dynamic heat map;

[0038] An indoor and outdoor positioning module for determining the current location of the user through multi-source fusion positioning technology;

[0039] A navigation terminal for generating an optimal navigation route and outputting navigation guidance according to user preferences and real-time passenger flow density through a path planning algorithm.

[0040] Furthermore, the cloud server adopts a distributed microservices architecture, including a data collection cluster, a computing cluster, and a storage cluster, which can realize the full-process automated processing of data collection, cleaning, storage, analysis, and prediction.

[0041] The cloud server further includes:

[0042] A data cleaning unit for filtering abnormal location data;

[0043] A clustering analysis unit for identifying high-density passenger flow areas using the DBSCAN algorithm;

[0044] The prediction module predicts the passenger flow distribution in the future period based on historical data and the LSTM model.

[0045] The multi-source fusion positioning technology adopts a loose coupling architecture, including Wi-Fi positioning and Bluetooth beacons. Wi-Fi positioning is based on RSSI fingerprint library matching with an accuracy of 3 - 5 meters, and the deployment density of Bluetooth beacons is 4 per 100 ㎡ with an accuracy of 1 - 3 meters.

[0046] Furthermore, the data cleaning unit uses an anomaly detection algorithm, specifically as follows:

[0047] def clean_data(points): # Speed filtering (exclude abnormal points > 5m / s)

[0048] valid = [p for p in points if p.speed < 5] # Trajectory smoothing (Savitzky-Golay filtering)

[0049] return savgol_filter(valid, window = 5, polyorder = 2);

[0050] The clustering analysis unit uses an improved DBSCAN algorithm:

[0051]

[0052]

[0053] Among them, the eps value is the neighborhood radius, and min_samples is the minimum number of clustering points.

[0054] The navigation terminal integrates an AR display function, which captures the user's surrounding environment in real time through a camera and overlays virtual navigation information on the screen. The virtual navigation information includes dynamic arrows, hotspot labels, and dynamic obstacle avoidance prompts. The dynamic arrows display direction guidance in the real scene according to the path planning result, and the user only needs to move following the arrows. At the same time, floating labels are overlaid at key positions such as shops, elevators, and restrooms, and clicking on them can view details. When a high-density passenger flow or a temporarily closed area is detected ahead, a warning sign is displayed on the AR interface.

[0055] The path planning algorithm is an improved A* algorithm, and its cost function fuses the passenger flow density weight, the user's walking speed preference, and the facility queuing duration. On the basis of the classical A* algorithm, the passenger flow density weight (D(n)) and the facility waiting time (W(n)) are introduced, and the cost function is adjusted to:

[0056] f(n) = g(n) + h(n) + α·D(n) + β·W(n);

[0057] where: g(n): the actual moving cost from the starting point to the current node;

[0058] h(n): heuristic estimated value;

[0059] D(n): the passenger flow density penalty value of node n;

[0060] W(n): the estimated waiting time of the facilities associated with node n;

[0061] α, β: adjustable parameters, by default α = 0.3, β = 0.2, dynamically optimized through machine learning

[0062] Please refer to Figure 2 , a method for location navigation based on mobile cloud passenger flow data, including the following steps:

[0063] S1. Use a mobile phone or smart watch to obtain the real-time location through GPS, Wi-Fi or Bluetooth beacon, remove sensitive information such as the MAC address and IMEI of the device, retain the timestamp, longitude and latitude coordinates and moving speed of the device, and upload the data to the cloud using the MQTT protocol;

[0064] S2. Analyze the passenger flow, use the DBSCAN algorithm, set the neighborhood radius to 5 meters and the minimum number of clustering points to 10, output the boundary of the high-density area, and at the same time calculate the dynamic congestion index, and calculate the number of people per square meter based on kernel density estimation;

[0065] S3. Generate a personalized route by combining user tags and the real-time congestion index.

[0066] The calculation method of the dynamic congestion index includes: first, calculate the regional population flow density based on kernel density estimation, perform spatial smoothing processing on discrete positioning points using a Gaussian kernel function, and the bandwidth parameter h is dynamically adjusted according to the type of venue. Further, introduce the variance of the moving speed as a correction factor. When the standard deviation of the crowd moving speed σ < 0.2 m / s is detected, the system automatically increases the congestion index by 20% to distinguish between the stationary queue and the slow moving state, and finally outputs a standardized congestion level of 0 - 100, and this calculation is updated every 30 seconds.

[0067] When it is detected that the user deviates from the route or there is a sudden peak in passenger flow, the system triggers a dynamic replanning mechanism, quickly updates the path through the edge computing node, and the replanning process adopts a hierarchical response strategy. When there is a slight deviation, it is quickly corrected locally by the terminal. When there is a major change, it requests the edge node to calculate. When it is an extreme situation, it switches to the manual service channel.

[0068] The dynamic replanning mechanism includes three triggering conditions. First, the user deviates from the planned route by more than 5 meters for 30 seconds. Second, the congestion index of the front path suddenly increases by more than 20%. Third, the status of the target POI changes.

[0069] During the guided tour, waiting time predictions and discount information of surrounding facilities are pushed in real time. The push frequency is negatively correlated with the user's movement speed. The push content types include facility waiting time, limited-time discounts and safety tips.

[0070] Furthermore, the core process of the present invention can be divided into three stages: data collection and processing, passenger flow analysis and prediction, and personalized route planning. The specific implementation steps are as follows:

[0071] The system first collects anonymous user location data in real time through multi-source fusion positioning technology (Wi-Fi / Bluetooth / GPS) of mobile terminals (smartphones, smart wearable devices). The sampling frequency is set to 1Hz. The original data is lightweight encrypted and transmitted to the edge computing node through the MQTT protocol. Data cleaning and standardization are performed to eliminate invalid data. Finally, a standardized trajectory point set is generated, which contains a triplet of timestamp, coordinates and moving speed, and the preprocessed data is uploaded to the cloud.

[0072] The cleaned real-time data is input into the three-layer analysis engine: first, based on the density clustering module of the improved DBSCAN algorithm (dynamic eps parameter), a heat map of the entire area is generated every 30 seconds, thereby identifying hot spots with a passenger density of >3 people / ㎡; second, the dynamic congestion index is calculated, integrating the kernel density estimation result and the moving speed variance (formula: CI=KDE×(1+1 / σ2)), and outputting a 0-100 standardized index; third, the LSTM prediction model uses the previous 6 hours of data as input to predict the passenger flow distribution trend in the next 15-30 minutes.

[0073] When a user initiates a navigation request, the system performs multi-objective decision-making: first, preset weights (α, β) are loaded according to the user's age, preference and mobility; second, an improved A* algorithm is used to calculate k candidate paths (k=3), and its cost function integrates distance, real-time congestion index, and facility waiting time (such as restaurant queues); third, the AR engine renders the optimal path: sub-meter positioning is achieved through SLAM technology, and dynamic navigation arrows (green for smooth traffic / yellow for warning / red for congestion) and POI bubble labels are superimposed on the camera screen. Environmental changes are continuously monitored during navigation. If a route deviation of >5m is detected or the congestion index ahead suddenly increases by 15%, local re-planning is immediately triggered to plan a new route.

[0074] Furthermore, for the calculation of the real-time congestion index, the spatial crowd density distribution is first calculated based on the Gaussian kernel density estimation (KDE). The bandwidth parameter h is adaptively adjusted according to the type of scene, such as h = 3m for shopping malls or h = 5m for airports, to generate the initial density value KDE∈[0,1]. At the same time, the speed variance σ2 is calculated through the trajectory data reported by the mobile terminal. When σ2<0.04m2 / s 2 When it is determined to be in a stagnant state, the reciprocal of variance weighting (1 / (σ2+ε), where ε = 0.01 is used to prevent division-by-zero errors) is triggered. Finally, the result is mapped to the 0-100 standardized interval through the normalization formula CI = 100×tanh(0.5×KDE×(1+1 / (σ2+ε))), and three levels of congestion are classified: 0-30 (smooth / green), 31-70 (slow / yellow), 71-100 (congested / red).

[0075] Through the above steps, by adopting real-time passenger flow analysis and dynamic route planning technologies, it is possible to accurately predict and avoid congested areas, greatly improving the navigation efficiency. In addition, the use of an AR augmented reality navigation interface combined with a personalized recommendation algorithm supports special needs such as barrier-free navigation, enabling different user groups to obtain a customized navigation experience, so as to solve the problem that traditional navigation systems rely on static map data and cannot perceive real-time passenger flow changes, resulting in possible congestion on the planned route.

Claims

1. A location navigation device based on mobile cloud passenger flow data, characterized in that, It includes: A cloud server for collecting and processing anonymous passenger flow data from mobile terminals in real time and generating a dynamic heat map; An indoor and outdoor positioning module for determining the user's current location through multi-source fusion positioning technology; A navigation terminal for generating an optimal navigation route and outputting navigation guidance according to user preferences and real-time passenger flow density through a path planning algorithm.

2. The location navigation device based on mobile cloud passenger flow data according to claim 1, characterized in that The cloud server further includes: A data cleaning unit for filtering abnormal location data; A clustering analysis unit for identifying high-density passenger flow areas using the DBSCAN algorithm; A prediction module for predicting the passenger flow distribution in the future period based on historical data and the LSTM model.

3. The location navigation device based on mobile cloud passenger flow data according to claim 1, characterized in that: The multi-source fusion positioning technology adopts a loose coupling architecture, including Wi-Fi positioning and Bluetooth beacons. Wi-Fi positioning is based on RSSI fingerprint library matching, and the deployment density of Bluetooth beacons is 4 per 100㎡.

4. The location navigation device based on mobile cloud passenger flow data according to claim 1, characterized in that: The navigation terminal integrates an AR display function, which captures the user's surrounding environment in real time through a camera and superimposes virtual navigation information on the screen. The virtual navigation information includes dynamic arrows, hotspot labels, and dynamic obstacle avoidance tips. The dynamic arrows display direction guidance in the real scene according to the path planning result. The user only needs to follow the arrows to move. At the same time, floating labels are superimposed at key positions such as shops, elevators, and restrooms, and details can be viewed by clicking. When a high-density passenger flow or a temporarily closed area is detected ahead, a warning sign is displayed on the AR interface.

5. The location navigation device based on mobile cloud passenger flow data according to claim 1, wherein: The path planning algorithm is an improved A* algorithm. Its cost function integrates the passenger flow density weight, the user's walking speed preference, and the queue waiting time of facilities. On the basis of the classic A* algorithm, the passenger flow density weight (D(n)) and the facility waiting time (W(n)) are introduced, and the cost function is adjusted to: f(n) = g(n) + h(n) + α·D(n) + β·W(n) Where: g(n): The actual movement cost from the starting point to the current node; h(n): The heuristic estimate value; D(n): The passenger flow density penalty value of node n; W(n): The estimated waiting time of the facility associated with node n; α,β: Adjustable parameters, with the default values of α = 0.3 and β = 0.2, dynamically optimized through machine learning.

6. A method for location navigation based on mobile cloud passenger flow data, using the location navigation device based on mobile cloud passenger flow data described in claims 1-5, characterized in that, It includes the following steps: S1, Use a mobile phone or smart watch to obtain the real-time location through GPS, Wi-Fi, or Bluetooth beacon, remove sensitive information such as the device's MAC address and IMEI, retain the device's timestamp, longitude and latitude coordinates, and movement speed, and upload the data to the cloud using the MQTT protocol; S2, Analyze the passenger flow volume, use the DBSCAN algorithm, set the neighborhood radius to 5 meters, and the minimum number of clustering points to 10, output the boundary of the high-density area, and at the same time calculate the dynamic congestion index and calculate the number of people per square meter based on kernel density estimation; S3, Generate a personalized route by combining user tags and the real-time congestion index.

7. The location navigation device and method based on mobile cloud passenger flow data according to claim 6, characterized in that, The method for calculating the dynamic congestion index includes: First, calculate the regional pedestrian flow density based on kernel density estimation. Use a Gaussian kernel function to perform spatial smoothing on discrete positioning points. The bandwidth parameter h is dynamically adjusted according to the type of venue. Further introduce the variance of the moving speed as a correction factor. When the standard deviation of the crowd moving speed σ < 0.2 m / s is detected, the system automatically increases the congestion index by 20% to distinguish between the stationary queuing and slow moving states. Finally, output a standardized congestion level from 0 to 100, and this calculation is updated every 30 seconds.

8. The method for location navigation based on mobile cloud passenger flow data according to claim 6, characterized in that: When it is detected that the user deviates from the route or there is a sudden peak in passenger flow, the system triggers the dynamic replanning mechanism, quickly updates the path through the edge computing node. The replanning process adopts a hierarchical response strategy. When there is a slight deviation, it is quickly corrected locally by the terminal. When there are major changes, it requests the edge node to calculate. When it is an extreme situation, it switches to the manual service channel.

9. The method of location navigation based on mobile cloud passenger flow data according to claim 8, characterized in that: The dynamic replanning mechanism includes three triggering conditions. First, the user deviates from the planned route by more than 5 meters for 30 seconds. Second, the congestion index of the forward path suddenly increases by more than 20%. Third, the status of the target POI changes.

10. The method of location navigation based on mobile cloud passenger flow data according to claim 6, characterized in that: During the navigation process, the waiting time prediction and preferential information of surrounding facilities are pushed in real time. The push frequency is negatively correlated with the moving speed of the user. The types of push content include facility waiting time, time-limited offers, and safety tips.

Images